IRC logs for #openttd on OFTC at 2014-08-22
⏴ go to previous day
00:01:01 *** Stimrol has joined #openttd
00:45:51 *** Supercheese has joined #openttd
01:11:26 *** Flygon_ has joined #openttd
01:47:38 *** luaduck is now known as luaduck_zzz
02:56:13 *** KWKdesign has joined #openttd
03:33:23 *** ccfreak2k has joined #openttd
04:56:16 *** Eddi|zuHause has joined #openttd
05:40:51 *** HerzogDeXtEr has joined #openttd
06:02:36 *** oskari89 has joined #openttd
06:25:38 *** HerzogDeXtEr1 has joined #openttd
06:25:50 *** Zmastershabeeb has joined #openttd
06:31:09 *** sla_ro|master has joined #openttd
07:28:49 *** jpierre03_ has joined #openttd
07:33:35 *** FLHerne has joined #openttd
07:37:14 *** Zmastershabeeb has quit IRC
07:51:51 *** ccfreak2k has joined #openttd
08:14:31 *** looptrooper has joined #openttd
08:15:15 *** Eddi|zuHause has joined #openttd
09:16:19 <Eddi|zuHause> hm weird. i set up a raid in bios, but it won't show up in linux
09:16:31 <Eddi|zuHause> also, the bios seems to have a 2TB bug
09:17:54 <Eddi|zuHause> also, i tried to flash the bios, but it says "bios id error", indicating i downloaded the wrong model... but i triple checked...
09:19:12 <peter1138> so it's just a software raid then
09:19:50 <V453000> why woyuld you setup a raid boss
09:20:30 <planetmaker> for the drops, obviously
09:20:34 <TinoDidriksen> Hardware RAID is dangerous. Unless you have several of the exact same card laying around for when it fries, just use software RAID.
09:20:42 <Eddi|zuHause> internet says it's better when you want to access it from windows
09:21:04 <Eddi|zuHause> TinoDidriksen: it's not an actual hardware raid
09:21:23 <peter1138> then you just answered your own question :p
09:21:39 <Eddi|zuHause> peter1138: that was not the question...
09:21:49 <peter1138> well it wasn't even a question
09:22:15 <peter1138> linux mostly does not bother supporting fakeraid
09:22:18 <V453000> probably some high lvl pwnzor raid
09:22:56 <peter1138> TinoDidriksen, raid1 should, in theory, be relatively safe. but you never know.
09:23:44 <TinoDidriksen> peter1138, it's not, because the metadata is in the card version's own special format. Been there, tried that, cursed a lot, went software.
09:25:30 <peter1138> if the card is sensible it will place the metadata at the end of the disk, but that's probably too much to ask
09:25:48 <peter1138> but still the data should be intact even if you can't directly boot it or whatever
09:25:59 <peter1138> which is sometimes easier than restoring from backup
09:26:23 * peter1138 has some hw raid6 cards lying around...
09:27:03 <Eddi|zuHause> what does raid6 do anyway?
09:27:59 <peter1138> makes it all more complex
09:28:23 <peter1138> and in theory gives you more redundancy than raid5
09:29:02 *** tokai|noir has joined #openttd
09:29:02 *** ChanServ sets mode: +v tokai|noir
09:35:10 <Xaroth|Work> Eddi|zuHause: tried ZFS?
09:35:25 <Eddi|zuHause> Xaroth|Work: for what?
09:35:35 <Xaroth|Work> instead of hardware raid
09:35:51 <Xaroth|Work> because it's.. somewhat better than hardware raid
09:36:15 <Xaroth|Work> and doesn't have some of the pitfalls that hardware raid has
09:36:47 *** luaduck_zzz is now known as luaduck
09:37:20 <Xaroth|Work> (like having to use dodgy raid cards with their own dodgy format of storing metadata that isn't compatible with other brand dodgy cards)
09:52:28 *** Eddi|zuHause has joined #openttd
09:54:17 <Eddi|zuHause> so now things are different kind of weird... i made a software raid, and it's now saying "activating md even if degraded" or something
10:00:14 <peter1138> Xaroth|Work, zfs supports different raid levels by itself?
