IRC logs for #openttd on OFTC at 2011-10-22
⏴ go to previous day
04:57:26 *** Eddi|zuHause2 has joined #openttd
06:28:35 <dihedral> i was up before you .. ;-)
06:32:35 <planetmaker> so are you proud or should I pity you (on a weekend)? :-P
06:43:11 <dihedral> and have since been tweeking my mail server :-P
06:53:10 <planetmaker> Eddi|zuHause2: I'm afraid your complaints wrt the Makefile are not reproducible for me.
06:53:34 <planetmaker> Please be more precise on how I could possibly reproduce the issue you experience
07:04:42 *** DayDreamer has joined #openttd
07:11:42 *** sla_ro|master has joined #openttd
07:22:33 <FFLaguna> Can I disable "you lose too much money, so we are offering your company to another player"?
07:29:40 *** andythenorth has joined #openttd
07:41:27 <z-MaTRiX> what are the benefits of a normal sized sd card if microsd+normalsd adapter costs same?
07:42:08 <z-MaTRiX> smaller in size, and same capacity
07:44:47 <dihedral> FFLaguna, by playing multiplayer ;-)
08:01:45 *** Alberth has joined #openttd
08:01:46 *** ChanServ sets mode: +o Alberth
08:09:26 *** Eddi|zuHause2 is now known as Eddi|zuHause
08:13:39 *** erik1984 has joined #openttd
08:24:08 *** Devroush has joined #openttd
08:28:55 *** Progman has joined #openttd
08:29:26 *** TinoDidriksen has joined #openttd
08:52:05 *** Progman has joined #openttd
08:57:17 *** Kurimus has joined #openttd
09:00:39 <z-MaTRiX> some zoom optics there
09:08:05 <z-MaTRiX> and everybody will think i'm going to take photos of stars
09:08:15 <planetmaker> no, why would they?
09:08:40 <andythenorth> Eddi|zuHause: he means the re4/4 no?
09:08:59 <Eddi|zuHause> andythenorth: no, those are listed elswhere
09:09:25 <Eddi|zuHause> RFe 4/4 sounds like a railcar
09:10:30 <planetmaker> z-MaTRiX: everybody would rather think you're going to make sports photography or wildlife shots
09:11:05 <planetmaker> it's not like the usual 200mm or maybe 300mm are a big zoom
09:11:22 <planetmaker> at least to me it feels quite limited to have only 300mm at hand :-)
09:12:05 <z-MaTRiX> yes well it might use a stand
09:12:40 <z-MaTRiX> take gun too next time
09:13:23 <planetmaker> nah, one can still carry this lenses. Not light with a proper camera, maybe 2.5kg together. But still
09:13:28 <planetmaker> feasible and worth it
09:13:33 <planetmaker> but indeed I need a monopod
09:14:04 <planetmaker> A tripod is what I have and use ;_)
09:14:11 <z-MaTRiX> and a comfortable seat
09:14:23 <planetmaker> grass and a mat is fine
09:15:24 <dihedral> ... i have a monopod ... *wink*
09:16:33 <z-MaTRiX> what are the benefits of a normal sized sd card if microsd+normalsd adapter costs same?
09:16:46 <z-MaTRiX> smaller in size, same capacity...
09:16:52 *** valhallasw has joined #openttd
09:17:27 <planetmaker> having one piece less which can fail?
09:17:28 <dihedral> i could not tell you any
09:18:28 <z-MaTRiX> not sure i'd want normal size sd if micro size is available
09:18:41 <z-MaTRiX> there is 16GB microsd too now
09:18:52 <planetmaker> my camera uses CF anyway ;-)
09:19:12 <planetmaker> but I possibly should get a bigger CF at one stage...
09:19:15 <z-MaTRiX> i see sd/sdhc everywhere
09:19:51 <planetmaker> my camera isn't terribly new. But I'm still content with 8MPix at 5fps and 12 bit resolution
09:20:19 <planetmaker> and good snr at 640 iso
09:21:50 <planetmaker> and I wished I could afford that camera ;-)
09:22:12 <z-MaTRiX> somehow i'd never prefer a canon
09:22:24 <planetmaker> somehow I always preferred them ;-)
09:22:58 <z-MaTRiX> because it 'looks alike' a canon?
09:23:07 <planetmaker> handling convinced me.
09:23:21 <planetmaker> and when I bought my first one it was the best buy for the buck as well
09:23:53 <planetmaker> and with some lenses around you don't switch camera manufacturer easily.
09:24:05 <planetmaker> especially when lenses keep their value ;-)
09:24:26 <planetmaker> and it's not like their cameras are bad or over-prices now as well. On the contrary
09:24:31 <planetmaker> especially looking at 7D
09:25:47 <planetmaker> one decision point was also the amount of lenses available in principle
09:25:57 <planetmaker> which makes for a reasonable 2nd hand market
09:28:35 <planetmaker> dihedral: jpg is for beginners
09:28:40 <planetmaker> per colour channel
09:29:10 <dihedral> who wants jpg when using a decent camera
09:29:13 <planetmaker> or was it 10bit? I might deceive myself for my camera
09:29:49 *** Biolunar has joined #openttd
09:29:51 <dihedral> i laughed my head of when a friend of mine was talking about having to 'edit the pictures' and was storing them in jpg format
09:31:09 <dihedral> i would have expected cannon raw to be more than 12 bit
09:32:01 <planetmaker> so I have it already for 5 years, I guess
09:32:40 <planetmaker> Kinda ordered it when it came out. And it wasn't my first camera either. Back then Samsung was totally out of question
09:33:00 <planetmaker> only Nikkon was the competition, less so Minolta
09:33:12 <dihedral> according to one source it's 12 bit
09:33:56 <planetmaker> then I remembered correctly
09:36:22 <dihedral> another source mentions 16 :-P
09:36:37 <planetmaker> 16bit is exagerated
09:41:35 <dihedral> you were just talking about the 1d :-P
09:42:01 <dihedral> what's not exaggerated about that? :-D
09:42:35 <planetmaker> 1D... yes, that might
09:42:45 <planetmaker> that's a sturdy and great beast :-)
09:44:38 <planetmaker> z-MaTRiX: and why do you choose the samsung?
09:46:57 <z-MaTRiX> that a mirrorless DSLR
09:47:20 <z-MaTRiX> and i think its better than the competitors
09:47:22 <planetmaker> I saw that. But... where's the advantage in that?
09:47:34 <z-MaTRiX> well it doesnt have the mirror
09:47:55 <planetmaker> 'no mirror' is not an advantage in itself
09:48:29 <planetmaker> how's the view finder handled with mirrorless cameras?
09:48:33 <z-MaTRiX> if you look at that part in a DSLR, focus is calculated
09:48:43 <z-MaTRiX> well its handled in digital, - if has one
09:48:53 <z-MaTRiX> there are some that does not have viewfinder
09:49:08 <z-MaTRiX> but live-preview is there
09:49:24 <z-MaTRiX> and autofocus can be done directly
09:49:38 <planetmaker> and everyone knows how great displays work or show under some conditions ;-)
09:49:41 <dihedral> that aint the best for the sensor :-)
09:49:43 <z-MaTRiX> you dont have to calibrate your mirror system for focus
09:50:03 <z-MaTRiX> this type i mentioned has oled display
09:50:28 <z-MaTRiX> it is active light matrix
09:51:10 <z-MaTRiX> the panasonic competitor uses lcd for example
09:51:26 <dihedral> it's good compared to normal digital cameras
09:51:43 <z-MaTRiX> since it has dslr sized sensor
09:52:30 <z-MaTRiX> well this is a long description with technical parameters if you interested
09:52:52 <z-MaTRiX> the samsung.com description lols
09:53:25 <z-MaTRiX> "life's easy use it stupid type"
09:54:41 <z-MaTRiX> dihedral<< yeah i saw some guys stating the mirror protects the sensor while not making photos
09:55:21 <z-MaTRiX> dont know yet if it has some other shutter for protection
09:55:36 <dihedral> but also all displays have pixels
09:55:49 <dihedral> so focus is never as good
09:55:51 <z-MaTRiX> but if it does live-preview then is possibly uses the sensor, so pointing it to sun will probably damage the main sensor
09:56:45 *** DorpsGek has joined #openttd
09:56:45 *** ChanServ sets mode: +o DorpsGek
09:56:46 <dihedral> though i am sure it's a neat way to get started ;-)
09:57:21 <dihedral> i'll get you to know some tricks ;-)
09:57:24 <planetmaker> my eye has a better focus in difficult situations than electronics. And then a VGA resolution or (whatever) just doesn't cut it
09:58:07 <planetmaker> but the main reason I'd not start with samsung: available lenses ;-)
09:58:20 <z-MaTRiX> well given the pentaprism is machined sub-micrometer accurate then its better option for you if you are concerned about focus correctness
09:58:22 <dihedral> but, considering the cost, i think you'll get a good portion of fun out of it, while getting into the topic deeper and depper
09:58:30 <planetmaker> most money should be invested in lenses
09:58:36 <dihedral> and soon (1 - 2 years) you'll want a better camera anyway ;-)
09:58:59 <dihedral> the only danger is, if you buy too many lenses for that thing, you'll be stuck with samsung
09:59:23 <planetmaker> yeah. switching systems is basically impossible once you started
09:59:26 <z-MaTRiX> i think i'll probably damage it and throw it out in a few years
10:00:21 <z-MaTRiX> this mirrorless thing is fairly new topic
10:00:52 <z-MaTRiX> what would you say if they make 10 megapixel oled viewfinder for you in 2 years?
