IRC logs for #openttd on OFTC at 2007-11-06
⏴ go to previous day
00:00:18 <SmatZ> TrueBrain: the way to reply to the discussion is by email, right?
00:00:33 <TrueBrain> SmatZ: yup :) That is the idea anyway ;)
00:00:33 <Unkown_Entity> I'm not sure about the graphics processing right now. i'll have to check some more. certainly something would have to be done by hardware otherwise the ARM9 shouldn't be enough to handle 3D
00:01:03 <TrueBrain> Unkown_Entity: depends on how much 3D is done :) After all, a GPU is a CPU, only optimized for graphics ;)
00:01:43 <ln-> how big a map is needed to implement a processor capable of running OTTD?
00:01:44 <TrueBrain> btw, Unkown_Entity, nice to see someone working on a DS port :)
00:02:01 <TrueBrain> ln-: sorry, I couldn't parse that sentence.. it can go two ways :)
00:02:13 <Unkown_Entity> TrueBrain: it's got "Hardware fog, lighting, and transformation along with non blending texture mapping, toon-shading, and edge anti-aliasing". so there must be some hardware graphics optimization in there ;)
00:02:16 <exe> TrueBrain: pathfinder just does readonly access to _m, can't hooks pause the pathfinder when changing rail graph?
00:02:25 <TrueBrain> Unkown_Entity: agree'd
00:02:39 <ln-> TrueBrain: implementing a processor with those
00:02:42 <TrueBrain> exe: yes, we call it locking ;)
00:02:50 <Unkown_Entity> TrueBrain: but i don't want to get anyones hopes up yet. apart from a pong game i haven't done any DS coding yet. :p
00:03:01 <TrueBrain> ln-: ah! You mean that :) I have been working on it myself, I only miss signal types :)
00:03:13 <TrueBrain> Unkown_Entity: doesn't matter :) PSP port stranded too
00:03:21 <TrueBrain> but to have someone working on it is nice :)
00:03:46 <TrueBrain> ln-: and I made it that trains were the electrons :)
00:03:53 <TrueBrain> the clock speed only was 3 game-days
00:04:05 <TrueBrain> (with rigger maglev to have instant acceleration :p)
00:04:07 <ln-> hmm, i think even a 32-bit register could take quiiiiite much space on the map.
00:04:17 <Unkown_Entity> a DS port would make much more sense than a PSP port. with the stylus and touchpad it would be much more comfortable
00:04:18 <TrueBrain> ln-: it should be 8bit CPU
00:04:23 <TrueBrain> anything bigger is just killing yourself
00:04:29 <TrueBrain> Unkown_Entity: I agree
00:05:27 <TrueBrain> ln-: I don't like tha tpage, it needs manual changing tracks :) That aint real!
00:06:37 <TrueBrain> hmm.. if we make trains electrons.. the circuit is as fast as the train
00:06:49 <TrueBrain> can we make signals the electrons.. hmm..
00:07:30 <SmatZ> are you reviving the NAND/NOR signal logic? :-)
00:07:49 <TrueBrain> just for fun, can it be done with current signals...
00:08:02 <TrueBrain> a ticker is simple :p
00:09:29 <TrueBrain> hmm, ln-, what kind of computer do you want?
00:09:30 <SmatZ> gate with delay in days (game days:)
00:09:41 <TrueBrain> building an ALU might be simple
00:09:45 <TrueBrain> so, the unit that does all the logic
00:09:51 <TrueBrain> but that doesn't make a CPU :)
00:10:09 <TrueBrain> no, you need trains to be electrons
00:10:28 <TrueBrain> so, you need a recycle path :p
00:11:08 <ln-> TrueBrain: a computer doing addition with given parameters could be nice
00:11:19 <TrueBrain> ln-: gimme example :)
00:11:30 <TrueBrain> 'with given parameters' still is too vague :)
00:11:36 <TrueBrain> lets limit ourself to a 4bit computer btw
00:13:01 <ln-> hmm, well, a computer that calculates 2+3, for example
00:13:20 <Ailure> have anyone ever used one?
00:13:29 <Ailure> lowest I gone was like, 8-bit processor
00:13:31 <TrueBrain> Ailure: if you can make a 4bit, you can make N bit
00:13:33 <Ailure> where I used 6502 assembly
00:13:39 <SmatZ> maybe calculators have 4bit CPUs
00:14:30 <Ailure> one amusing thing I thought on was like
00:14:46 <Ailure> programming a i386 virtual machine for 6502 processors
00:15:16 <Ailure> there will so much fun with bankswitching
00:15:29 *** mist`shoo has joined #openttd
00:15:37 <SmatZ> I had an idea about implementing 80386 at 80286
00:15:49 <SmatZ> I ended with 80286 emulator running at 80286
00:15:57 <SmatZ> so it was like 20 times slower 80286
00:16:35 <Ailure> that's always the case for virtual machines
00:16:42 <mist`shoo> why my aircrafts have (after time) "empty command" when I set "depot service" early?
00:16:43 <Ailure> there are tricks to reach great speed
00:17:23 <mist`shoo> i mean "void order"
00:17:40 <SmatZ> Ailure: it depends what you call 'virtual machine' :) some processors allow running of legacy programs, like VM86, or even real-mode on new x86, amd64, CPUs
00:17:53 <Ailure> void order should only appear after you destroyed the destination of said order
00:17:56 <ln-> TrueBrain: feature request: ability to have wagons from several companies in one train, and drop them off one by one at stations.
00:17:58 <SmatZ> mist`shoo: what version are you running?
00:17:59 <TrueBrain> ln-: if you revert, that red is signal, and green is no-signal, it becomes much easier :)
00:18:02 <Ailure> you might have stumbled upon a bug if that's not the case
00:18:25 <TrueBrain> ln-: a real game request, or? As I don't see what you mean...
00:18:33 <TrueBrain> mist`shoo: you removed your airport?
00:18:35 <Ailure> SmatZ: Heh true, and I suspect that's how the case for most virtual machines
00:18:48 <Ailure> Some virtual machines do fun stuff
00:18:48 <SmatZ> Ailure: as my bachelor work, I made a 80x86 emulator, that run ~7x slower than original machine... it uses dynamic recompilation
00:18:57 <ln-> TrueBrain: additionally, "public" tracks, and tracks that only allow trains with a wagon of the same company as last wagon.
00:18:59 <SmatZ> it could be done better...
00:19:04 <Ailure> like translating the machine code of the spefic processor
00:19:09 <Ailure> into machine code of another processor
00:19:19 <ln-> TrueBrain: Finite State Automata
00:19:30 <Ailure> The SNES emulator ZSNES does that I belive
00:19:38 <TrueBrain> I think I know what you are getting at
00:19:48 <TrueBrain> that would allow you to make a microcontroller
00:20:22 <SmatZ> Ailure: I think even DOSBox does that, but it is really slow :-/
00:20:23 <TrueBrain> I HATE that Shift is fastforward in debug mode...
00:20:24 <mist`shoo> TrueBrain no, airport is still there
00:20:43 <mist`shoo> TrueBrain i have 0.4.8 version
00:20:48 <Ailure> I thought on doing that as a Bachelor project
00:21:00 <TrueBrain> 0.4.8 is long long ago :) 0.5.3 is latest
00:21:21 <Ailure> but i'm thinking about taking another year to get the equilant of Master degree
00:21:31 <Ailure> but I still have to do a project
00:21:34 <Ailure> well, it's either a project of my own
00:21:41 <Ailure> or something assigned by a external company
00:21:50 <mist`shoo> because i was playing TTD long time ago and im back - fresh, smart but lame in all new open stuff :)
00:21:59 <ln-> TrueBrain: i'm not sure what automata can do in hardware, but they can be used for equivalents of regular expressions.
00:22:32 <Ailure> I actually started on a small X86 emulator
00:22:34 <TrueBrain> hmm, is a NOT port possible..