10:01:57 <Eddi|zuHause> "You can safely ignore the message"
10:02:22 <Xaroth|Work> peter1138: in what way do you mean exactly?
10:02:42 <peter1138> rather than say, using mdraid like normal people do
10:02:52 <Xaroth|Work> it does 1, 0, 5(z), 6(z2)
10:03:28 <Xaroth|Work> it's not just a raid-like system, it's also a volume manager
10:04:02 <peter1138> i prefer my file systems to be file systems, my volume managers to be volume managers, and my software raid to be software raid.
10:04:12 <Eddi|zuHause> which brings us back to the question... why?
10:04:32 <Xaroth|Work> peter1138: if you have 3 split systems, one is not fully aware of the other
10:05:14 <peter1138> i don't think i've ever needed them to be
10:05:42 <Xaroth|Work> so when your raid array has silent corruption, and figures it out
10:05:52 <Xaroth|Work> how does the rest know what has suffered bitrot?
10:06:12 <peter1138> if zfs is handling volumes and raid, then where do you put swap? on a file? heh
10:08:12 <Eddi|zuHause> i finally need to install grub on my ssd... that never worked...
10:09:28 <peter1138> never had any problem with that, though i'm on a bios system
10:09:58 <Eddi|zuHause> peter1138: whenever i tried to put grub on a new disk, it made things ever worse...
10:10:09 <Eddi|zuHause> like using the wrong drive order
10:10:18 <Eddi|zuHause> or getting into an endless restart loop
10:10:26 <peter1138> zfs sounds horrible, i'm glad i don't use it
10:10:33 <Eddi|zuHause> and let's not speak of grub2 :/
10:10:49 <Xaroth|Work> so you see 2 things on something, and you call it horrible? that's a well educated oppinion right there
10:13:05 <peter1138> i see the bit about it not being in linux
10:13:28 <peter1138> if it's not in debian i don't give a shit :)
10:13:40 <Xaroth|Work> seriously, do you even google?
10:13:47 <peter1138> packages FOR debian is not IN debian
10:14:11 <^Spike^> sorry bad experiences... :)
10:14:20 <^Spike^> better say way too much bad experiences :)
10:14:24 <peter1138> will not touch with a giant bargepole
10:15:02 <Xaroth|Work> ^Spike^: what kind of issues?
10:15:12 <Xaroth|Work> i've been fiddling around with it for a while and everything's been running just fine
10:15:14 <^Spike^> stuff with large disk sets
10:15:19 <^Spike^> stuff with raid controllers
10:15:29 <Xaroth|Work> raid controllers for a software raid system?
10:15:32 <^Spike^> stuff with it being f***** slow on freebsd with lots of snapshots
10:15:46 <^Spike^> stuff with it killing a production server bi-weekly atm
10:16:42 <Xaroth|Work> lots of snapshots being how many?
10:16:47 <^Spike^> in the end... zfs is A file system and not THE filesystem
10:17:00 <^Spike^> 400+ subvolumes with each i think 8-10 snapshots
10:17:13 <Eddi|zuHause> so, why does the software raid consume like 20% cpu and constant disk accesses, even though nothing is actually accessing the disk?
10:17:32 <^Spike^> because software raid.....
10:17:45 <Eddi|zuHause> but... what is it doing?
10:18:00 <^Spike^> if i knew... ottdc server would be optimized already :)
10:18:37 <^Spike^> lots of disk actions from sw raid
10:18:47 <Eddi|zuHause> also, why does google open in a noscript-y browser consume so much cpu?
10:18:52 <^Spike^> cause we don't constantly try to fill disk/clean disk... :)
10:19:44 <Xaroth|Work> ^Spike^: so you cram it full of crap (400+ subvols), and then you claim it's full of crap? yeh, I see how that's ZFS' issue :P
10:20:06 <^Spike^> it's said it could handle it... :)
10:20:20 <^Spike^> it's called user separation when you offer a service... ;)
10:20:23 <peter1138> Eddi|zuHause, probably synchronizing, maybe.