10:01:37 <planetmaker> I'd still say "I'll need to test it in nighttime"
10:01:52 <planetmaker> or for delicate macro
10:02:07 <z-MaTRiX> nighttime performance depends on sensor size
10:02:24 <z-MaTRiX> i saw largest ones are 18k$+
10:02:29 <planetmaker> mostly yes. And sensor performance
10:02:56 <planetmaker> canon used to be much better there than others. But others caught up
10:03:19 <z-MaTRiX> btw if you use larger optics you get better light collection efficiency no?
10:03:47 <planetmaker> that's what the f/x.y tells you
10:03:53 <planetmaker> the "opening ratio"
10:04:04 <planetmaker> the smaller x.y the more light you get
10:04:35 <planetmaker> focal length over opening diameter
10:05:49 <dihedral> the more lenses inside the optic the more light you lose
10:06:07 <dihedral> unless you want to spend a bunch of cash, but then you'd not want that camera ;-)
10:06:15 <planetmaker> dihedral: yes... but for normal lenses that doesn't really matter
10:06:31 <z-MaTRiX> well that camera includes an optics
10:06:38 <planetmaker> any somewhat decent commercial lense has sufficient lambda/4 coatings there
10:06:44 <Eddi|zuHause> hm... the "double decker" engine, which realistically does only pull double decker push-pull trains, offer that as fixed unit, or as single engine?
10:07:13 <planetmaker> go for fixed now with refits in length
10:08:24 <z-MaTRiX> so if there is too much light i close the aperture a bit
10:08:41 <z-MaTRiX> my webcam currently is too sensitive
10:09:21 <planetmaker> yes. At the same time you increase the depth of field though
10:09:31 <planetmaker> you might as well reduce sensor sensitivity of shutter time
10:09:45 <planetmaker> depends on what you want to achieve wrt image result
10:09:59 <z-MaTRiX> unfortunatly video recording does not have shutter time
10:10:05 <dihedral> personally i'd go for something like a 60D
10:10:19 <z-MaTRiX> yeah i was looking at nikons but...
10:10:20 <dihedral> or something along those line
10:10:28 <z-MaTRiX> dont want mirrors now
10:10:29 <dihedral> or preferably something with a big sensor :-)
10:10:44 <dihedral> z-MaTRiX, you do, you just don't know it yet
10:10:44 <planetmaker> dihedral: yes... but unfortunately they killed the nice handling of the 30D...50D from the 60D and adopted the xy0D handling
10:10:47 <planetmaker> which I like much less
10:11:07 <z-MaTRiX> and its large, its clicking
10:11:07 <planetmaker> thus... if I'd buy a camera now, it'd be the 7D ;-)
10:11:10 <dihedral> z-MaTRiX, mirror is the best you can do ;-)
10:11:21 <dihedral> unless you want to shot movies
10:11:41 <dihedral> a friend of mine bought one
10:11:49 <planetmaker> and it's only 1.2k€...
10:11:57 <dihedral> eh ... wait ... no - he went to nikon i think
10:12:04 <dihedral> bought a complete new set
10:12:18 <dihedral> z-MaTRiX, because they get money for that
10:12:22 <planetmaker> then the d300 from nikon
10:12:50 <planetmaker> D300 is probably what I'd get, if I used Nikon
10:13:08 <z-MaTRiX> oh look what i found
10:14:05 <planetmaker> well... I need a very tiny one for the trouser pocket ;-)
10:16:50 *** DorpsGek has joined #openttd
10:16:51 *** ChanServ sets mode: +o DorpsGek
10:17:04 <z-MaTRiX> you may wonder what will you photo underwater
10:17:11 <z-MaTRiX> and wat will be the quality
10:18:04 <Alberth> water quality is probably more of a concern :)
10:18:20 <planetmaker> makes it easily usable on canoe tours or so
10:19:06 <planetmaker> last canoe tour I didn't dare take mine
10:19:18 <planetmaker> rain forest is no issue with good but normal equipment
10:19:32 <planetmaker> it's sufficiently weather-sealed
10:19:39 <z-MaTRiX> but your dslr ... ;/
10:19:52 <planetmaker> it's no issue to use and handle it in heavy rain
10:20:25 <planetmaker> only water on the lense may be detrimental... but a lense hood may help there
10:20:56 <planetmaker> also a difference to the cheap entry cameras, though ;-)
10:21:14 <z-MaTRiX> btw how about those leafshadowthings i saw before on an expensive optic end?
10:21:45 <z-MaTRiX> those will provide protection against raindrops too right?
10:21:45 *** Brianetta has joined #openttd
10:21:55 <planetmaker> I guess that's what I meant with lense hood
10:22:51 <planetmaker> they avoid stray light coming from out-of-image area entering the optics and giving nasty bright spots
10:23:13 <dihedral> planetmaker, do you have a 'pol-filter'?
10:23:42 <planetmaker> helps a lot when you try to take photos of fish
10:24:08 <planetmaker> but can be funny sometimes in other places as well. But I don't use it often
10:24:32 <dihedral> i love the colours you get with those filters
10:25:47 <planetmaker> I mostly use it to get rid of reflections I don't want
10:26:20 <z-MaTRiX> horizontal polarization filter underwater?
10:27:54 <dihedral> did you check the example picture z-MaTRiX
10:29:21 <dihedral> find the 1 difference ;-)
10:29:46 <z-MaTRiX> yey well thats rather interesting
10:30:03 <dihedral> and the window is still closed ;-)
10:30:06 <planetmaker> I find it a rather disconcerting example :-P
10:31:15 <z-MaTRiX> yeaah was reading about the quarter-wave plate in cdrom
10:31:46 <planetmaker> shunting use only?
10:31:51 <z-MaTRiX> and it has a polarisation splitter cube too
10:32:40 <Eddi|zuHause> yeah, mainly shunting
10:32:54 <dihedral> pabst-mobil on rails :-P
10:33:04 <planetmaker> interesting engine, indeed :-)
10:33:20 <Eddi|zuHause> can be switched between track class C and D by removing some weights
10:33:29 <Eddi|zuHause> so which one do we take?
10:35:02 <planetmaker> or would we duplicate then much?
10:40:50 <TrueBrain> All OpenTTD Webservices (with the exception of the masterserver) are down. There is some terrible performance from the MySQLd, causing huge issues throughout the network.
10:43:22 <z-MaTRiX> i have a new project now
10:43:45 <z-MaTRiX> i have ~5m cat-5e cable, and a webcam with ~1m cable with audio
10:44:07 <z-MaTRiX> and willing to cross the 2 ;>
10:45:02 <z-MaTRiX> usb + audio over cat-5e
10:45:20 <valhallasw> it's just a different cable :p
10:46:01 <valhallasw> and 5m is probably a bit long for USB
10:54:34 <TrueBrain> for everyone working with MySQL: add skip-name-resolve in your [mysqld] in your config .... #@$#$%#@$@#%#$
10:55:07 <Whity> lol, it made a big difference?
10:55:34 <Alberth> between working and not working big enough? :p
10:55:46 <TrueBrain> between connection accepted in 60 seconds and directly? :P
10:56:37 <Whity> haha, that looks more like a DNS issue to me
10:56:49 <TrueBrain> events of the day, for who is interested: all webservice were terribly slow. An initial scan revealed a user on a VERY slow connection doing MANY requests, and over time he took too many active connections, limiting availability
10:57:09 <TrueBrain> then MySQL decided to hang (looking back, it had DNS issues), it got killed, which corrupted one of our archive table
10:57:22 <TrueBrain> which means it consumes 100% CPU and consumes the whole disk very fast
10:57:42 <TrueBrain> and then it turns out our ISPs DNS stalled connections with rDNS requests, causing MySQL to be VERY slow
10:57:51 <Alberth> murphy at its best :)
10:57:53 <TrueBrain> at least they are coming all at the same time, I guess
10:58:09 <TrueBrain> anyway, all webservices should be available again :)
10:58:18 <Whity> MySQL is being a bitch lately
10:59:16 <Whity> i just solved an issue at work with our SUN Cluster randomly rebooting MySQL
10:59:29 <Whity> because of timeouts in checking permissions
10:59:55 <TrueBrain> we have an archive table of 100 MiB
10:59:59 <TrueBrain> believe it contains 40M records
11:00:07 <TrueBrain> every time MySQL does not 100% what it should, the table crashes
11:00:15 <TrueBrain> I don't know why we still use MySQL tbh
11:00:55 <TrueBrain> you would think, yeah ...