00:23:16 <Ailure> the way I did it was kind of clumsy though lol
00:23:57 *** mist`shoo has joined #openttd
00:23:57 <Ailure> the whole emulator was basically a huge switch statement inside a infinite loop
00:24:10 <Ailure> very simplistic processor too
00:24:11 <TrueBrain> Ailure: aren't all emulators, at some level? :p
00:24:19 <TrueBrain> I am creating a MIPS emulator, also fun fun :p
00:24:30 <SmatZ> TrueBrain: like, you are emulating CPU in hardware? :)
00:24:34 <Ailure> but I just looked it up and there was better ways
00:24:35 <TrueBrain> how to get a signal GREEN when something is on the track?
00:24:40 <Ailure> for the programming language I was using (Java)
00:24:46 <TrueBrain> SmatZ: hmm.. not really :p
00:24:48 <Ailure> someone have written a fast Playstation emulator in Java apparently
00:24:52 <ln-> btw, did you know that C can be converted to run under Java virtual machine?
00:24:54 <Ailure> and I read his presentation slides
00:24:55 <TrueBrain> we made a datapath,w e made microcode, and now implementing it in C++ :P
00:25:35 <SmatZ> hey this channel is full of people writing emulators! :-D
00:25:38 <ln-> correction: _compiled_ C can be converted to run under JVM
00:25:40 <Ailure> but I know it's bit... unusual
00:26:26 <Ailure> C is only used when you need to break out of the sandbox
00:26:32 <SmatZ> ln-: even on a different architecture?
00:26:40 <Ailure> also called native classes
00:27:31 <Ailure> you kinda lose the point
00:27:38 <Ailure> with Java by using JNI
00:29:19 <exe> it's nowhere as fun as {Console::WriteLine("Hello World - Managed");std::cout << "Hello World - Native" << std::endl;} of c++/cli
00:29:41 <TrueBrain> bah, I hate NOT-ports :p
00:30:17 <ln-> SmatZ: the idea was to (cross)compile C to a MIPS binary, and somehow convert that to native java bytecode. (i only read about this, didn't try)
00:31:18 <SmatZ> ln-: MIPS instruction set and Java bytecode are someway similiar? (I don't know)
00:32:17 <Ailure> bytecode is made to be interpreted by software
00:32:26 <Ailure> machine code is made for being interpreted by hardware
00:32:41 <Ailure> bytecode tend to make a few shortcuts becuse of this
00:32:59 <ln-> weren't there supposed to be processors that run Java natively?
00:33:47 <TrueBrain> lol, I can stop the clock... now that is just wrong...
00:33:52 <TrueBrain> signals in OpenTTD really are fucked :p
00:33:55 <TrueBrain> (directional signal)
00:34:08 *** mist`shoo has joined #openttd
00:36:13 <Ailure> the java bytecode isn't too widely docuemanted I guess
00:36:18 <Ailure> even if you can probably find a reference for it
00:36:25 <SmatZ> I am not sure whether Java is a good platform to develop CPU emulators... if you demand speed
00:36:38 <Ailure> well it's a good language
00:36:58 <Ailure> some of the problems related with java have been resolved too
00:37:11 <ln-> only it's not a good language for teaching programming, for which it is often used.
00:37:32 <Ailure> been proved it can be as fast as nativly compiled code, but things like startup time is still needed
00:37:32 <Tefad> ugh java makes me want to punch things when i use it
00:37:42 <Ailure> it made me a better programmer
00:37:47 <Ailure> even if learning it was hell
00:38:07 <Tefad> ok, for the most part i understand streams and the like, however the whole GUI thing is annoying
00:38:08 <Ailure> I realized how much that was possible in C but not in Java
00:38:12 <SmatZ> I never got used to OOP.... Java is a real pain for me :-/
00:38:12 <Ailure> usually was bad practice to do anyway
00:38:19 <Tefad> i think GUI programming in general is an annoyance to me though : \
00:38:36 <Ailure> Windows API is supposed to be hell
00:38:39 <Ailure> but I never touched it
00:38:53 <Ailure> Java API isn't too hard to use
00:38:56 <Tefad> well, i've poked it i guess, but never got down to nittygritty
00:38:57 <Ailure> the only problem is that well
00:39:02 <Ailure> it assumes you know OOP
00:39:08 <Ailure> don't grasp OOP and you will have problems
00:39:16 <Tefad> i wrote a program that would beep the PC speaker in 9x or NT for x86 platforms.
00:39:29 <Ailure> you don't need a program for that
00:39:54 <Ailure> you can write a infinite loop bat that outputs the bell signal
00:39:56 <ln-> class names in Java are at least 500 letters long and due to the stupidity of having to allocate everything with new, each of them needs to be typed at least twice.
00:39:58 <Ailure> was really fun to annoy people with
00:40:14 <Tefad> bell signal can be interpreted by windows to play "ding.wav" or whatever
00:40:49 <Ailure> I thought the bell signal
00:40:59 <Ailure> it's mostly a leftover from early computing days
00:41:04 <Ailure> where the huge mainframe actually had a bell
00:41:08 <Ailure> that would ring at signal
00:41:26 <Ailure> I think most of the early giants to computers had even
00:41:26 <Tefad> win32 API was developed initially for NT iirc. win32 has Beep(). however when win32 was ported to win3.1 (and carried on into win9x) Beep() was only a stub.
00:41:36 <Ailure> so researchers would be able to tell when it was done
00:41:44 <Ailure> since it took a half eternity to get done sometimes
00:41:51 <Tefad> so to make a beep in 16bit derived windows (win32s win9x/me) one has to directly interface with the timer hardware.
00:42:30 <Ailure> Win95 wasn't supposed to be made apparently
00:42:38 <ln-> do modern day printers have bells?
00:43:11 <Ailure> I doubt even my first printer does <<
00:43:16 <ln-> i had one in the 90's that did have
00:43:32 <Ailure> then it's something I never got to try
00:43:36 <Ailure> since it's something I didn tknow about
00:43:59 <ln-> it would beep if it was given the bell byte
00:44:16 <mist`shoo> TrueBrain - ive installed newest version and i have "bushes" in polish language - i mean: ??ywno?c
00:44:53 <TrueBrain> mist`shoo: language problems should go to the language depertmant :)
00:44:59 <TrueBrain> (see www.openttd.org, I believ eunder development)
00:45:07 <TrueBrain> argh, this NOT port is driving me crazy :)
00:45:16 <TrueBrain> I think it is easier to make an OpenTTD modification :p
00:45:54 <mist`shoo> but i have used 0.4.8 and there was correct polish lang
00:46:10 <TrueBrain> mist`shoo: I don't understand you, sorry
00:46:18 <TrueBrain> you mean you see ? chars
00:46:25 <SmatZ> mist`shoo: download unifont.grf and include it in openttd.cfg, inder newgrf_static ... or how is it called
00:46:46 <TrueBrain> SmatZ: no, we have easier solutions for that these days
00:46:47 <mist`shoo> questionmark is an example
00:46:56 <TrueBrain> mist`shoo: I suggest to read the README, it tells you to download a font
00:46:59 <Ailure> this is probably not too hard to do
00:47:07 <SmatZ> TrueBrain: ok, sorry then
00:47:08 <Ailure> attach a bell to the computer
00:47:19 <Ailure> make it ring when the bell byte been printed out
00:47:28 <Ailure> this probably wouldn't work well on any closed systems :)
00:47:44 <Ailure> but should be easy for open systems
00:47:57 <SmatZ> honestly, 0.5 was the version when I stopped using Czech language, because of '?' everywhere... even unifont doesn't have all required characters, and I don't want to use different font...
00:48:57 <mist`shoo> SmatZ - but, you know, i dont want to change my language because of upgrade... was good, is wrong now :/
00:49:25 <SmatZ> mist`shoo: I totally understand
00:49:59 <Rafagd> mist`shoo: get the source, correct the errors, send the patch =]
00:50:07 <SmatZ> TrueBrain: did I say something wrong?
00:50:07 <TrueBrain> Rafagd: it aint an error :)
00:50:12 <TrueBrain> SmatZ: no, I was wondering your logic
00:50:20 <TrueBrain> load a proper font file and you loose the ?