10:20:24 <^Spike^> and quota management :)
10:20:48 <^Spike^> we are switching over to get rid of the crappy sub volume stuff.... cause software understands quotas in ldap now...
10:20:55 <Xaroth|Work> ^Spike^: ever seen what happens if you create 200+ ext4 volumes to a system? I'm somewhat amazed that it actually has some performance with 400 :P
10:21:24 <^Spike^> ehm... considering the storage backend @ work.... + proper lvm....
10:21:31 <^Spike^> i think it would come quite far :)
10:24:20 <^Spike^> although lvm overhead could be killing it in one way... but with a proper hw raid controller it should work quite well
10:24:31 <Xaroth|Work> ^Spike^: side question; how much ram and what features were enabled?
10:25:07 <^Spike^> almost no features as far as i remember and about 16G ram
10:25:20 <^Spike^> and about 100T storage space
10:25:29 <Xaroth|Work> 16g on 100t? that's a bit low tbh
10:26:01 <Xaroth|Work> should work, but if your arc isn't tuned then it might cause issues
10:26:03 <^Spike^> i already noticed zfs does too much in ram when we were calculating deduplication vs ram needed
10:26:34 <Xaroth|Work> yeh, but that's just how (and why) ZFS works
10:26:38 <Xaroth|Work> RAM ALL the things :P
10:26:57 <^Spike^> it was better to buy a new EMC with more features... then buy an enterprise server for that price....
10:27:30 <Xaroth|Work> that somewhat depends on what kind of hardware you're running it on
10:27:58 <Xaroth|Work> i mean, 100T on big disks, add a few SSDs for L2arc and a ZIL and you'll still be at half the cost of an enterprise storage device
10:28:13 <Xaroth|Work> because they all want the 'enterprise' 15k sas drives
10:29:18 <^Spike^> hmmm will keep that in mind... :)
10:29:32 <^Spike^> doesn't mean we might still change things... but we are working with the current platform... and still don't like zfs :)
10:29:39 <^Spike^> as said it's a filesystem and not the filesystem :)
10:33:32 <Xaroth|Work> and I don't like paying 600 euro for 1TB of space, if I can get 12TB of space AND a SSD for L2arc/zil at the same price :P
10:45:05 <SpComb> ZFS has a lot to offer in terms of management for large disk systems
10:45:36 <Eddi|zuHause> hm, there see,s to be a process called "ext4lazyinit" running that consumes i/o
10:45:38 <SpComb> like I'm not really sure I'd want to mdraid/lvm/ext4/rsync some box with two dsik shelves and 127 disks
10:46:10 <SpComb> whereas with ZFS it's not really any more difficult than a box with 2 disks
10:46:56 <SpComb> it's not like LVM is perfect either..
10:47:18 <SpComb> Eddi|zuHause: are you going to be dual-booting between windows and linux and want the same raid on both?
10:48:08 <Eddi|zuHause> SpComb: when i now formatted it ext4, probably not...
10:48:55 <Eddi|zuHause> sometimes i really wished windows could just read all the linux file systems :/
10:49:38 <SpComb> I think I've mostly given up on dual-booting these days
10:49:46 <SpComb> laptop is just Ubuntu, with Windows in a VM inside
10:49:57 <SpComb> desktop is still dual-boot but I haven't booted it into Ubuntu for a while
10:50:00 <Eddi|zuHause> i've tried windows VM
10:50:17 <Eddi|zuHause> but kvm/qemu somehow uses 30% cpu when idle
10:50:39 <SpComb> kvm/qemu graphics imo kinda sucks, VirtualBox seems to just work better for desktop stuff
10:50:56 <Eddi|zuHause> how long is this ext4lazyinit going to run?
10:50:59 <SpComb> but yeah, the Windows 8.1 inside likes to peg the cpu once in a while
10:51:15 <Eddi|zuHause> i only have XP to play with
10:51:33 <Eddi|zuHause> most stuff i do in wine, but some things just don't work
10:51:34 <SpComb> Eddi|zuHause: never encountered ext4lazyinit before
10:57:28 *** SylvieLorxu has joined #openttd
11:02:54 <peter1138> Hmm, I don't think it should be sticking around, unless it's waiting for something else to finish, maybe.