11:01:04 <TrueBrain> 40M records is on the large size, sure, but .. really? :P
11:01:04 <Whity> using the ARCHIVE engine? or plain MyISAM/InnoDB ?
11:01:15 <TrueBrain> I just said that: an archive table :P
11:01:30 <TrueBrain> MyISAM bailed out years ago, and InnoDB didn't like it either :P
11:02:03 <Whity> i am working with 1.5G inno tables :S
11:02:22 <TrueBrain> size of Inno is bigger than ARchive :D
11:02:29 <TrueBrain> this is a table which keeps track how many downloads are done
11:02:31 <TrueBrain> so it is INSERT only
11:02:35 <TrueBrain> much faster as ARCHIVE
11:03:06 <Whity> you could do inserts on a separate DB :p
11:03:20 <Whity> so it doesnt fuck up the query cache :p
11:03:58 <TrueBrain> silly that it is needed tbh
11:04:10 <Whity> your query cache is not 1GB or so is it?
11:04:14 <TrueBrain> but okay .. that a non-clean-shutdown crashes this table really annoys the fuck out of me
11:04:23 <TrueBrain> euh .. query cache is huge
11:04:43 <Whity> that is on the big side for cache
11:06:11 <TrueBrain> I know little of MySQL tuning
11:06:20 <TrueBrain> this performance app tells me what I should icnreas, and I did that till it shut up :P
11:07:34 <TrueBrain> kinda lost where it is .. hmm ...
11:07:52 <Whity> other tables are isam or inno ?
11:09:04 <Whity> as long as its mainly read only that's okay, if your key_buffer_size is big enough
11:09:06 <TrueBrain> Innodb is disabled I believe
11:09:19 <TrueBrain> 256 MiB of KeyBuffer :P
11:09:22 <TrueBrain> no, we run our own shit :P
11:09:26 <TrueBrain> disabled by configuration
11:10:03 <Whity> slow_query_log showing anything out of the ordinary?
11:10:15 <TrueBrain> but what can you do ...
11:10:28 * TrueBrain gives Whity MediaWiki :P
11:10:32 <TrueBrain> also FlySpray has a few
11:10:42 <TrueBrain> can't be arshed to optimize them tbh :)
11:10:55 <TrueBrain> and our stats (which is run every 24h) screams :P
11:11:02 <TrueBrain> that is why we run it at night :P
11:11:18 *** andythenorth has joined #openttd
11:11:19 <TrueBrain> the query is HUGE :P
11:11:32 <TrueBrain> I tihnk I removed the tool I used to tune mysql, I cannot find it :(
11:11:34 <Whity> im on a mysql roll anyway this month
11:11:45 <TrueBrain> ah! in /var/tmp ... lolz
11:12:48 <TrueBrain> haha: Data in ARCHIVE tables: 9G (Tables: 1)
11:12:54 <TrueBrain> I strongly doubt that value is correct :P
11:13:15 <Whity> how did you get that value? :p
11:13:53 <TrueBrain> Ah, the total amount of indexes is 111.8M, and keybuffer is 256M .. should fit easily :D
11:14:02 <Whity> total over the full DB?
11:14:05 <TrueBrain> 97.4% keybuffer hit rate :D:D
11:14:16 <Whity> that means you have near to none indexes lmao
11:14:21 <TrueBrain> 74.2% query cache efficiency :D
11:14:34 <TrueBrain> Whity: every table is indexed, no worries
11:14:36 <Whity> unless your DB is only 120MB offcourse
11:14:48 <TrueBrain> no, 4 times as big or something
11:15:14 <Whity> what is your query cache ratio?
11:15:25 <TrueBrain> [13:14] <TrueBrain> 74.2% query cache efficiency :D
11:15:44 <TrueBrain> Joins performed without indexes: 33 :P
11:16:08 <TrueBrain> 18%, 178 on disk, 973 in total
11:16:41 <Whity> tmp_table_size prob too small
11:17:22 <TrueBrain> not even in the settings .. there was something with that entry in MySQL 5 .. hmm
11:17:28 <TrueBrain> table_cache is 1000
11:17:40 <Whity> and open_tables is currently?
11:17:58 <TrueBrain> much more will be very hard :P
11:18:17 <TrueBrain> we don't have that much tables :P
11:18:19 <Whity> we have 300 tables and 4000 open :P
11:18:30 <TrueBrain> and not many concurrent connections :P
11:19:22 <TrueBrain> I believe something like 30 at most
11:19:28 <TrueBrain> in worst case scenarios (read: slashdot)
11:19:42 <Whity> 30 is already getting bigger lol
11:20:39 <TrueBrain> last time we got slashdotted, we got up to 40 http requests per second or something, well below the max .. I am not that worried :P
11:20:47 <Whity> no reverse proxy in front of the site?
11:20:48 <TrueBrain> believe the peek was 80 ... *yawn*
11:21:06 <TrueBrain> there is, but how does it matter?
11:21:32 <Whity> well that should take care of the HTTP requests that serve non-dynamic content
11:21:49 <TrueBrain> it does; our http stuff is fairly optimized
11:22:21 <TrueBrain> for example, we use php-fpm for PHP pages, nginx for all static content, for the wiki alone
11:22:26 <TrueBrain> flyspray has an identical setup
11:22:35 <TrueBrain> all proxied by one huge nginx in front of that
11:22:46 <TrueBrain> gives a huge performance, we have seen :)
11:23:33 <TrueBrain> MediaWiki has at max 7 php-fpm instances, an average page takes 250ms to generate, so when more than 28 dynamic pages per second are requested, mediawiki starts to complain .. the first one who will :P
11:23:53 <TrueBrain> of course it is heavily cached, so only the first page takes 250ms, every other request after that takes 30ms ..
11:24:11 <TrueBrain> which would mean ... 200+ dynamic pages per second
11:24:30 <Whity> you dont have the pages cached for some seconds?
11:24:52 <TrueBrain> mediawiki caches its pages on several levels
11:25:00 <TrueBrain> depending on what and when, it can cache up to a few hours
11:26:23 <TrueBrain> hmm, in fact, our load balancer will complain first .. takes him 14ms to handle a connection, and it runs in 1 thread (lol)
11:28:15 <TrueBrain> but yeah, I like our system :)
11:28:48 <Whity> it sounds like a good setup, just some minor flaws kicking it in the balls sometimes
11:28:59 <TrueBrain> yeah, like a rDNS which goes out for a piss
11:29:17 <TrueBrain> one thing I need to figure out, is how I can avoid very slow users hogging up all connections ...
11:29:43 <Whity> but how is it setup that a slow user can hog a connection?
11:30:23 <TrueBrain> 1 IP had 10+ connections all with a Send-Q of huge, with active connection attached
11:31:00 <TrueBrain> (that user is banned now btw :p He was crawling all pages, in a very bad and slow way :P)
11:31:31 <Whity> the user was called BaiduSpider? :P
11:31:39 <Whity> i banned that chinese mofo last week too
11:33:08 <Whity> those connections get hogged on the nginx level?
11:33:14 <TrueBrain> on the first proxy, yeah
11:33:23 <TrueBrain> not sure why .. somehow they stayed alive
11:33:46 <Arafangion> Perhaps they trickled data through slowly?
11:33:56 <Whity> i'm not a nginx expert :P
11:34:07 <Arafangion> It's the usual way to keep a connection open.
11:34:14 <TrueBrain> Arafangion: yeah, but even then. Normally data is put in the Send queue, and nginx just closes the connection
11:34:18 <TrueBrain> it is up to the OS to take care of it
11:34:36 <TrueBrain> its normal to have a shitload of connections with a queue and CLOSE_WAIT
11:34:39 <Arafangion> TrueBrain: The OS wouldn't know, as it does have data.
11:34:40 <TrueBrain> those I don't care about
11:34:59 <TrueBrain> Arafangion: when you request a page, nginx serves you the page (it hands it over to the OS) and it closes the connection
11:35:05 <Arafangion> Basically you manipulate the TCP connection.
11:35:07 <TrueBrain> nginx no longer cares about you, and continues to serve others
11:35:20 <Arafangion> Ah, ok - then neat.
11:35:35 <TrueBrain> Whity: yeah, earlier today I had the limiter active, but it turns out my browser alone opens 5+ connections to open the wiki :D
11:35:51 <TrueBrain> and nginx then returns a bad return-code when the limit is reached, causing partial pages :(
11:36:01 <TrueBrain> so that was even worse :P
11:36:27 <TrueBrain> I assumed it would queue the connections, but it doesnt :P
11:37:01 <TrueBrain> reading on the internet, I am not the only one who had issues with JikeSpider :P
11:37:07 *** Cybertinus has joined #openttd
11:37:37 <TrueBrain> Arafangion: if a http (or any other network service btw) wouldn't do this, a DoS would be very very easy ;)
11:39:43 <Whity> and your max connections is slightly over 10 or how come it hogs all connections? ;)
11:40:09 *** JVassie has joined #openttd
11:40:57 <TrueBrain> I have no idea why it hogged them ... seems keepalive is the cause ..