00:50:31 <TrueBrain> and there are font files which look just like the default OpenTTD ones :)
00:50:45 <Ailure> reminds me about a stable release
00:50:48 <SmatZ> TrueBrain: I know only unifont, but it miss some glyphs
00:50:50 <TrueBrain> the whole reason why we introduced Unicode :)
00:50:50 <Ailure> that did default to russian for some fun reason
00:51:04 <TrueBrain> SmatZ: 0.5 and on supports Unicode, you know that, right? :)
00:51:57 <Ailure> it reminds me about how TTO did stuff
00:51:57 <TrueBrain> so mist`shoo, there is no reason to change languages, just set a font, and you will see all your lovely chars :)
00:51:57 <TrueBrain> Ailure: are you in a 'remember' mood? You sound like Bjarni :)
00:51:57 <Ailure> since the station symbols technically are special charcthers
00:52:04 <_42_> SmatZ, Bjarni (~Bjarni@0x50c79a03.virnxx14.adsl-dhcp.tele.dk) was last seen quitting #openttd.notice 1 day 23 hours 9 minutes ago (04.11. 01:42) stating "Quit: Leaving" after spending 3 hours 19 minutes there.
00:52:14 <Ailure> so using would sometimes result into station icons being written
00:52:33 <Ailure> I do live next to Denmark.
00:52:37 <SmatZ> yes, you could insert that characters into savegame, maybe even company name
00:52:42 <Ailure> Live in southest part of Sweden.
00:53:57 <mist`shoo> TrueBrain - ok, so im on wiki and? shoud i change e.g. small_font = to small_font = Arial ?
00:53:57 <Ailure> I hadn't seen many signs of Bjarni being stereotypically danish though.
00:53:57 <Ailure> Then, the Swedish stereotypical image of a danish person is that of a drunk person.
00:55:06 <exe> TrueBrain: i still think pathfinder could be possibly multithreaded :) all the pathfinder threads get a message before map change and after it. Pathfinder threads can look for a new message in outermost loop (not much locking) and It's not a problem if the main game has to wait even 0.5 milisecond for pathfinders to acknowledge map change.
00:55:17 <SmatZ> Ailure: Danish must love you :-)
00:55:39 <Ailure> the history between denmark and Swedish
00:57:22 <TrueBrain> exe: I suggest, create it :)
00:58:16 <Ailure> Called Tyrant by Swedes
01:00:25 <TrueBrain> now my NOT port gives signals at the speed of the clock ;)
01:00:38 <TrueBrain> which... is not what I wanted :p
01:00:48 <exe> TrueBrain:i will try to simulate that without game
01:02:14 *** LeviathNL has joined #openttd
01:03:58 <SmatZ> I spent too much time here today
01:08:37 <huma> he's gone for another couple of weeks..
01:09:48 <TrueBrain> bah bah bah, and again, bah
01:09:54 <TrueBrain> it looks impossible :p
01:11:23 <TrueBrain> I guess we need the NAND patch :)
01:14:12 *** Mussolini has joined #openttd
01:14:15 *** Markkisen has joined #openttd
01:14:39 <TrueBrain> hmm, the one on the website has the same problem
01:14:42 <TrueBrain> sometimes it does give a signal
01:20:37 <TrueBrain> well, I guess it has to be..
01:20:55 <TrueBrain> btw, most part of the logic of the NOT gate on that website, is redundant
01:21:03 <TrueBrain> (the whole 'b' channel)
01:21:19 *** Mussolini is now known as Benito
01:21:36 *** Markkisen is now known as MarkSlap
01:31:20 *** Eddi|zuHause2 has joined #openttd
01:34:38 <TrueBrain> trains without orders can do strange things
01:40:02 <TrueBrain> I don't udnerstand combo signals :)
01:40:07 <TrueBrain> who can explain them to me? :p
01:43:18 *** Jello is now known as Guest27
01:43:38 <TrueBrain> I really did it.... :)
01:43:50 <TrueBrain> a synchronized clock fixed it all :)
01:48:34 <TrueBrain> besides the NOT gate, they are all very space efficient
01:48:41 <TrueBrain> so.. now we can make circuits :)
01:54:54 <TrueBrain> the only restriction is: while clock is ticking, all data is faulty :)
01:55:04 <TrueBrain> bah, everyone went sleeping
01:56:16 <Unkown_Entity> one gameloop takes 58 ms on emulator. i'll better go to bed and wait for the postman with my hardware tomorrow. g'night ;)
01:56:26 <TrueBrain> night Unkown_Entity
01:56:31 <TrueBrain> btw, 58ms aint bad ;)
02:04:30 *** DaleStan_ has joined #openttd
02:05:29 *** DaleStan_ is now known as DaleStan
02:42:45 *** Maarten has joined #openttd
02:55:48 *** ThePizzaKing has joined #openttd
04:37:10 *** nairan_zZZ has joined #openttd
05:38:09 *** eQualizer has joined #openttd
05:41:03 *** ThePizzaKing has joined #openttd
05:47:57 <mikk36|work> what's the shortcut for level land ?
06:31:04 *** Deathmaker has joined #openttd
06:43:08 *** XeryusTC has joined #openttd
06:56:40 *** Frostregen_ has joined #openttd
07:02:20 *** Frostregen_ is now known as Frostregen
07:06:33 *** Rygrass has joined #openttd
07:10:27 <huma> i was just about to help.. eh..
07:10:52 <Noldo> they are quite quick sometimes
07:17:07 *** Farden123 has joined #openttd
07:23:47 *** Farden123 is now known as Farden
07:36:25 *** Zaviori has joined #openttd
07:58:03 *** |Bastiaan| has joined #openttd
08:02:24 *** Brianetta has joined #openttd
08:25:14 *** stillunknown has joined #openttd
09:18:10 *** Celestar has joined #openttd
09:33:19 <mikl> jaeh, det kan vel diskuteres :)
09:39:23 *** Cool-Matte has joined #openttd
09:53:24 *** Vikthor has joined #openttd
10:05:27 <Rubidium> Celestar: haven't seen you for a long time ;)
10:06:50 <Rubidium> did you do some work on making the 'new' aircraft speeds customisable back when you were still active?
10:10:19 <Celestar> Rubidium: yes, I think so
10:19:17 <Celestar> Rubidium: I'll dig for the stuff later today, k?
10:29:18 *** mucht_work has joined #openttd
10:30:34 <Rubidium> Celestar: that's fine
10:37:03 *** Klanticus has joined #openttd
10:54:09 *** Gonozal_VIII has joined #openttd
10:57:26 *** Tino|R152 has joined #openttd
10:59:25 <Gonozal_VIII> would it be possible to let you buy/sell land based vehicles only at depots, that are connected to one of your docks and then at the end of the month a ship arrives and unloads your new vehicles and picks up the ones you sold? and ships/planes also not instant buy/sell, they come in from outside the map
10:59:48 *** Prof_Frink has joined #openttd
11:00:16 *** dev|ant has joined #openttd
11:00:48 *** egladil has joined #openttd
11:04:41 <Noldo> Gonozal_VIII: or from a factory on the map
11:40:11 *** dihedral has joined #openttd
11:45:21 *** ChanServ sets mode: +v tokai
12:04:06 <TrueBrain> bah, creating a good logic circuit with OpenTTD is hard, because signals are stupid
12:05:54 <TrueBrain> main problem is that signals totally ignore tracklayout
12:06:02 *** Progman has joined #openttd
12:06:18 <dihedral> could they be combined with yapf?
12:06:26 <TrueBrain> at a waste of CPU, yes
12:06:38 <dihedral> some limited yapf :-P
12:06:44 <TrueBrain> ha, finally I got myself a holding line (clock-based)
12:07:18 <TrueBrain> to make a logic circuit, things need to synchronize
12:07:33 <TrueBrain> so, now I made a 4 bit bus, which waits for the clock before the signals are 'unleached' :)
12:08:59 <TrueBrain> for some reason only over time it desynchonizes..
12:12:42 <TrueBrain> it worked a moment ago! :( Now it fails again...
12:12:46 <TrueBrain> either a bug, or something is fishy :)
12:17:49 <TrueBrain> sometimes my 2 trains do go together to the signal block
12:17:52 <TrueBrain> sometimes they don't
12:18:11 * Vikthor hopes that TrueBrains interest in logical circuits in OpenTTD will drive him to code PBS:)
12:18:29 <TrueBrain> Vikthor: doesn't help for logic circuits
12:19:03 <Vikthor> Ok, so maybe newsignals?