11:03:16 <^Spike^> Xaroth|Work: what helps in the end (special in the environment i work in) with EMC for example if it breaks they come by to fix it... if shit hits the fan really hard... they come to help out :)
11:03:16 <Eddi|zuHause> SpComb: apparently it's some process that zeros stuff that was left out from formatting
11:03:56 <^Spike^> with zfs if it breaks... you're on your own basicly (and a community perhaps) as far as i've seen
11:04:12 <^Spike^> and telling that to customer when you got 2k VMs offline... doesn't go to well i can tell you :)
11:04:56 <Eddi|zuHause> man, mount.ntfs is some inefficient shit...
11:06:24 <SpComb> ^Spike^: that's normal... if you have the money, you should concentrate on the actual product and not the infrastructure around it... outsource that part
11:07:05 <^Spike^> it also depends on what you use it for in the end
11:07:15 <^Spike^> in this case running clusters of VMs
11:07:45 <^Spike^> for something easier/simpler it might be enough to kick 3 boxes into a DC put for example gluster on it and hf
11:08:11 <Xaroth_> ^Spike^: there's always Nexenta
11:08:21 <Xaroth_> they are some of the core contributors for the BSD branches
11:08:27 <Xaroth_> and they procide extra crap like replication
11:08:34 <Xaroth_> which isn't in 'normal' zfs
11:08:35 <^Spike^> trying to think.. cause i recognize the name.... but i think we kicked that out aswell
11:09:30 *** Pikkaphone has joined #openttd
11:10:07 <^Spike^> yup deleted device...
11:22:49 <SpComb> oh, so someone does actually do zfs live replication?
11:23:26 <SpComb> my stack is ganeti, so something like lvm/ext4/drbd/rsync... it would be nice if ZFS really did all of that... and then someone implemented support for it in ganeti :)
11:31:43 *** Supercheese is now known as Guest1659
11:31:44 *** Supercheese has joined #openttd
11:46:46 *** Klanticus has joined #openttd
11:53:35 <Xaroth|Work> SpComb: Nexenta apparently does it
11:53:45 <Xaroth|Work> but that's their enterprise edition
12:06:01 *** KWKdesign has joined #openttd
12:18:29 *** kostiak has joined #openttd
12:22:01 * planetmaker doesn't see anyone
12:22:16 <kostiak> I'm having a bit of a problem
12:22:41 <kostiak> I'm reading that trains are supposed to be availble as early as 1925 in temperate
12:23:03 <kostiak> yet in 1940 the railway constuction option is still greyed out
12:23:42 <planetmaker> that depends on climate. And if you play with NewGRFs
12:24:02 <kostiak> temperate, no newgrfs
12:24:29 <kostiak> kirby engine is supposed to be available since 1925
12:24:46 <kostiak> yet in 1940 no railway constuction option
12:25:00 <planetmaker> got a savegame for me?
12:25:13 <kostiak> where should i upload it?
12:25:23 <planetmaker> wherever. Dropbox maybe?
12:27:59 <planetmaker> that's arctic climate
12:28:32 <planetmaker> earliest train: 1945
12:28:48 <kostiak> well I'm stupid (as always)
12:28:56 <kostiak> i was 100% sure i picked temperate
12:30:32 <planetmaker> btw, you see the climate even before you load the save. The saveload dialogue tells you the climate, date and some other stuff about the save
12:30:55 <planetmaker> and you see it from the colour of the grass when you loaded it. But that, of course, needs some knowledge how stuff looks like :)
12:31:19 <kostiak> tbh, haven't played in like a year
12:31:24 <kostiak> and forgot all this stuff :)
13:03:56 *** Midnightmyth has joined #openttd
13:32:29 <planetmaker> why do you sell it, __ln__ ? :D
13:33:52 <SpComb> __ln__: you just need to pay a 50000€ deposit into an unnamed nigerian bank account to clear legal issues before your payment can be processed?
13:45:10 <^Spike^> they usually also offer to send pictures of their wife for some reason with those mails...