11:41:15 <TrueBrain> oeh, btw, talking about this shit: what is a good dns proxy / cache / resolver for on a gateway?
11:41:28 <Arafangion> TrueBrain: So it was trickling data? :)
11:42:00 <Arafangion> TrueBrain: bind is the traditional choice.
11:43:04 <TrueBrain> I run dnsmasq at home .. dont really like it either
11:43:21 <Whity> you asked for a good one :P
11:44:39 <TrueBrain> bind is not good :P
11:44:42 <TrueBrain> bind is 'common' :P
11:45:15 <Whity> most common things are not 'bad' :P
11:45:17 <Rubidium> maybe the samba project has a dns server
11:45:26 <Rubidium> then you can run it in an AD
11:45:33 *** |Jeroen| has joined #openttd
11:46:07 <TrueBrain> Whity: how many issues have there been with bind? :P
11:46:10 <TrueBrain> how many of them were critical?
11:46:14 <TrueBrain> now take, for example, powerdns
11:46:58 <TrueBrain> DebianAdministration suggests pdnsd
11:49:37 <TrueBrain> <3 simple configuration scripts :D
11:50:59 <TrueBrain> haha, 692 KiB of RAM in use :D
12:06:52 *** valhallasw has joined #openttd
12:10:04 <michi_cc> Seems to be somewhat exotic :)
12:10:26 <TrueBrain> why does OpenTTD always have more servers online than clients to play on them? :P
12:11:48 <Arafangion> TrueBrain: A lot of those games seem to be very old, or saturated with AI bots.
12:11:50 <TrueBrain> Whity: I changed the tmp_table_size to 128MiB (!), and still it makes tables on the disk ... I hate MySQL :P
12:12:28 <Arafangion> TrueBrain: You're not alone there. :) Many people hate MySQL.
12:13:13 <TrueBrain> now made it 256 MiB :P
12:13:16 <Whity> did you also increase the heap size+
12:13:21 <TrueBrain> now MySQL can use 101% memory, shit
12:13:22 <Whity> otherwise it doesnt matter
12:13:28 <TrueBrain> those should always be identical :)
12:13:29 <|Jeroen|> many people hate orcale :-p
12:13:58 <TrueBrain> okay, that is why I didn't change it: even at 256MiB, it still works on disk :P
12:14:10 <Whity> so i guess you are doing joins that create tables over 256MB
12:14:10 <Eddi|zuHause> michi_cc: only things i miss now for SBB is the very old engines, the cisalpinos and the very new engines (511/512). and occasionally some weight values
12:14:21 <Whity> are you sure those joins are needed? :p
12:14:29 <TrueBrain> Whity: which is hard, as I doubt we have that much data to JOIN :P
12:14:47 <Whity> those queries should be in your slow query log
12:15:54 <Whity> well it depends, what is your long query time?
12:16:16 <Eddi|zuHause> next we should maybe think about which of these go to core
12:16:44 <TrueBrain> not in show_status?
12:16:45 <Whity> the time after which queries go to the slow query log
12:16:52 <TrueBrain> owh, configure option :P
12:17:10 <TrueBrain> I guess it is in seconds
12:17:25 <Whity> i mean 1 second is a slow query imho
12:17:43 <Arafangion> Unless it's Oracle.
12:17:49 <Whity> (im in e-commerce; clients talk in ms)
12:18:04 <TrueBrain> Whity: yeah, I agree, 1s is insanely slow tbh :P
12:18:18 <Whity> so 10s is not telling you a lot
12:18:22 <Arafangion> Seriously, I know of an oracle database that takes 3 *minutes* to respond.
12:18:33 <TrueBrain> hahaha, 1 of the queries:
12:18:41 <TrueBrain> select IFNULL(SUM(...) from INFORMATION_SCHEMA! :P
12:18:44 <Arafangion> Whity: Not my responsibility. :)
12:19:12 <Whity> where does that query come from lol
12:19:19 <Whity> never seen that in the wild
12:19:31 <Arafangion> Probably a wierd ORM.
12:19:34 <Whity> and that is taking over a second? :s
12:19:52 <TrueBrain> ah! I also log non-indexed :)
12:22:41 <TrueBrain> turns out our frontpage is without index on the order by field, lolz
12:22:51 <TrueBrain> query time is 0.5ms, I don't care
12:23:06 <Whity> your frontpage should be in the reverse proxy cache anyway
12:23:22 <TrueBrain> Whity: it is in a cache, just not in thatone :P
12:23:41 <Whity> then why is it hitting your DB ? :p
12:24:05 <TrueBrain> k .. 66 tables on disk, no slow queries :P
12:24:47 <Whity> im gonna check a production system of ours
12:24:50 <TrueBrain> I guess I should just accept it and live with it :P
12:25:20 <Whity> probably on disk means in /tmp which is RAM :p
12:25:47 <TrueBrain> nope, /tmp is disk on this VPS
12:26:02 <TrueBrain> I wonder why, but okay
12:26:03 <Whity> then i cannot explain it lol
12:26:52 <TrueBrain> so then we just accept it :P
12:27:50 <Whity> until it gives you issues
12:27:50 <TrueBrain> tnx anyway Whity :)
12:27:59 <Whity> but it doesnt explain why your DB crashes
12:28:09 <TrueBrain> owh, that is just MySQL being MySQL :)
12:28:29 <Whity> ive never seen it crash it self in a production system
12:28:42 <TrueBrain> lol; I can't say the same :D
12:28:47 <TrueBrain> mostly it self-repairs
12:28:53 <TrueBrain> but MySQL 5 can't repair ARCHIVE tables
12:29:16 <Whity> mysql> show status like '%tmp%';
12:29:16 <Whity> +-------------------------+-------+
12:29:16 <Whity> | Variable_name | Value |
12:29:16 <Whity> +-------------------------+-------+
12:29:16 <Whity> | Created_tmp_disk_tables | 0 |
12:29:17 <Whity> | Created_tmp_files | 42 |
12:29:17 <Whity> | Created_tmp_tables | 0 |
12:29:19 <Whity> +-------------------------+-------+
12:29:32 <TrueBrain> use pastebin for so many lines :P
12:29:36 <Whity> this is a heavy e-com DB
12:29:44 <TrueBrain> we are special :D:D
12:30:00 <Eddi|zuHause> "Das pflücken von Blumen während der Fahrt ist verboten" :p
12:30:08 <TrueBrain> Eddi|zuHause: ich spreche keine Deutch
12:30:35 <Eddi|zuHause> TrueBrain: that is entirely your own fault :p
12:31:20 <TrueBrain> but to call it a fault ...
12:32:28 <Whity> but TrueBrain if i were you i would run that ARCHIVE table in a separate instance so the full table scan when you read from it does not overwrite any caches
12:33:04 <TrueBrain> its only read rarely
12:33:20 <TrueBrain> like once in the 4 months or so
12:33:33 <TrueBrain> no, the ARCHIVE table is not, it is a manual action Rubidium has to do :P
12:33:42 <TrueBrain> daily stats go in another table
12:34:13 <TrueBrain> 37M BaNaNaS downloads :D
12:34:39 <TrueBrain> since 2009 somewhere?
12:35:49 <TrueBrain> I had to look it up
12:35:51 <TrueBrain> but no, not really :P
12:36:24 <Whity> so, if the ARCHIVE is not being read... what is causing your crash :/
12:36:37 <TrueBrain> and because the ARCHIVE is written often and a lot
12:36:44 <TrueBrain> chances it corrupts seems very high
12:36:58 <TrueBrain> oeh, OpenTTD is downloaded 3M times! :D
12:37:20 <TrueBrain> owh, no, that is wrong
12:37:24 <TrueBrain> that is total downloads of everything
12:37:30 <Whity> you still outperform minecraft
12:37:30 <TrueBrain> 2.5M times a release is downloaded :P
12:37:32 <Eddi|zuHause> michi_cc: i have not looked at the BLS vehicles yet
12:37:46 <TrueBrain> Whity: just it took us a few years :P
12:37:57 <TrueBrain> stats started at 2008-09-14
12:38:15 <Whity> that is 3 years; still decent
12:38:23 <TrueBrain> 3000 downloads every day
12:38:26 <TrueBrain> that is not bad, no
12:38:50 <Arafangion> And it wouldn't count people like me who started with OpenTTD from the debian repo.