12:19:31 <TrueBrain> that would be cheating :)
12:19:39 <Celestar> this Fortran code sucks, and my svnserve is not working either
12:21:08 <Celestar> I mean not even svnadmin does anything
12:21:17 <Celestar> restarting the server doesn't help a lot either
12:23:46 <TrueBrain> this clearly makes it impossible to make any logic circuit in OpenTTD.... :( :(
12:29:05 <TrueBrain> newsignals, or more, routingrestriction is pretty useful
12:30:04 * dihedral is already looking forward to seeing TrueBrains feature... :-)
12:30:20 <TrueBrain> I want to see how TTDp did it :)
12:30:43 <TrueBrain> no, that is trying to keep things a bit the same :)
12:34:35 <dihedral> is ttdp open source?
12:35:01 <TrueBrain> I really never even read about routingrestrictions, I have to say
12:37:13 <TrueBrain> and it is, what, a year old? Lol :)
12:38:57 <Ammler> TrueBrain: can you remember abaout a year or longer ago, OwenS played with NAND
12:39:15 <TrueBrain> but that makes me wonder why doing NAND, and not doing general signals
12:39:44 <Ammler> its a nice thrad, with flipflops etc.
12:41:58 <Celestar> TrueBrain: do you have any idea how long an "svnadmin hotcopy" should take on a REALLY small repo?
12:42:03 <Celestar> because I'm waiting for 3 hours now
12:42:44 <Celestar> (it's 17 megs on disk)
12:48:00 <TrueBrain> Celestar: euh, it takes 30 minutes for the OpenTTD SVN
12:48:14 <Celestar> it seems I have a problem
12:48:38 <Celestar> TrueBrain: svnadmin dump doesn't do anything either :o
12:48:49 <TrueBrain> bdb repos or fsfs repos?
12:49:06 <Celestar> disks are maybe not ok
12:49:12 <Celestar> this is what I'm checking at the moment
12:49:56 <Celestar> the permissions are weird as well
12:50:17 <TrueBrain> if bdb starts to fuck up, it is almost impossible to recover, in my experience
12:50:20 <Celestar> only the one repo doesn't work
12:51:22 <Celestar> drwxr-s--- 2 svn svn 4096 2007-05-09 13:21 db
12:51:34 <TrueBrain> group stick, group read, user read/write
12:51:49 <TrueBrain> allows anyone in group SVN to read
12:51:52 <TrueBrain> and allows svn to write
12:52:11 <TrueBrain> if svn isn't in svn group ;)
12:52:49 <Celestar> isn't there any way to even get an error message out of svn/svnadmin :P
12:53:57 <Celestar> what do I do now? :P
12:54:07 <TrueBrain> as I said, you are fucked :)
12:54:11 <TrueBrain> it might be possible to recover
12:54:19 *** |Bastiaan| has joined #openttd
12:54:22 <TrueBrain> but last time I tried a bdb recover.. well... I ended up switching to fsfs froma backup :)
12:54:23 <Celestar> svnadmin recover just hangs
12:54:31 <TrueBrain> svnadmin recover in fact can take REALLY long
12:54:35 <Celestar> how simple is it to set up fsfs ?
12:54:42 <TrueBrain> svnadmin create with 1.3+
12:54:49 <TrueBrain> default is fsfs nowedays... with good reason
12:55:24 <Celestar> this thing seems really broken
12:55:31 <Celestar> I can't even activate DMA on the disk :o
12:55:46 <dihedral> ouch - that must hurt
12:55:54 <Celestar> I have backups somewhere
12:56:07 * dihedral likes the _somewhere_
12:56:16 <Celestar> unless they've been overwritten because the "svnadmin hotcopy" fucked up
12:57:54 <TrueBrain> as said, it takes a while :p
12:58:03 <Celestar> no, it took 15 seconds
13:00:57 <Celestar> note to self: make sure svnserve is not running when running svnadmin
13:01:15 <TrueBrain> I do it all the time :p
13:01:17 <Celestar> TrueBrain: yeah, it worked. because I just found the backups anyway :P
13:01:35 <TrueBrain> newsignals really are cool :)
13:02:36 <Celestar> svn: Berkeley DB error while opening environment for filesystem /srv/svn/mglet/db:
13:02:46 <TrueBrain> or anyway, the routingrestrictions :)
13:02:53 <TrueBrain> Celestar: swtich to fsfs :)
13:03:03 <Celestar> TrueBrain: I'm trying to
13:03:10 <Celestar> but I first need to dump all the repos in there
13:03:14 <Celestar> and upgrade the server
13:03:20 <Celestar> which is the only mail server at this department
13:03:43 <Celestar> plus the system drive cannot do DMA anymore
13:03:55 <Celestar> an svndump on a 17MB repo drives the load up to 12
13:04:10 <Celestar> and no one can send mail while I do it
13:04:48 <TrueBrain> but by the looks of it, if you don't do it now, it will shut down itself very soon :p
13:04:59 <Celestar> yes, I'm waiting for itself to do so
13:08:50 <Celestar> because that fucking system still runs suse 9.1
13:09:26 <dihedral> Celestar: never change a running system :-D
13:09:55 <dihedral> it was a helping excuse to why it had not been updated...
13:09:56 <TrueBrain> [13:39] <Ammler> its a nice thrad, with flipflops etc. <- where? :p
13:10:19 <Celestar> dihedral: because $BOSS things a 7-year old mailserver is ok
13:12:13 *** Farden123 has joined #openttd
13:12:47 <dihedral> what kind of Boss do you have Celestar
13:12:56 <dihedral> that must be a really greedy person
13:13:05 <dihedral> not even wanting to invest a little in an upgrade
13:13:10 <TrueBrain> dihedral: not really.. just any boss
13:13:15 <Celestar> I'm a grad student ...
13:13:20 <TrueBrain> if it aint broken, don't fix it
13:13:41 <Celestar> TrueBrain: how much more broken than "DMA doesn't work" does it have to get?
13:13:52 <TrueBrain> Celestar: failure to spin-up :p
13:14:14 <dihedral> at least a defect in 50% of the available sectors :-D
13:14:54 <Celestar> I'm going to put a screwdriver through the mainboard
13:15:13 <dihedral> and share it with us :-P
13:19:41 *** Farden123 is now known as Farden
13:20:34 *** Ammller has joined #openttd
13:28:57 <dihedral> what possible advantage would nand signalling bring...?
13:29:06 <TrueBrain> all logic circuits become possible
13:30:43 <TrueBrain> so you can make any programmable signal with it
13:30:48 <TrueBrain> just it consumes space :p
13:30:57 <TrueBrain> in fact, a NAND isn't needed, a NOT does the same
13:31:00 <TrueBrain> but... takes more room :p
13:32:30 <dihedral> how likeley for that to make it's way into trunk?
13:33:21 <TrueBrain> dihedral: who knows :)
13:33:43 <dihedral> was just wondering :-P
13:33:55 <Celestar> "I have activated the remote thingy" --- Me: "Ok it doesn't work, let's check your network cabling" --- "Oh, I need to activate the network for the remotedesktop to work?"
13:34:13 <Celestar> no dumbfart, it works by telepathy
13:34:42 <dihedral> Celestar: sit down and have a mug of hot tea :-P
13:34:48 <dihedral> with a ginger nut cooky
13:40:06 <TrueBrain> dihedral: btw, I more wonder what is more useful... NAND is nice and all, but via, for example, Squirrel, it would be even more useful
13:40:19 <TrueBrain> but, that would be kind of slow, somehow :p
13:40:21 <Celestar> dihedral: I'm having tea already :P
13:40:35 *** HerzogDeXtEr has joined #openttd
13:40:52 * dihedral hopes Celestar is not drinking earl grey
13:45:30 <Celestar> dihedral: why's that?