13:45:34 <^Spike^> and tell how you could really help her out
13:45:56 <V453000> what I want pictures of their wife
13:47:56 <^Spike^> .... i shall forward one of the mails when i have stand-by next time and need to keep an eye on the abuse mailbox :)
13:56:59 *** TheMask96 has joined #openttd
13:57:29 *** sla_ro|master has joined #openttd
14:20:20 <Eddi|zuHause> ext4lazyinit still running :/
14:27:40 <^Spike^> kill it all.......... ;)
14:27:44 <^Spike^> it's just a friday afternoon.. ;)
14:27:47 <^Spike^> nobody will notice :D
14:28:52 <^Spike^> hmmm lazy init should be gradual without impact... technically...
14:29:00 <^Spike^> lazy zeroing should do that normally atleast
14:37:21 *** Defaultti has joined #openttd
14:47:16 *** DanMacK has joined #openttd
15:05:38 *** Progman has joined #openttd
15:12:56 *** Hazzard has joined #openttd
15:13:32 *** Midnightmyth has joined #openttd
15:23:00 *** Progman_ has joined #openttd
15:24:27 <DorpsGek> Eddi|zuHause: 0.3312
15:24:31 <DorpsGek> Eddi|zuHause: 0.3264
15:24:47 <Eddi|zuHause> @calc .6*.76*.75
15:24:52 <Eddi|zuHause> @calc .6*.76*.70
15:24:52 <DorpsGek> Eddi|zuHause: 0.3192
15:24:55 <Eddi|zuHause> @calc .6*.76*.73
15:24:55 <DorpsGek> Eddi|zuHause: 0.33288
15:24:59 <Eddi|zuHause> @calc .6*.76*.72
15:24:59 <DorpsGek> Eddi|zuHause: 0.32832
15:28:48 *** Progman_ is now known as Progman
15:33:31 <keoz> toying out Eddi|zuHause ? :)
16:12:51 *** oskari89 has joined #openttd
16:13:58 *** Midnightmyth has joined #openttd
16:18:13 *** frosch123 has joined #openttd
16:28:39 *** davidstrauss has joined #openttd
16:30:53 *** Xaroth|W1rk has joined #openttd
17:03:02 *** trendynick has joined #openttd
17:09:53 *** Xaroth|W1rk is now known as Xaroth|Work
17:12:06 <DorpsGek> Commit by rubidium :: r26753 /trunk/src (saveload/station_sl.cpp station_base.h) (2014-08-22 17:11:59 UTC)
17:12:07 <DorpsGek> -Fix: desync due to not always properly restoring game state from the savegame
17:15:07 *** planetmaker has left #openttd
17:15:11 *** planetmaker has joined #openttd
17:15:11 *** ChanServ sets mode: +o planetmaker
17:36:37 <frosch123> LSky`: your desync log files were useful :)
17:45:35 <DorpsGek> Commit by translators :: r26754 trunk/src/lang/catalan.txt (2014-08-22 17:45:28 UTC)
17:45:36 <DorpsGek> -Update from WebTranslator v3.0:
17:45:37 <DorpsGek> catalan - 9 changes by juanjo
17:47:10 *** Alberth has joined #openttd
17:47:10 *** ChanServ sets mode: +o Alberth
17:49:03 *** luaduck is now known as luaduck_zzz
17:52:15 *** gelignite has joined #openttd
18:16:39 *** luaduck_zzz is now known as luaduck
18:29:25 *** andythenorth has joined #openttd
18:36:58 *** Zmastershabeeb has joined #openttd
18:37:39 <andythenorth> Alberth: any ideas how to run a python 3 nml in a python 2 compile?
18:38:18 <andythenorth> I’ve been relying on the coincidental fact that my scripts and nml both work with 2.6/2.7
18:38:21 <Alberth> if you run it as a separate process, how is it a problem?
18:38:37 <andythenorth> how will I teach nml to use python 3?
18:39:15 <andythenorth> I don’t want to modify the shebangs, that’s not good
18:39:15 <Alberth> #! /usr/bin/env python3 <-- by the first line in nmlc
18:40:14 <Alberth> but current nml2 has the above #!