12:42:16 <TrueBrain> that is why BaNaNaS is a slightly better indication
12:42:23 <TrueBrain> 37M ingame downloads is .. a lot :P
12:42:50 <TrueBrain> sadly, we don't keep track of unique connections made, so we can't really count how many people are active on a day :P
12:42:54 <TrueBrain> would be nifty, but we are not Facebook
12:44:13 <amix> just put my 4 computers on facebook
12:44:52 <michi_cc> Hmm, what to do with modified engines? The BLS Ce 4/6 for example was modified to a Ce 4/4 from 1956 on. So should the decomissioned date be that of the original 4/6 (1972) or of the 4/4 (1984 or so)
12:45:24 <TrueBrain> amix: I guess it is a good thing; if you ever might die, we at least have enough stories to tell about you :D
12:49:05 <amix> TrueBrain: the biggest screen is for the morphos machine
12:49:26 <michi_cc> Eddi|zuHause: The BLS Ae 8/8 is an articulated engine with two parts, no idea how to put that into the spreadsheet :)
12:49:52 <Eddi|zuHause> michi_cc: like the SBB 8/14
12:50:22 <amix> TrueBrain: what are you doing?
12:50:23 <Eddi|zuHause> you have the "front" part with an id > 128, and the "back" part with an id <128
12:50:34 <Eddi|zuHause> and 0 middle parts
12:50:39 <TrueBrain> amix: enjoying my saturday; you?
12:50:59 <amix> waiting for MorphOS 3 for powerbooks
12:55:54 <michi_cc> Eddi|zuHause: Have to step out a bit, but will do so unless you get there first.
12:56:25 <amix> the new Amiga is released
12:56:43 <amix> for the cheap price of 1699 ;P
12:57:39 *** sla_ro|master has joined #openttd
13:01:48 *** Devroush has joined #openttd
13:04:34 *** sla_ro|vista has joined #openttd
13:10:08 *** Devroush has joined #openttd
13:30:09 *** andythenorth has joined #openttd
13:40:29 *** frosch123 has joined #openttd
13:43:45 *** HerzogDeXtEr1 has joined #openttd
13:57:13 <z-MaTRiX> can some pro recommend me something?
13:57:44 <z-MaTRiX> my modded webcam takes too much light and picture is oversaturated
13:58:12 <z-MaTRiX> should i put a piece of aluminium foil with a smaller hole in place of the IR filter?
14:06:18 *** |Jeroen| has joined #openttd
14:07:46 <Arafangion> Depends on the mod I guess.
14:07:59 <z-MaTRiX> only removed the IR filter
14:08:14 <z-MaTRiX> so sensor gets more light
14:08:48 <Arafangion> You'd need to tweek the calibration
14:09:04 <Arafangion> Most likely in software.
14:09:15 <z-MaTRiX> but sensor is getting oversaturated
14:10:02 <z-MaTRiX> even if i could get brightness lower i have lost all colors
14:10:06 <Arafangion> The sensor itself? You'd probably have to increase the frame rate.
14:10:14 <Arafangion> But that's probably limited by hardware.
14:10:42 <Arafangion> Perhaps add a filter?
14:11:05 <Arafangion> Back to square one, I guess. :)
14:12:04 <z-MaTRiX> there is a filter that takes a little outof visible portion?
14:12:28 <Arafangion> The white balance may assume a particular distribution, though.
14:19:11 <andythenorth> z-MaTRiX: too much light is falling on the ccd?
14:19:16 <andythenorth> try an ND filter
14:19:35 <andythenorth> sounds unusual though, typically there is too little light for a sensor
14:20:01 <andythenorth> the gain on the ccd might be turned up in software
14:20:21 <Arafangion> andythenorth: He's removed a filter, though.
14:20:29 <Arafangion> andythenorth: So it makes sense for there to now be too much.
14:20:42 <z-MaTRiX> andythenorth<< i have removed the IR filter
14:21:03 <z-MaTRiX> because it can see in the dark this way
14:21:14 <z-MaTRiX> "dark" IR irradiated
14:21:19 <andythenorth> try replacing IR with an ND then
14:21:30 <andythenorth> the IR filter will have a loss factor, probably not much
14:21:38 <andythenorth> the ccd will be calibrated for that
14:21:47 <z-MaTRiX> is an ND filter common thing?
14:21:52 <andythenorth> an ND is 'neutral density'
14:22:02 <z-MaTRiX> will it let through IR ?
14:22:13 <z-MaTRiX> i'd like to block a bit from visible spectrum
14:22:48 <andythenorth> there are filters for that
14:23:18 <andythenorth> possibly the issue can be fixed in software in your case
14:23:21 <Arafangion> You could also use multiple cameras.
14:23:28 <andythenorth> but sometimes it's easier just to stick a filter on
14:24:04 *** pegdelf has joined #openttd
14:24:47 *** pegdelf is now known as amix^
14:25:15 <z-MaTRiX> i put down brightness to zero and contrast, but still white
14:25:43 <andythenorth> more complex software might be able to handle changing the gain
14:25:45 <z-MaTRiX> this might be because sensor is more sensitive to IR
14:26:02 <andythenorth> but ultimately if there's too much light falling on the ccd, it will clip at the peak
14:26:32 <z-MaTRiX> so it cannot be corrected using software
14:27:03 <Arafangion> z-MaTRiX: you could correct it with software using an additional camera (with IR filter)
14:27:31 *** Alberth has joined #openttd
14:27:32 *** ChanServ sets mode: +o Alberth
14:28:00 <andythenorth> z-MaTRiX: got anything you can use as a filter to test?
14:28:17 <z-MaTRiX> sure i have used a welding google, that worked
14:28:20 <andythenorth> glass, acetate, tracing paper...
14:28:26 <andythenorth> cigarette rolling papers
14:28:32 <z-MaTRiX> but i believe it blocks some IR and a bit green
14:28:41 <andythenorth> a welding google is quite dark :)
14:29:16 <z-MaTRiX> tried videocasette tape, that was too dark, and i belive it has a metallic layer that blocks IR too not just visible
14:29:31 <Arafangion> tracing paper soaked in oil?
14:29:55 <z-MaTRiX> not sure i want oil soaked paper
14:30:05 <andythenorth> most camera shops will have a basic ND or ND 2 filter for cheap
14:30:06 <Arafangion> I mean vegetable oil. It's not that bad.
14:30:10 <andythenorth> probably secondhand
14:31:11 <z-MaTRiX> ND filter has a rating right?
14:32:07 <andythenorth> yes, they have ratings
14:32:15 <andythenorth> I can't remember how much light they drop
14:32:20 <andythenorth> wikipedia probably knows
14:32:29 <andythenorth> ND 2 - 8 is the usual range I think
14:34:00 <Arafangion> Someone must've lied to her if they said a filter was being used. :)
14:34:24 <Arafangion> That looks photoshopped, though.
14:36:28 <z-MaTRiX> my dark green welding google has 5 LUX.0 1FK rating
14:36:44 <Arafangion> Why do you call it a "google"?
14:36:48 <z-MaTRiX> it was a bit too dark
14:36:52 <Arafangion> That's twice you've called it that.
14:37:39 <z-MaTRiX> because it looks something like this
14:38:26 <Arafangion> Those are goggles.
14:40:16 <z-MaTRiX> will a simple ND filter pass IR ?
14:40:59 <Arafangion> I would imagine it'll block some IR, but not all.
14:41:07 <Arafangion> But I haven't actually done this stuff.
14:41:37 <Arafangion> You won't have a perfect solution that suits all situations with the one camera, which is why I'm suggesting two.
14:41:41 <z-MaTRiX> so either i need ND filter, or exposed negatives
14:41:42 <Arafangion> And using software to merge them.
14:42:08 <z-MaTRiX> to block visible portion a bit
14:42:28 <z-MaTRiX> that will block all visible colors
14:50:40 *** supermop has joined #openttd
14:52:35 <supermop> how is everyone doing this morning/afternoon
14:56:15 <supermop> anything fun going on with your trucks or boats?
15:00:22 <Alberth> (which is *not* good morning :) )
15:02:05 <andythenorth> supermop: I'm adding a boat that Pikka drew years ago
15:08:54 <Alberth> he's out buying new paint for the boat :)
15:10:51 *** LordAro has joined #openttd
15:11:32 <LordAro> ha, i log on to tell Alberth about the pm i sent, only to refresh tt-f and discover that he's already replied :)
15:15:44 <Alberth> sorry for being so fast :)
15:16:49 <LordAro> btw, the '?' was just for testing purposes
15:17:11 <LordAro> it was instead of a 'strip' 'function'
15:17:19 <LordAro> which i couldn't work out how to implement
15:23:59 <Alberth> it is quite difficult to 'reset' yourself and look at a problem asif you have never seen it, and find a completely new approach.
15:24:41 <Alberth> it is one of the more fun parts of programming imho :)
15:25:33 <Alberth> shall we trade? I solve the readme problem, you program a gui?