13:46:00 <dihedral> you could just as well desolve some soap in hot water :-P
13:46:44 <dihedral> just a personal dislike
13:50:14 <Celestar> I like Earl Grey, but not every day
13:55:06 <Celestar> TrueBrain: about the spam .... :)
13:57:39 <Celestar> the stuff we're getting on the mailing lists?
13:57:58 <TrueBrain> I wouldn't know, not signed on to it
13:58:38 <TrueBrain> if you received spam on it, please email me the header
13:58:47 <Celestar> but you wanted to reduce it "big time", right?
13:59:24 <TrueBrain> so you didn't receive any last days?
13:59:56 <Celestar> I think .. *goes checking*
14:00:26 <Celestar> TrueBrain: about 30 messages today
14:00:38 <TrueBrain> from the SVN maillist?
14:00:44 <TrueBrain> don't you mean your openttd.org account?
14:01:02 <TrueBrain> the devmaillist should be dead
14:01:10 <Celestar> where is my account? :P
14:01:18 <Celestar> I really have to sort out my mail accounts
14:01:26 <TrueBrain> no, the devmaillist still is dead
14:01:42 <Celestar> I even get some to truelight@openttd.org :o
14:01:43 <TrueBrain> so I really have no idea about wha tmails you talk
14:01:52 <TrueBrain> Celestar: ah, so that is via your celester@openttd.org address
14:02:01 <TrueBrain> now we are both on a spam-list again
14:02:26 <Celestar> but I think these addies are on plenty of lists anyway already :P
14:02:36 <TrueBrain> it aint that bad, but yes
14:02:43 <TrueBrain> glx: I misspelled his on purpose :)
14:02:56 <TrueBrain> he didn't mine, I hate Celestar for that now :p
14:03:14 <Celestar> so where do I deactivate the forwarding of the openttd.org addy?
14:03:21 <TrueBrain> Celestar: if you want, I can do that for you
14:03:33 <Celestar> TrueBrain: _please_ do so for the time being
14:03:41 <varun> just got one stupid question
14:04:33 <TrueBrain> Celestar: despite the 2 spam-filters, spam nowedays is hard to filter :(
14:04:50 <varun> when i wanna run a dedicated server behind a router ... which ip adress does connect_to_ip have to have? my internet ip or the "lan" ip adress of the router?
14:04:51 * glx needs to check the junk box (76 mails in it)
14:05:58 <TrueBrain> varun: your LAN IP should be fine
14:06:26 <glx> connect_to_ip is only used by client
14:07:18 <varun> well the wiki sez i have to set it to my routers ip adress ...
14:07:30 <TrueBrain> varun: then the wiki is wrong :)
14:07:39 <TrueBrain> for the server-listing it uses the IP it gets from your advertise
14:07:55 <TrueBrain> so you only have to bind to the right IP, which can't go wrong with 1 IP :)
14:13:58 <unkie> i have installed the ECs new indutry vectors. What do i have to do to get vehicles which can actually transport the new goods?
14:16:39 <Belugas> isn't that question been answered on the forums??
14:17:55 <Celestar> TrueBrain: greylisting does a good job
14:18:11 <TrueBrain> Celestar: you are right, the greylist results on an other server are pretty good
14:18:20 <TrueBrain> openttd just runs exim, which didn't had greylisting support last time I tried :(
14:18:52 <Celestar> TrueBrain: I see, we're using greylisting @ work and we get less than one spam message per 24 hours
14:19:06 <TrueBrain> I wish I would reach that number ;)
14:19:24 <TrueBrain> even with greylisting the spam-hit is 60+% last time I checked :)
14:19:32 <Celestar> I wish we would finally all ditch SMTP and use something more sane :P
14:19:35 <TrueBrain> (but okay, on a volume of 10000+, over 100+ domains
14:20:15 <Celestar> don't google it, I just made it up :P
14:20:27 <TrueBrain> make it, and you might get rich ;)
14:22:53 <Celestar> well, let'S so it together :P
14:23:52 <DorpsGek> TrueBrain: I have not seen Scotty.
14:25:34 <Belugas> Scotty is in orbit, right now :P
14:25:41 <Belugas> and for a lng time, i think!
14:28:08 <unkie> Belugas: probably, i guess i'll have a look there
14:29:48 <unkie> yes, my craft leaves in a few mins
14:30:00 <unkie> currently packing some extra astronaut food
14:33:54 <dihedral> you can move faster if you leave off the heat shield
14:37:15 *** Diabolic-Angel has joined #openttd
14:44:54 <Belugas> dihedral, the guy who played Scotty died and asked (by testament) for his hashes to be sent in orbit, which it has been granted "recently".
14:45:03 *** mucht_work has joined #openttd
14:45:48 <dihedral> Belugas: crazy world... i prefer keeping my 'hasches' on my keyboard :-P
14:48:03 <TrueBrain> Celestar: and a tnx to you, I finally came around finding a greylistd for exim, and installed it on openttd.org :)
14:50:21 *** Gonozal_VIII has joined #openttd
14:51:05 <TrueBrain> hmm, it only doesn't work :p Lol
14:51:08 <TrueBrain> it keeps on greylisting
14:51:53 <Belugas> listing of extra terrestrials on earth ;)
14:56:02 *** Guest27 is now known as Jello
14:59:50 *** LeviathNL has joined #openttd
15:02:57 <Belugas> hashes to action, funk to funky, we know Major Tom's a junky...
15:03:30 <TrueBrain> yippie, openttd.org has working greylisting; now that is useful :)
15:06:26 *** DaleStan_ has joined #openttd
15:06:33 *** stillunknown has joined #openttd
15:19:30 *** Digitalfox has joined #openttd
15:20:37 *** Digitalfox has joined #openttd
15:29:14 *** orudge` has joined #openttd
15:29:37 *** orudge is now known as Guest85
15:29:37 *** orudge` is now known as orudge
15:29:46 *** ChanServ sets mode: +o orudge
15:32:30 *** egladil has joined #openttd
15:38:28 <TrueBrain> greylistd, the python version debian uses, really is very nice
15:39:51 <TrueBrain> Celestar: talking about cool, have you seen my webtt already? :p
15:52:15 <TrueBrain> you ever sent the email about profiling? :)
15:52:57 <SmatZ> I haven't found any way how to do it
15:53:10 <SmatZ> I didn't receive any email I could reply to...
15:53:22 <TrueBrain> so you signed on after the conversation? :)
15:53:23 <SmatZ> maybe I have bad settings somewhere
15:53:59 <SmatZ> I received just Patch Digest, Vol 1, Issue 1
15:54:11 <TrueBrain> ah, you have digest on :)
15:54:16 <TrueBrain> so you receive a summary every day
15:54:32 <TrueBrain> that is what digest is :)
15:54:54 <TrueBrain> so replying to that is hard
15:55:49 <SmatZ> yes, I will turn off Digest
16:00:59 <LeviathNL> Ammler, I noticed there was a little typo in the tempsnow.grf description. I fixed it you may want to replace the one in the coop-pack with this new one. Only problem is it's ID changed
16:20:03 * SpComb requests for his digest to be sent in orbit
16:29:04 *** Vikthor has joined #openttd
16:29:42 <valhallasw> TrueBrain: the python version debian uses? is that not like... version 2.2 or something? :P
16:30:19 * valhallasw still things CS should be taught in python -_-
16:30:42 <Eddi|zuHause2> CS should not be taught in any specific language at all
16:31:13 <hylje> from my experience python makes one fight the language less
16:31:19 <hylje> so i think it's justified
16:31:19 <Eddi|zuHause2> and i am not a big fan of dynamic python
16:32:18 <Eddi|zuHause2> of course, python has a lot of good things
16:32:33 <Eddi|zuHause2> like the native lists and dictionaries
16:32:35 <Belugas> as any languages/devving environnement
16:33:24 <Eddi|zuHause2> but really, CS is not there to teach people to be <specific language>-masters
16:34:02 <Eddi|zuHause2> because whatever language is "hip" when you start studying, will be totally out of date by the time you finish and want to enter a real company
16:34:26 <Eddi|zuHause2> what CS should teach is to easily adapt to the next "hip" language
16:35:55 <Eddi|zuHause2> btw. especially for newbie programmers, it should be more appropriate to have a really "strict" language
16:36:32 <Eddi|zuHause2> the more dirty tricks they learn at the beginning, the more they cry when the next language does not support that trick
16:37:30 <hylje> python is very strict on a level
16:37:55 <hylje> when one does stuff that's deemed stupid by the greater community (or just a seasoned user), ridicule ensues
16:38:14 <valhallasw> Eddi|zuHause2: python is dynamic, but strongly typed
16:38:29 <valhallasw> but warnings when re-using a variable would be nice :)
16:38:41 <valhallasw> because re-using a variable - except in a loop, a counter, etc - in general is bad
16:39:07 <hylje> would changing a string institute reusing a var?