18:40:17 <andythenorth> that’s interesting
18:40:27 * andythenorth tries something again
18:41:12 * Alberth bets it give the same result
18:41:38 <andythenorth> so either I change the shebang or alias python 3
18:41:42 <andythenorth> I’d misunderstood an error
18:42:41 <Alberth> I lost you, but that's probably fine :)
18:43:37 <Alberth> I tried to update the fish translation yesterday, and got stuck on the roster thingie iirc
18:43:51 <andythenorth> my python isn’t in /usr/bin/env so I need to alias from there
18:44:54 <Alberth> /usr/bin/env is a program
18:45:34 <Alberth> as parameter you give it the program you look for, and it returns the path to it, iirc
18:46:10 <andythenorth> so it knows about python3.3 but not python 3
18:46:23 <andythenorth> so I was going to symlink (alias) python3.3 to python 3
18:48:33 <Eddi|zuHause> wouldn't the package installer do that automatically?
18:48:59 <Alberth> Eddi|zuHause: you expect sane package installers at a mac box?
18:49:19 <andythenorth> we don’t use the package installers for python
18:49:23 <andythenorth> they’re all fucked for OS X
18:49:32 <andythenorth> we use python buildout, it’s reliable
18:50:08 <Eddi|zuHause> > dir $(which python3)
18:50:09 <Eddi|zuHause> lrwxrwxrwx 1 root root 9 6. Aug 23:39 /usr/bin/python3 -> python3.3
18:51:20 <andythenorth> make test should pass with nml tip and python3.3?
18:51:21 <Eddi|zuHause> yes, if you do things manually, do them either in
18:51:28 <Eddi|zuHause> ~/bin or /usr/local/bin
18:52:19 <Eddi|zuHause> i never understood the name "usr" anyway... certainly doesn't have anything to do with "user"
18:53:06 <Alberth> (20:51:20) andythenorth: make test should pass with nml tip and python3.3? <-- I just checked, it does
18:53:53 <Alberth> Eddi|zuHause: / is single user mode, /usr is mounted in multi-user mode, ie normal operation
18:54:26 <Eddi|zuHause> Alberth: that makes absolutely no sense
18:55:40 <Alberth> they are run states of the OS, probably hysteric rasins made /usr and it stuck around, as these things do
18:56:14 <andythenorth> 0.3.1 also fails
18:56:19 <andythenorth> but looks like missing PIL
18:56:20 <Alberth> andythenorth: no nml/tokens.py? (it's 9333 bytes here)
18:57:30 <andythenorth> 9333 bytes here too
18:57:51 <Alberth> python3 ; from ply import lex
18:58:24 <andythenorth> I’ll sort the deps out first
18:58:31 * andythenorth has to read how to make a virtualenv
18:59:00 <Alberth> pillow, actually, I don't think there is a real PIL for python3
18:59:08 <andythenorth> yeah I’ll get pillow
18:59:09 <Alberth> make a few notes today :p
18:59:58 <andythenorth> first I have to get virtualenv for python3
19:00:35 <andythenorth> which means I need pip for python3
19:00:42 <andythenorth> which means I need distribute for python3
19:03:02 <andythenorth> I’m posting here in case it’s some kind of use in future :P
19:03:36 <Alberth> the Internet is of course the ideal place to store such personal notes :p
19:04:18 <andythenorth> hth do I get pip to run under python3
19:04:29 <andythenorth> I don’t want to change my system python just for this
19:04:34 * andythenorth *hates* python packaging
19:04:38 <Eddi|zuHause> sure, just name your backup "Lost.all.seasons" and put it in a torrent. free distributed backup :p
19:05:07 <Eddi|zuHause> andythenorth: there is a pip3 of course
19:11:00 <andythenorth> and that is how we do it
19:11:18 * andythenorth should probably document that for other poor OS X users
19:12:00 <andythenorth> or I coud draw a boat :(
19:12:35 <Alberth> instead of getting an OS that can do these things, you manage your packages manually
19:12:53 <Alberth> and you pay big bucks to get to do it as well
19:14:03 <andythenorth> I work with linux devs who think it’s normal to recompile the kernel to try and fix a crashing wifi driver
19:14:17 <andythenorth> or who only have VGA support on their DVI-out laptop :P
19:14:37 <andythenorth> we all have our curses to bear
19:14:50 <Alberth> I did such things when I was young :)
19:15:30 <andythenorth> there’s nothing stopping one of us (OS X users) writing a decent package manager
19:15:37 <andythenorth> we just prefer to whine about it seems
19:16:27 <andythenorth> 32 nml tests passed
19:16:30 <Alberth> much easier borrow one from linuces
19:16:39 <Rubidium> andythenorth: isn't the OS X package manager called iTunes?