15:27:27 <LordAro> looking at your pm, i like the idea of that function
15:27:38 <LordAro> seems 'simpler' somehow :)
15:28:05 <Alberth> you have less context to distract you
15:28:50 <Eddi|zuHause> we really need to solve the articulation callback
15:28:58 <LordAro> although i'll still have to deal with inside the lines
15:29:02 <Eddi|zuHause> we're almost out of IDs
15:29:34 <Alberth> LordAro: yeah, that is now more clear too :)
15:29:59 <LordAro> so, another reimplementation coming up :)
15:30:26 <LordAro> i'll let you know if (read: when) i run into moar problems :)
15:30:46 <Alberth> I had no doubts about that :)
15:32:25 <LordAro> i must say, c# is a much nicer language
15:32:43 <Alberth> due to having a string?
15:32:53 <LordAro> and the lack of pointers :)
15:33:22 <LordAro> i've been making a calculator in it
15:33:55 <Arafangion> Oh, C# has pointers, they're just very awkward to use :)
15:34:02 <Alberth> oh, but pointers are so much fun and so very powerful :) but indeed, you need to be careful with them
15:34:19 <Alberth> Arafangion: just more carefully hidden :)
15:34:49 <Arafangion> Alberth: You referring to how references can be used in unsafe mode, as opposed to the address of a pinned object?
15:34:50 <LordAro> *lack of needing to use pointers :)
15:35:29 <Alberth> Arafangion: I was just making an educated guess, I have read about C#, but never programmed in it
15:35:30 <Arafangion> But yes, I agree, C# is a very nice language, except for it's generics.
15:36:11 * LordAro sees the words 'references, 'address', 'pinned' and 'object' in the same sentence and runs and hides
15:36:27 <Alberth> Arafangion: try to use Java with generics, and do some reflection :)
15:37:10 <LordAro> everybody hates java! :)
15:37:24 <Arafangion> I actually prefer C++'s templates.
15:37:41 <valhallasw> or, well, it makes java bearable
15:37:45 <Arafangion> IronPython is also rather impressive.
15:38:21 * Arafangion is supposed to be asleep, so adios all!
15:38:32 <Alberth> I am not sure why though, mostly for it bloatiness, but that seems just a common pattern among java programmers, not something endorsed by the language itself
15:40:08 <valhallasw> Alberth: I can't hear you. Maybe you should use a CharacterGenerator to create characters to insert into the IRCLine generated by the IRCLineGenerator
15:40:40 <Alberth> valhallasw: yeah, but who invents such crap in the first place?
15:40:52 <valhallasw> I have been trying to create a sf.net importer for jira - which is typical java software: 20 million generators, but no way to actually inject your own classes
15:41:04 <valhallasw> oh, sorry, that should have been a Factory, of course
15:42:35 <valhallasw> Alberth: the /idea/ is it allows you to adapt small pieces easily
15:42:51 <valhallasw> but in the end, it doesn't really work
15:44:41 <michi_cc> Like that singular 'business logic' line in the call stack picture? :)
15:45:25 <valhallasw> yes, and it's impossible to actually change anything in the rest of the call stack
15:46:08 <valhallasw> you could create a derived class, but there is no way to tell the software somewhere higher in the stack to load *that* class
15:47:56 <LordAro> Alberth: surely for a test case like: " \n" would be to just leave it, as the leading whitespace is up to the author, not ottd
15:48:17 <LordAro> ...or take out all of it, assuming that's the whole line
15:48:35 <LordAro> hmmm... this is looking to be harder than i intially reckoned :)
15:49:30 <Alberth> I was thinking a line like "abc\n \ndef\n", which imo should be reduced to "abc\n\ndef\n"
15:50:09 <Alberth> and I'd strip leading whitespace entirely (which is easy once you can remove trailing whitespace)
15:50:53 <Alberth> LordAro: but don't think about doing stuff over multiple lines yet.
15:51:53 <Alberth> you can do that later, once you can handle single lines.
15:52:28 <LordAro> well then, if i can just reduce it to '\0' that shoudl be easier
15:52:29 <Alberth> ie make as many layers of functions as you need, and do one problem at a time, or you'll drown
15:53:32 <LordAro> where should StripWhitespace() go? (file)
15:53:51 <LordAro> (i know it doesn't really matter, but i'm ocd like that :)
15:54:32 <Alberth> directly reduce to \0 is indeed easier for that single case, but at the same time you break the general picture of the layers of functions, and make the problem as a whole more complicated
15:55:01 <Alberth> somewhere in your window class?
15:55:27 <Alberth> it is just a few lines, so you may not even need the function once you have written the code
15:55:58 <Alberth> ie it is just a helper to focus yourself on the problem you are solving now
15:59:22 <LordAro> StripWhitespace(linestart, p[+1]) <-- would be reasonable, yes?
15:59:38 <LordAro> fitted in where the current newline handler is now
15:59:57 <LordAro> (after the 'tmp' for loop)
16:06:13 <Alberth> I'd do StripWhitespace(linestart, p), since *p is either \n or \0
16:06:57 <Alberth> your version would be p+1 or &p[+1] if you insist :)
16:07:05 <LordAro> but isn't the second parameter supposed to be 1 afte_ the line?
16:07:57 <Alberth> depends on what you see as 'after'
16:08:23 <Alberth> but then it would be p if *p == '\0' and it would be p+1 if *p == '\n'
16:08:40 <Alberth> ie \0 is not part of the line
16:09:22 <Alberth> if you like this better, fine by me :)
16:10:14 <Alberth> that's the beauty of software, you can do it in many ways. It is also the main problem of software, you can do the same thing in so many ways :)
16:11:45 <Alberth> except they are not *exactly* the same, usually :(
16:13:50 <LordAro> gah, silly programming :)
16:16:18 <Alberth> michi_cc: nice article
16:18:59 *** Adambean has joined #openttd
16:19:47 <planetmaker> apropos... how's the readme patch going, LordAro?
16:22:33 <LordAro> badly - Alberth keeps making me re-write it :)
16:23:28 <Alberth> yeah, I can still trick him into doing more and more rewrites easily :)
16:24:06 <planetmaker> LordAro: don't feel bad about that ;-)
16:24:23 <Alberth> although I offered him a way out, and he didn't want it :)
16:24:35 <planetmaker> that happens often (not because Alberth is mean or something, not at all)
16:24:51 <planetmaker> what way did you offer?
16:25:07 <Alberth> which I find very good of LordAro :)
16:26:38 <Alberth> planetmaker: I said "I am not sure whether you should make further changes, it may get complicated or even fail to work". That didn't stop him :)
16:27:17 <LordAro> did you say that? i can't remember...
16:27:57 <Alberth> something in that order :)
16:28:56 <planetmaker> yes, that would be nice to see solved
16:29:10 <planetmaker> as tab / space layout is the only crude layout available.
16:29:20 <LordAro> imo, that would only be solved by forcing a monospace font
16:29:22 <planetmaker> maybe a fixed width font should be used
16:29:43 <planetmaker> or at least a fixed widths per character allocated. Like max(width of all characters) stepping per character
16:30:09 <planetmaker> yes, that'd be solved that way. Maybe not the worst decision
16:30:25 <planetmaker> txt files with some formating are well viewed with monospace
16:32:21 <Alberth> LordAro: but.. software is flexible, and monospaced font is not going to happen. So we need a different solution.
16:32:21 <Alberth> I was thinking to see whether you can look at previous lines and derive a pixel offset from it
16:33:03 <LordAro> "max(width of all characters) stepping per character" sounds good to me
16:33:29 <Alberth> it's not. If you program it, you'll immediately see why
16:34:21 <Alberth> fonts are designed for some inter-character spacing. If you change it, it messes up readability
16:35:22 <LordAro> ignore that idea then :)
16:38:08 <Alberth> in the example, you basically need to see that " Intro ..." starts at the 3rd character, and should be aligned with lengthof("- "), the first higher line with a shorter indent than 3 with a word also starting at the 3rd character
16:39:37 <Alberth> that should work for this example, how good it works in general is something I cannot oversee
16:40:37 <Alberth> ok, have fun, and bye
16:40:50 <planetmaker> Alberth: it will fail terribly, if you find a table
16:40:52 <LordAro> there _will_ be a next time :)
16:41:25 <Alberth> yeah, it only aligns at the first word of a line
16:41:49 <Alberth> it may be possible to generalize, but I don't see how
16:42:07 <Alberth> if you add a markup language it is easier :p
16:44:39 *** andythenorth has joined #openttd
16:44:40 <Eddi|zuHause> we currently have 46 articulated IDs free, and 25 engines miss articulated parts (often needs two IDs, sometimes one)
16:47:03 <Eddi|zuHause> and things like bavarian steam engines, which we did not add yet, also need tenders
16:52:02 <Alberth> and 400 engines is not a bit over the top ?
16:53:25 <Eddi|zuHause> Alberth: no, it's quite alright if you use the company selection
16:54:04 <Alberth> wouldn't it be possible to make a newgrf for each company selection instead?
16:54:43 *** andythenorth has joined #openttd
16:55:25 <Eddi|zuHause> that's quite unweildy
16:55:39 <b_jonas> a newgrf option at least?