16:40:01 <valhallasw> += etc defenitly not
16:40:03 <Eddi|zuHause2> valhallasw: such a warning probably requires static types
16:40:07 <valhallasw> but what about text = text.replace(...)
16:40:20 <valhallasw> Eddi|zuHause2: it's possible to warn about type changing while interpreting
16:40:43 <hylje> i looked at some OP code the other day
16:40:48 <hylje> it was quite alike to python
16:41:14 <Eddi|zuHause2> anyway, especially from a compiler building view, i do not like dynamic typing
16:41:34 <valhallasw> or rather... long live swig :)
16:41:47 <valhallasw> Eddi|zuHause2: from a compiler building view dynamic typing is evil, true.
16:41:50 <hylje> there are folks who intend to work all of python in a compiled/JIT way
16:42:05 <valhallasw> but from an interpreter view it is not important, really
16:42:18 <valhallasw> or rather less important
16:42:43 <Eddi|zuHause2> i have a project where they want to check order of function calls against a protocol
16:43:13 <Eddi|zuHause2> and that requires to get to know which functions may actually be called at a certain program point
16:43:29 <Eddi|zuHause2> and that is really tricky, when you can do stuff like:
16:43:57 <Eddi|zuHause2> valhallasw: statically decidable
16:43:57 <Belugas> Pascal = comfy programming. Lots of stuff been handled under the hood
16:43:59 <hylje> dynamic stuff is harder to optimize, yes
16:44:02 <valhallasw> just let the function check whether it is allowed to be called
16:44:21 <Eddi|zuHause2> again, statically
16:44:22 <valhallasw> Eddi|zuHause2: as in: hard to determine for a compiler
16:45:03 <valhallasw> but still it should be possible to test against some state machine?
16:45:15 <Eddi|zuHause2> meaning, you compile the code without running it, and the result is either "good" or "here is a way to break the protocol: a(), b(), x()"
16:45:27 <valhallasw> yeah, okay, then I get the problem
16:45:36 <hylje> the programmer is expected to have a clue
16:45:47 *** egladil has joined #openttd
16:46:35 <Eddi|zuHause2> the actual check is somebody elses job
16:47:52 <Eddi|zuHause2> my job is to convert python and c++ code into a "process rewrite system" of the structure "in state A, there can be executed function b, which spawns a new thread in state B and continues in state C"
16:48:14 <Eddi|zuHause2> (this would look like: "A-b->B||C"
16:48:54 <Eddi|zuHause2> a conventional function call would look like "A-b->B.C"
16:50:09 <Eddi|zuHause2> the tricky part about that, in the above example, you have to get to statically decide that when you call y(), you actually call x()
16:51:45 <Eddi|zuHause2> note, that the assignment y=x and the call y() may be in completely different parts of the program
16:52:12 <Eddi|zuHause2> like in one function you do:
16:52:26 <Eddi|zuHause2> and in f you do
16:53:25 <SmatZ> mmmm function pointers :-)
16:53:48 <Eddi|zuHause2> honestly, pointers are way worse ;)
16:54:53 <SmatZ> maybe ... I am happy with what I know :-p
16:55:22 <Eddi|zuHause2> in the above example you could at least say: "well, i know here is an assignment to y, so for all calls y(), i nondeterministically call either x or y"
16:55:40 <Eddi|zuHause2> try to do that with (x+y)*()
16:56:05 <Eddi|zuHause2> (note, i have no clue about C syntax)
16:56:08 <SmatZ> what would that be good for?
16:57:34 *** DaleStan_ is now known as DaleStan
16:57:38 <Eddi|zuHause2> it's not about "what is this good for", it's about "the language allows this"
16:58:55 <SmatZ> what about function that returns pointer to itself?
16:59:27 <varun> can anybody tell me why my dedicated server aint accessible via the internet ?
16:59:44 <Eddi|zuHause2> in context of a alias-analysis, pointers are probably the worst case
16:59:49 <varun> is there some kinda trick or something (besides opening the port thingie )
17:00:44 <SmatZ> varun: I don't know... often somebody says he has this prrrrrrrrrrroblem, but I don't know the solution... maybe you have blocked some ports you shouldn't have
17:01:08 <glx> varun: server advertised?
17:01:52 <glx> port open and forwarded for UDP and TCP?
17:02:43 <varun> (sorry for me not being to precise )
17:02:57 <varun> (what does error 22 or error 101 mean ? ?)
17:10:02 <varun> SmatZ: if someone would make error messages which ppl could understand i wouldnt be her askin you these stupid questions ....
17:10:41 <glx> these errors are OS specific errors I think
17:12:39 <varun> they only appear when the server is running and someone tries to connect to it
17:14:31 *** Greyscale has joined #openttd
17:16:23 <glx> what are the exact messages?
17:18:11 <varun> just rebooted the machine
17:20:33 <varun> well it keeps tellin me that somebody queried it ?oO
17:22:29 <varun> (search for varun in that page ;) )
17:25:09 <glx> password needed but I have it in the list
17:38:50 *** |Bastiaan| has joined #openttd
17:39:48 *** |Jeroen| has joined #openttd
17:52:42 *** dihedral_ has joined #openttd
17:54:20 *** dihedral is now known as Guest98
17:54:20 *** dihedral_ is now known as dihedral
18:15:34 <SmatZ> TrueBrain: I have finished the profiling, now I would like to post it to the maillist... could you please post something I could reply to? I hope I will receive that email
18:20:19 *** LeviathNL has joined #openttd
18:24:16 *** |Jeroen| has joined #openttd
18:25:08 *** AntB is now known as Guest103
18:29:21 *** |Bastiaan| has joined #openttd
18:29:40 <Ammler> SmatZ: maillist? why not in the forums? is that a secret?
18:30:16 <glx> Ammler: patch tracker is now a maillist
18:46:54 *** Osai is now known as Osai^Kendo
19:00:48 *** Ben_1 is now known as _Ben_
19:07:53 *** mikk36|w has joined #openttd
19:12:24 *** henkiepenkie has joined #openttd
19:13:32 <henkiepenkie> Hi all. Any developpers/coders online?
19:14:41 <SmatZ> henkiepenkie: don't ask to ask, just ask
19:14:49 <SmatZ> somebody will reply sooner or later
19:16:54 *** henkiepenkie has left #openttd
19:21:47 <Belugas> a strange fellow, for sure
19:22:38 *** Wezz6400 has joined #openttd
19:39:45 <dihedral> amusing - that's what it is :-P
19:53:11 <TrueBrain> SmatZ: random seeds are stored in savegame, so that is fine
19:53:45 <TrueBrain> (oh, and it didn't make the thread, but that is no suprise to me :)
19:53:50 <TrueBrain> don't worry about it :)
20:17:36 *** LeviathNL has joined #openttd
20:25:58 *** skidd13 has joined #openttd
20:27:03 <skidd13> SmatZ: Thanks for profiling. I'm currently at a 'small' perl script wich compares all the functions of two profiling sessions :)
20:29:43 <skidd13> SmatZ: Any values or ratios that might be usefull to calculate?
20:30:33 <SmatZ> skidd13: I don't know... I just look into the profile and compare the time some function took
20:30:48 <SmatZ> compare version without and with a patch
20:31:08 <skidd13> The script will compare every function :)
20:31:23 <SmatZ> I just compared your patch against trunk, and it seem to be the same speed with maximum optimisation, and it is faster with default optimizations
20:31:51 <SmatZ> skidd13: to compare the time took xxx ticks of game, I use 'time'...
20:32:03 <SmatZ> the data are not perfect
20:32:10 <SmatZ> it may differ among different runs
20:32:33 <SmatZ> so ... if you compare every function, it will print little differences where there should be none
20:33:08 <skidd13> SmatZ: Maybe a little filter does the job therefore ;)
20:34:09 <skidd13> I'll release the sript under CC BY NC if anyone likes to use it.
20:35:06 <SmatZ> a nice name for a licence :)
20:40:59 *** Cool-Matte is now known as Matte
20:58:29 <skidd13> Anyone can help explain me this: Unmatched ( in regex; marked by <-- HERE in m/TileLoopClearHelper( <-- HERE ()/ at comparer.pl line 168.
20:59:03 <Rubidium> ( starts a grouping of some kind
20:59:13 <Rubidium> you probably have to escape it
20:59:32 <skidd13> I want to search for the ( and did it $line =~ s/\w{2,}\(//;
21:00:33 <skidd13> Rubidium: Is there a special escape for ( ?
21:08:13 <Rubidium> no idea; haven't coded a single line of perl in my life.
21:08:25 <Rubidium> all my regexp experience is in sed and awk ;)
21:11:26 <skidd13> Hmm. I recoded it different. Works so.
21:24:47 *** Guest115 has joined #openttd
21:24:54 *** Guest115 is now known as chu_
21:26:46 <chu_> is there a known problem with the pathfinder of the current svn-head?
21:27:03 <chu_> all my busses went crazy today
21:27:44 <chu_> it looks as if the drive a depth-first-search
21:28:04 <chu_> as if the pathfinder was interrupted in it's work to find the shortest way
21:28:06 <Rubidium> skidd13? Did you do that?
21:28:07 <skidd13> chu_: There has been a problem recently but it has been fixed.
21:28:24 <chu_> skidd13: already in svn?
21:28:33 <Rubidium> depends on what "current" is
21:28:36 *** dihedral has joined #openttd
21:28:43 *** Rexxars has joined #openttd
21:28:51 <chu_> skidd13: ok. i'll try it
21:28:58 <Rubidium> still, depends on what chu_ means with "current svn-head"
21:29:19 <Rubidium> then nothing has been fixed
21:29:38 <chu_> (funny- that's my user-number with my isp too :-)
21:29:39 <Gonozal_VIII> why is there no new nightly?
21:29:39 <Rubidium> or at least, the fix skidd13 was talking about didn't fix it
21:29:53 *** Digitalfox has joined #openttd
21:29:53 <Rubidium> Gonozal_VIII: want to guess?
21:30:19 *** Digitalfox has left #openttd
21:31:30 <chu_> Rubidium: oh. it seems, it did fix the problem.
21:31:48 <Rubidium> then you weren't running "current svn-head"
21:31:49 <chu_> alt least the buses go on tour again *G*
21:32:07 <chu_> Rubidium: yes. you are right. i forgot to "make" the game
21:32:21 <chu_> but the source was "svn up"ed
21:33:19 <chu_> ok. thanks for all your work. and
21:35:13 <Gonozal_VIII> [22:29:53] Rubidium: Gonozal_VIII: want to guess? <-- nothing changed? but there are always changes...
21:35:56 <Rubidium> not the last 27:35 hours
21:42:20 <TrueBrain> [22:29] <Rubidium> then nothing has been fixed <- I triple checked it, so I was sure it was ;)
21:42:31 <TrueBrain> (I wouldn't dare to leave such a bug too long in trunk ;))
21:43:34 <Rubidium> TrueBrain: under the assumption that he was running r11386 and trunk was at r11386 at that moment there would not be a fix in trunk that would solve his problem
21:44:17 <TrueBrain> Rubidium: your assumption is logic, no doubt about it, but I was sure he did something wrong :)
21:44:20 <|fjb|> He was running r11384. :-)
21:49:11 <skidd13> is ther something like ceil or round in perl?
21:51:53 <Rubidium> otherwise it couldn't be that popular ;)
21:56:47 <skidd13> Hmm you have to cast it to an in. Bah. Or include it from the system but only on posix.
21:58:22 <Rubidium> then it's a stupid API
21:59:57 <fjb> skidd13: Math::BigFloat has the fuction fround. Maybe that helps you.
22:00:16 <skidd13> fjb: use POSIX qw(ceil floor); worked for me
22:04:04 <skidd13> SmatZ: Do you like to try my script. The first working version is ready
22:04:45 <skidd13> Where can I upload the script to compare two profiling outputs?
22:04:57 <fjb> Hm, what Do I do if there is no room left for planting trees? :-(
22:05:13 <skidd13> fjb: remove all and start again
22:06:29 * fjb doupts that that is a proper solution...
22:13:13 <SmatZ> skidd13: I dont know where to upload, don't you have some web server?
22:13:59 <skidd13> TrueBrain: I hope the result is enough profiling for you ;)
22:14:29 <skidd13> SmatZ: Nope. Now it comes via maillist.
22:15:45 <skidd13> There's some room for optimisation, but it works... ;)
22:16:25 <fjb> Does the AI never do upgrades on their vehicles?
22:20:42 <skidd13> SmatZ: Did you recieved the mail, cause I don't see it on the patch-archieve
22:21:33 <TrueBrain> skidd13: if you can post your new patch ;)
22:21:36 <TrueBrain> (with the -- stuff ;))
22:21:52 <skidd13> TrueBrain: Satisfied?
22:23:40 <SmatZ> StartStopIndustryTileAnimation(): 3438
22:24:07 <SmatZ> TrainCheckIfLineEnds(): 200
22:24:45 <skidd13> Did you got other experiences?
22:26:17 <SmatZ> I wonder if it is caused by inexact measurements
22:26:23 <SmatZ> or if it really is 3 times faster
22:27:27 <skidd13> Might both be, but you should get an impression if it's faster.
22:28:18 <SmatZ> it is hard to notice, the difference was ~2% when compiled with default settings
22:29:25 <skidd13> SmatZ: The problem is that even if it is that fast. There are also some functions that have a loss of performance.
22:31:34 <SmatZ> it is hard to decida how things changed, when you have so few data
22:31:53 <SmatZ> eg. if it chages from 1,5 seconds to 1,2, it is good
22:32:07 <SmatZ> but when it changes from 0.09 to 0.03... it is hard to say
22:33:24 <skidd13> SmatZ: If you like you can adapt my perl script that way that only the stuff, of that you think wich is important, is noticed
22:33:35 <SmatZ> skidd13: the generated ASM code is almost the same in both cases...
22:33:51 <SmatZ> it cannot cause 3x faster code at all
22:35:23 <skidd13> SmatZ: I removed a few GB in FindFirstBit2x64 might be that this causes a bit speedup.
22:36:23 <SmatZ> skidd13: but is seems to be better
22:36:38 <SmatZ> GB(~t, 4, 4) was compiled as
22:37:32 <skidd13> SmatZ: I've to say that I never learned assembler :(
22:38:03 <SmatZ> just - originally, it did shift and did one useless mask
22:38:08 <SmatZ> now it does just the shift
22:38:43 <SmatZ> but it may be just a luck :)
22:38:54 <SmatZ> compilers are very complicated piece of software
22:39:45 <Eddi|zuHause2> so you are now skid13? :p
22:40:14 <SmatZ> return (x >> s) & ((1U << n) - 1);
22:40:20 <SmatZ> isn't it a bit dangerous?
22:40:28 <SmatZ> when working with 64bit values
22:40:52 <SmatZ> eg., if n >= 32, how will it behave?
22:41:12 <SmatZ> but ok, compiler will give a warning about it
22:41:23 <Eddi|zuHause2> n should never be bigger than the bit size
22:41:40 <Eddi|zuHause2> above that, behaviour is undefined
22:41:53 <SmatZ> x can be 64bit, but 1U is usually 32bit
22:42:13 <skidd13> The problem with the stuff is that templates could be used on custom incomplete types and therefore sizeof() is not realy helpfull
22:42:26 <SmatZ> I am not sure how C defines this situation - isn't the shift value masked with (bit_width-1)?