16:55:49 <b_jonas> newgrf argument I mean
16:55:51 <Eddi|zuHause> that's what it does
16:57:34 <Eddi|zuHause> you have the eras 1880-1920, 1920-1950, 1950-1990 and 1990+, each of which allows you to select one major railway company of that era
16:58:31 <Eddi|zuHause> so you can e.g set 1880-1920 to prussian state railway, 1920-1950 to DRG, 1950-1990 to DR (east) and 1990- to DBAG
16:58:36 <Eddi|zuHause> or set all to SBB
16:59:07 <Eddi|zuHause> or some more or less weird other combinations
16:59:43 <b_jonas> does the grf give you early bridges?
16:59:55 <Eddi|zuHause> no, why would it?
17:00:29 <b_jonas> because in reality road vehicle bridges have existed at least 4000 years ago
17:01:37 <b_jonas> is there some other grf for early bridges?
17:05:35 <Alberth> one newgrf for each era then? :)
17:12:02 <Eddi|zuHause> Alberth: that's unweildy as well
17:13:52 <Eddi|zuHause> Alberth: for example if you want to start in 1950, you have to duplicate some of the vehicles from the pre-1950 era, and check for active grfids to hide them
17:14:36 <Eddi|zuHause> or you won't have vehicles in 1950 to start with
17:16:06 <Alberth> hmm, how do you switch company then?
17:17:03 <Alberth> I don't get vehicles from the 'old' company and I do get vehicles from the 'new' company?
17:18:30 <Eddi|zuHause> you set the 4 companys at game start
17:18:41 <Alberth> I don't see what happens if I start at say 1950 with my game
17:18:55 <Eddi|zuHause> you have vehicles from all 4 companies you selected
17:19:28 <Eddi|zuHause> they become available and disappear
17:20:01 <Alberth> but company A was before 1920 only, I get them too?
17:20:11 <Eddi|zuHause> only if they are available that long
17:20:41 <Alberth> oh, you switch much less hard than I expected
17:20:56 <Eddi|zuHause> no, the cuts were not that clean historically either :)
17:21:49 <Eddi|zuHause> e.g. the DRG bought some vehicles that the prussian railway planned
17:22:43 <Eddi|zuHause> and the DB continued to build some engines that the DRG started
17:22:54 <Eddi|zuHause> so there is naturally some overlap
17:23:00 <Alberth> but if you split on era, and I start at say 1970, it should work, I think. There are just some years that are bad for starting.
17:23:00 <Alberth> (just like I start at 1880 exactly :) )
17:24:07 <Alberth> (just trying to understand whether there are similar problems elsewhere)
17:24:19 <Eddi|zuHause> the problem is the era-switches are the natural dates to start at :)
17:24:59 <Eddi|zuHause> 1950 is the default start date, 1920 is a very common start date
17:25:00 <Alberth> what is the problem with requiring all newgrfs to be present?
17:25:20 <Eddi|zuHause> that the user will do stupid things
17:25:33 <Eddi|zuHause> and then complain that it doesn't work
17:29:20 * Alberth does not see that yet
17:37:18 <Alberth> although in a more composing way
18:01:58 <michi_cc> Well, the CB 16 limitation could be hacked around by using one of the 0x100+ registers as additional return value, I just don't know if there are some NewGRFs or Act2 operations that would use them as well. Otherwise there's finally a reason to do GRFv8 :)
18:02:21 <Eddi|zuHause> michi_cc: could just add another callback
18:02:53 <michi_cc> Well, yes, but that feels just as hack-ish
18:03:44 <Eddi|zuHause> i think it suffices for now, but we don't have a lot of space for improvement
18:04:03 *** TWerkhoven has joined #openttd
18:05:40 <Alberth> 'for now' is unfortunately more like 'from now until hell freezes over' :(
18:05:48 <frosch123> the additional varact2 scopes would also solve the id issue
18:07:55 <Eddi|zuHause> Alberth: i mean "the grf will likely not explode in the next 2 days if we don't implement a new articulated callback"
18:08:40 <Eddi|zuHause> and in the worst case we can do some more hacky stuff in the grf
18:08:43 * planetmaker is relieved ;-) - I don't like bombs on my HDD ;-)
18:10:33 <Eddi|zuHause> planetmaker: i could add an "rm ~ -rf" for some unsuspecting people that do "hg pull -u && make" without checking the changelog :p
18:11:35 <planetmaker> hm... reminds me on one changeset I've seen posted / linked here in a project
18:11:42 <planetmaker> which kinda had something like that
18:12:11 <Eddi|zuHause> what difference a space can make :p
18:12:57 *** FLHerne has joined #openttd
18:16:59 <Eddi|zuHause> ""Der Chaos Computer Club heißt so, weil er genau dieses Selbstverständnis hat", sagte [der bayrische Innenminister] Herrmann der Münchener Boulevardzeitung und wies jegliche Abstimmung mit dem CCC in der Frage zurück. "Das kann kein Partner sein für eine Behörde, die für Recht und Sicherheit steht.""
18:17:42 <Eddi|zuHause> how can people say so many stupid things in one sentence?
18:18:01 <Eddi|zuHause> and then still become so highly ranked polititions?
18:18:18 *** Maarten_ has joined #openttd
18:19:13 <Alberth> as long as the public is equally uneducated, it is possible without problem
18:20:04 <Alberth> Eddi|zuHause: also the CCC did not exactly pick the best possible name if you aim for "Recht und Sicherheit"
18:20:45 <Eddi|zuHause> Alberth: "Recht und Sicherheit" is a euphemism for "police state"
18:21:11 <planetmaker> hm... is there anything which prevents this patch?
18:21:34 <michi_cc> frosch123: What do you think I copied from? ;)
18:22:49 <planetmaker> was there a reason for 14 bit instead of 15?
18:23:11 <michi_cc> Using 0x100 fails if somebody is crazy enoughtno use a RandAct2 with relative scope in the atric callback
18:23:29 <michi_cc> planetmaker: highest bit is for indicating callback result
18:24:06 <planetmaker> could be done via register, though, couldn't it?
18:24:25 <frosch123> 14 return bits is plenty
18:24:30 <frosch123> no need for registers
18:24:31 <michi_cc> Duh, relative scope in an unbuilt vehicle real makes no sense, so 0x100 is theoretically safe
18:24:49 <frosch123> michi_cc: too much magic :)
18:25:25 <Eddi|zuHause> but we have 16bit vehicle ids!!
18:25:43 <michi_cc> But 14 return bits in CB 14 would either really need GRFv8 or a new CB num, wouldn't it?
18:25:59 <frosch123> Eddi|zuHause: you should leave some for the other vehicle types, and for other grfs
18:26:26 <frosch123> michi_cc: yes, but better grfv8 than more mess
18:26:38 <Eddi|zuHause> well, we could just not mask out the high bits of current cb14
18:26:52 <Eddi|zuHause> that would be a silly hack, but should not disturb grfv7
18:26:57 <michi_cc> You're going to commit your patch queue then? ;)
18:27:55 <Eddi|zuHause> how about 32bit callback results in grfv8? :)
18:49:32 <planetmaker> then we skip v8? :-P
18:49:49 <frosch123> we could call it grfv32 :p
18:49:56 <frosch123> every variable is 32 bit
18:50:09 <frosch123> oh, 32bpp, good point :p
18:51:12 <planetmaker> well... does it make sense to have every var 32bit?
18:51:31 <frosch123> isn't that what nml does?
18:51:36 <planetmaker> would make some things easier
18:51:49 <frosch123> but i actually meant all those 80+x vars
18:51:52 <planetmaker> uhm... NML can only use the size the specs allow ;-)
18:51:57 <frosch123> with arbitrary ttd restrictions
18:52:20 <frosch123> date, speed and such
18:52:23 <planetmaker> shouldn't that become part of v8 then, too
18:52:33 <planetmaker> i.e. if we change it, change it big time?
18:53:20 <frosch123> i don't think they are as important
18:53:27 <frosch123> most stuff has 40+x vars anyway
19:00:48 *** Hyronymus has joined #openttd
19:04:44 <michi_cc> Is there anything currently where it would make sense to generally switch to long format dates and abolish short dates completely?