22:42:39 *** Unknown_Entity has joined #openttd
22:42:42 <Eddi|zuHause2> c defines that it is undefined
22:42:55 <Eddi|zuHause2> the processor does whatever it does
22:43:55 <skidd13> SmatZ: A cast migth secure that stuff
22:44:08 <skidd13> return (x >> s) & (T)((1U << n) - 1);
22:44:46 <SmatZ> maybe (x >> s) & (((T)1) << n) - 1);
22:45:05 <SmatZ> when the '1' is shifted out too much, then the mask will be 0xfffff...
22:50:57 <skidd13> TrueBrain: Do I need to update the patch with the stuff SmatZ mentioned or ..
22:52:11 <SmatZ> skidd13: I changed 1U to ((T)1)
22:52:18 <SmatZ> and it seems to compile exactly same
22:52:59 <Eddi|zuHause2> all templates should be resolved by the time you get to code generation
22:53:33 <SmatZ> well.. I heard something about using typename
22:53:41 <SmatZ> eg. situations where typename is required
22:53:42 *** Osai^Kendo is now known as Osai
22:53:50 <SmatZ> that it is very hard to optimise
22:54:13 <SmatZ> generated ASM is combined with inlined C code...
22:54:14 <Eddi|zuHause2> that was probably about polymorphy
22:54:21 <SmatZ> I don't know, maybe things have changed
22:54:28 <SmatZ> I don't know, never seen that actually
22:54:40 <Eddi|zuHause2> polymorphy is resolved at runtime
22:54:49 <Eddi|zuHause2> templates are resolved at compile time
22:55:13 *** Unknown_Entity has quit IRC
22:56:11 <SmatZ> I would have to imagine the situation
22:56:24 <SmatZ> but I am not skilled in C++...
22:57:11 <Eddi|zuHause2> it's like this in all other object oriented languages
22:57:58 *** Brianetta has joined #openttd
22:58:12 <Eddi|zuHause2> if you have a class A with function f(), and a function B derived from A, which overrides f() with its own version
22:58:24 <Eddi|zuHause2> you can assign objects of B to variables of type A
22:58:40 <SmatZ> do you mean virtual functions etc.?
22:59:32 <Eddi|zuHause2> whereas templates are an improved version of the preprocessor
23:00:07 <Eddi|zuHause2> every template type is replaced by a real type before compilation
23:00:27 <Eddi|zuHause2> the code is copied for each different type that template is instantiated to
23:00:40 <SmatZ> I am just talking what I heard, I don't know "anything" about it, so...
23:00:50 <SmatZ> this is what I would think, too
23:06:00 <SmatZ> I get this warning, when run with -Weffc++:
23:06:11 <SmatZ> /mnt/svn/openttd-rev/src/ai/../helpers.hpp:211: warning: postfix 'OverflowSafeInt<T, T_MAX, T_MIN>& OverflowSafeInt<T, T_MAX, T_MIN>::operator++(int) [with T = long long int, T T_MAX = 9223372036854775807ll, T T_MIN = -0x08000000000000000ll]' should return 'OverflowSafeInt<long long int, 9223372036854775807ll, -0x08000000000000000ll>'
23:06:16 <SmatZ> what do you think about it?
23:06:50 <Eddi|zuHause2> what's that supposed to mean?
23:06:59 <SmatZ> it returns the increased value, doesn't it?
23:07:17 *** Brianetta has joined #openttd
23:07:26 <SmatZ> postfix ++ returns reference to this
23:07:42 <SmatZ> but it should return OverflowSafeInt(*this)
23:08:06 <Eddi|zuHause2> yeah, it looks like a type issue
23:08:07 <SmatZ> FORCEINLINE OverflowSafeInt& operator ++ (int) { return *this += 1; }
23:08:39 <SmatZ> it returns the new value
23:08:45 <SmatZ> but it should return old value
23:08:56 <SmatZ> Money++ is not used anywhere in the code
23:11:23 <SmatZ> I will post a bugreport...
23:15:18 <Roujin> a tool is more intelligent than me
23:15:47 <Roujin> buildottd manages to build the latest revision and i don't >_<
23:16:25 <Roujin> i get some strange issues with heightmap.o ...
23:17:36 <Eddi|zuHause2> ./configure --with-png?
23:18:01 <Eddi|zuHause2> you should probably report that
23:18:37 <Roujin> what's that with configure...?
23:19:12 <Roujin> i'm not really experienced with all that...
23:20:08 <Eddi|zuHause2> well, you need pnglib installed to open png files (e.g. heightmap images)
23:20:47 <Eddi|zuHause2> configure should figure out if you have it installed or not
23:21:01 <Eddi|zuHause2> and if you configured without png support, you should not get errors
23:21:25 <Roujin> i don't run linux... so i guess i don't have this configure
23:21:30 <Eddi|zuHause2> so unless you manually altered your configuration, there should be no problems
23:21:43 <CIA-1> OpenTTD: truelight * r11387 /trunk/src/macros.h: -Codechange: changed many macros.h #defines to static inline functions (patch by skidd13, with a big tnx to SmatZ for profiling!)
23:21:54 <Eddi|zuHause2> mingw also uses configure
23:22:55 <Roujin> okay.. i'm using mingw.. so how do i do this configure then?
23:23:19 <Eddi|zuHause2> at the same place where you run make
23:27:23 <SmatZ> wow big tnx to me :-))))
23:28:03 <Eddi|zuHause2> better have big THX ;)
23:29:02 <TrueBrain> SmatZ: FS is for bugs
23:29:12 <TrueBrain> in case you misesd all the topics
23:29:38 <SmatZ> I though it the maillist is still in testing phase
23:29:47 <TrueBrain> and how else to test? :)
23:31:33 <TrueBrain> hmm, bug in Pipermail I created..
23:31:42 <TrueBrain> will fix it some other day :) At least it shows patches now ;)
23:34:38 <SmatZ> "Patches no longer should go to this bugtracker, but to patch at openttd.org. See http://www.openttd.org/dev.php for more info." is written with very small font... but I admire, I wouldn't read it even if it were written in a big red font...
23:35:04 <TrueBrain> SmatZ: I hate that red font ... who reads that? :)
23:35:13 <TrueBrain> so I thought a small text would get more attention ;)
23:35:37 <TrueBrain> too bad it failed ;) Hehe :)
23:35:42 <Eddi|zuHause2> it's like naming a file "donotreadme" :p
23:35:53 <SmatZ> it failed just for me... so far :)
23:36:03 <SmatZ> "Do not push this button"
23:36:12 <TrueBrain> hmm, Pipermail is weird... your patch now is inline
23:37:58 <SmatZ> but it was sent as attachment
23:38:02 <TrueBrain> Pipermail should make it not-inline
23:38:29 <Eddi|zuHause2> post them to the bugtracker ;p
23:39:31 *** Osai is now known as Osai^zZz
23:39:56 <SmatZ> what about making all base classes' destructors virtual?
23:40:38 <Eddi|zuHause2> are there non-virtual destructors?
23:40:54 * Sacro considers trying openttd on his solaris machin
23:43:06 <SmatZ> OldMemoryPoolBase is base class for OldMemoryPool
23:43:14 <SmatZ> but doesn't have defined destructor
23:44:30 <Eddi|zuHause2> maybe it's an abstract class?
23:45:25 <SmatZ> I am reporting this because
23:45:30 <Eddi|zuHause2> there's no need for destructors if you can't create objects of this class
23:45:31 <SmatZ> compiler gives me a warning
23:46:02 <SmatZ> /mnt/svn/openttd/trunk/src/oldpool.h:100: warning: base class 'struct OldMemoryPoolBase' has a non-virtual destructor
23:46:05 <Eddi|zuHause2> well, _somebody_ was supposed to provide a reimplementation of the pools anyway :p
23:46:34 <Eddi|zuHause2> task pending since like 2 years or so :p
23:46:48 <Eddi|zuHause2> i never get this right...
23:47:51 <SmatZ> I really can't do that, I have different tasks assigned
23:55:31 *** Brianetta has joined #openttd
23:59:08 *** Vikthor has joined #openttd
continue to next day ⏵