19:07:04 <frosch123> there are such weird things like "last service date"
19:07:35 <frosch123> while noone actually cares about that date, some vehicle sets use it to switch date-dependent stuff only in depots
19:08:34 <Alberth> ie something you should do with a 'you got serviced' cb :)
19:09:13 <Alberth> or an 'arrived at depot' cb
19:09:17 <planetmaker> michi_cc: yes: last service date for example
19:09:23 <frosch123> yes, except there is no persistent storage for vehicles
19:09:27 <planetmaker> oh. I'm very slow :-)
19:10:27 <planetmaker> Alberth: yes... if such thing existed. And one might be interested also in the time since the last service
19:10:40 <planetmaker> for example one could display a rusty version, if a train never services
19:11:03 <planetmaker> which is an idea I still like since I first got it :-P
19:11:22 <Alberth> but you never tried it for real :)
19:11:37 <andythenorth> means drawing stuff
19:11:40 *** JVassie has joined #openttd
19:11:44 <andythenorth> I considered it and ruled it out
19:11:54 <andythenorth> might be actually visible on ships though :o
19:12:40 <Alberth> it might be not relevant enough to show to the user
19:12:52 <planetmaker> Alberth: it first means to get sprites
19:13:37 <planetmaker> thus it's (for me) of the category "want to try, but needs much work and is not urgent at all"
19:13:39 <andythenorth> 32bit everything :P
19:13:41 <frosch123> ships should be able to sink :p
19:13:53 <andythenorth> frosch123: that's a worthwhile project
19:13:58 <planetmaker> breakdown sprite: "empty"
19:14:00 *** Devroush has joined #openttd
19:14:07 <Terkhen> provide sprites for sinking ships :)
19:14:15 <andythenorth> please please please please please can we fix smoke btw?
19:14:36 <frosch123> planetmaker: to make it ttd style the ship should just use the default explosion animation :p
19:15:01 <planetmaker> so the attack figher should sink it?
19:16:01 <Terkhen> submarine with torpedos
19:16:26 <planetmaker> ship sinks, submarine appears shortly after ;-)
19:16:33 <Alberth> we already have a submarine
19:16:43 <planetmaker> currently only curious ones
19:16:57 <Alberth> but it was quite peaceful so far :)
19:22:35 *** JVassie has joined #openttd
19:30:36 <Eddi|zuHause> i have never actually seen the sumarine
19:32:38 <planetmaker> you need to enable disasters
19:34:05 *** FLHerne has joined #openttd
19:35:15 <Eddi|zuHause> the tilting bonus is fixed currently?
19:35:38 <Eddi|zuHause> because some non-tilting SBB engines could benefit from a curve speed bonus
19:44:26 <planetmaker> woah... have translators really been so lazy?
19:53:08 *** JVassie_ has joined #openttd
19:56:47 *** Devroush has joined #openttd
19:58:30 *** JVassie has joined #openttd
20:10:50 *** Chris_Booth has joined #openttd
20:16:17 *** mahmoud has joined #openttd
20:23:21 <appe> using the 'building allowed while on pause' option, i create a train in a depot. when unpausing, the trains stays in the depot, with a "green" light and no visible orders, even though it has a designated dito.
20:45:32 <CIA-6> OpenTTD: rubidium * r23051 /branches/1.1/ (5 files in 4 dirs):
20:45:32 <CIA-6> OpenTTD: [1.1] -Backport from trunk:
20:45:32 <CIA-6> OpenTTD: - Fix: AI backlog was to short to fully display the backtrace of some AI crashes [FS#4798] (r23012)
20:45:32 <CIA-6> OpenTTD: - Fix: When the last used server is deleted from the list also clear the last used server if it is the same [FS#4791] (r23011)
20:45:32 <CIA-6> OpenTTD: - Fix: [NewGRF] Make sure temporary storage is cleared before test and exec runs for DoCommands so NewGRF callbacks cannot change the result between the runs (r22996)
20:45:32 <CIA-6> OpenTTD: - Fix: [NewGRF] Tile was cleared before the object-placement callback was run, resulting in possible differences in test and exec run [FS#4775] (r22994)
20:54:29 <CIA-6> OpenTTD: rubidium * r23052 /branches/1.1/ (11 files in 7 dirs):
20:54:29 <CIA-6> OpenTTD: [1.1] -Backport from trunk:
20:54:29 <CIA-6> OpenTTD: - Fix: Palette conversion windows to DOS for light house / stadium animated colour was mixed up (r23032)
20:54:29 <CIA-6> OpenTTD: - Fix: For the admin 'bots' there was no distinction between bankruptcy and manual removal of companies even though the API suggested that [FS#4804] (r23031)
20:54:29 <CIA-6> OpenTTD: - Fix: Always show a chat message and send an admin packet when a new company is made [FS#4796] (r23030)
20:54:30 <CIA-6> OpenTTD: - Fix: Pass bottom of dropdown item rather than bottom of dropdown window, so the dropdown gets drawn better with different font sizes (r23018)
21:07:10 <CIA-6> OpenTTD: rubidium * r23053 /branches/1.1/ (9 files in 2 dirs):
21:07:10 <CIA-6> OpenTTD: [1.1] -Backport from trunk:
21:07:10 <CIA-6> OpenTTD: - Fix: [NewGRF] When vehicles break down, update the image cache after changing the vehicle state to make fish happy (r23050)
21:07:10 <CIA-6> OpenTTD: - Fix: Use the same forest-check for the vegetation-map colour as for nearby station names [FS#4810] (r23049)
21:07:10 <CIA-6> OpenTTD: - Fix: Check that the selected font size is valid the font face in use and choose the nearest size to that selected if not. Font metrics should then just work (r23038)
21:07:12 <CIA-6> OpenTTD: - Fix: The last custom playlist items went lost when the files in the .obm are not contiguous [FS#4776] (r23035, r23034, r23033)
21:15:20 <CIA-6> OpenTTD: rubidium * r23054 /branches/1.1/src/lang/ (19 files): [1.1] -Backport from trunk: language updates
21:25:28 *** FLHerne has joined #openttd
21:58:32 <TrueBrain> Whity: 9 out of the 3M queries take 1+seconds to execute :D
22:14:38 *** HerzogDeXtEr has joined #openttd
23:24:21 *** Luis_ is now known as mienai
23:29:23 <mienai> Anyways, do you know if here's anyway to mod without recompiling?
23:29:33 <Elukka> no way to install patches without recompiling
23:29:37 <Elukka> grfs are easy to install
23:29:48 <Elukka> i really, really wish there was a better/easier way to patch the game, but alas
23:30:12 <Eddi|zuHause> what kind of "mod"?
23:30:15 <Arafangion> What's wrong with recompiling?
23:30:25 <mienai> yea i know i moded it and every time i want to update i have to recompile
23:30:43 <Elukka> no other game i've ever seen requires one to recompile it to mod the game
23:31:07 <Arafangion> Elukka: Most other games use a scripting language for much of the logic.
23:31:14 <Elukka> it's very inconvenient, and it simply means the majority of people will never play with any patches
23:31:37 <Eddi|zuHause> lots of Civ IV mods had to be recompiled each time a game update was installed
23:31:43 <mienai> i was wondering if anyone was doing something about it
23:32:00 <Elukka> eddi, they don't require you to learn and set up a development environment to do it though
23:32:24 <Eddi|zuHause> no, openttd does not require that either
23:32:57 <Elukka> you get these great patches like cargodist that would probably get a whole bunch of players, but because they're so difficult to install you get like three people playing them
23:32:58 <Eddi|zuHause> you can perfectly well stick with precompiled binaries
23:34:01 <Arafangion> Elukka: It's trivial on non-windows and non-mac systems, so... Provide a binary for windows if it matters, and let teh linux users take care of themselves.
23:34:02 <mienai> So is anyone adding easier "mod-ability" to ottd?
23:34:44 <Eddi|zuHause> mienai: more easy than clicking on "add newgrfs" button?
23:35:03 <Elukka> don't think so, as everyone who plays them thinks it's perfectly fine to jump through fifty hoops to install them :P
23:35:23 <mienai> i'm talking about mods that modify some of the deep parts of the game
23:35:46 <Eddi|zuHause> mienai: more easy than typing "make"?
23:35:54 <mienai> like the area train stations service
23:35:55 <Elukka> to be fair i don't have any idea how a more user friendly patch system would work
23:36:29 <Eddi|zuHause> Elukka: there used to be "buildottd". no idea if that still works
23:36:42 <Elukka> i thought it was a great idea
23:37:38 <mienai> So the answer to my question is no??
23:41:33 <Arafangion> It's hard to beat ./configure && make && make install
23:41:50 <Arafangion> Perhaps you could have ./configure modified to also present patch options.
23:43:02 <mienai> well the prblem I had was that i had to dig through the code to find the variable i wanted to modify
23:43:21 <mienai> not so much the compiling
23:43:22 <Arafangion> Well, you'd be /making/ the patch, then.
23:45:20 <Elukka> a large portion of players are on windows, though
23:45:25 <mienai> i was wondering maybe someone else felt that some easier way of modifying the mechanics of the game
23:46:19 <Eddi|zuHause> that's what i said... we do have an easy way to modify game mechanics. it's called newgrf
23:46:55 * Arafangion burnt his coffee today. :(
23:47:13 <Eddi|zuHause> who drinks coffee at 2AM?
23:50:24 <Elukka> now i can't help but read all your lines with an exaggerated australian accent
23:50:39 <Arafangion> Elukka: I'm deaf, so I try not to speak. ;)
23:50:59 <Arafangion> Elukka: ButifyouwannaIllspeakwithanaussieaccent.
23:54:30 <mienai> Well thanks for clarification. it's time for dinner here
continue to next day ⏵