19:16:49 <andythenorth> App Store now :P
19:17:04 <andythenorth> Alberth: I think they started out by copying BSD Ports Tree, might have been a bad place
19:17:30 <andythenorth> updating a simple package can lead to a recursive port update that takes hours and leaves gcc tying up the cpu
19:17:38 <andythenorth> I guess that’s how a depgraph works though :P
19:17:55 <Alberth> obviously compiling all programs yourself is very useful :p
19:18:20 <andythenorth> anyway, what was I actually doing?
19:18:40 <Alberth> trying to run nml with pythno3 in something python2
19:19:01 <andythenorth> let’s see if I broke my compiles
19:19:06 <Alberth> as well as trying to answer my roster question, if you have seen it
19:19:37 <andythenorth> FISH compile works, and the deprecation messages about PIL tostring() stuff have gone
19:19:49 <andythenorth> what’s the roster question?
19:20:20 <Alberth> I don't understand what "roster" is
19:20:40 <Alberth> for translating to dutch
19:21:11 <Taede> roster: selection of vehicles available (for instance, dutch ships, british ships)
19:21:24 <Taede> not unlike industry set selection in firs?
19:21:38 <andythenorth> yes similar to FIRS economy
19:21:55 <andythenorth> same concept in Iron Horse and Road Hog
19:21:58 <Alberth> This Sceptered Isle / The Blue Danube refer to something?
19:22:07 <andythenorth> This Sceptered Isle <- Brit
19:22:14 <andythenorth> think it’s a Shakespeare line
19:22:49 <andythenorth> This precious stone set in the silver sea,
19:22:50 <andythenorth> Which serves it in the office of a wall
19:22:51 <andythenorth> Or as a moat defensive to a house,
19:23:06 <andythenorth> The Blue Danube is generic european river boats
19:23:24 <andythenorth> france/germany/switzerland/austria/hungary etc
19:23:35 <andythenorth> rhine / rhone / elbe / danube
19:25:32 <Alberth> nice reference :) thanks
19:28:27 <andythenorth> let’s try frosch’s patch...
19:31:19 <andythenorth> frosch123 so I should use the effect_spawn_model prop, not visual_effect?
19:31:23 <andythenorth> seems the obvious conclusion
19:43:44 <andythenorth> so can’t use python3 nml and python2 build
19:43:57 * andythenorth overlooked that nmlc is installed into the virtualenv
19:46:46 <andythenorth> I guess I could have the makefile switch virtualenv
19:49:49 <andythenorth> make install fails I think
19:53:24 <frosch123> that is not in my nml
19:55:07 <frosch123> your quote uses python2 syntax
19:55:13 <frosch123> so, what nml are you running?
19:59:34 <andythenorth> this is nml 0.4.0
19:59:42 <andythenorth> nml==0.4.0.r5307M-0be320b4fa1a
20:03:12 <andythenorth> well the smoke works
20:03:19 <andythenorth> frosch left too early :P
20:06:44 <andythenorth> steam hovercraft ftw
20:22:00 <andythenorth> maybe I should convert my compiles to python 3
20:26:47 <Eddi|zuHause> how does that help anybody?
20:27:42 <andythenorth> I don’t see how a python3 project will run in a python2 environment
20:27:50 <andythenorth> fundamentally, how is that supposed to work?
20:28:31 <andythenorth> surely it’s just a fool’s errand?
20:29:40 <Eddi|zuHause> the point is that you can easily have python2 and python3 simultaneously
20:30:16 <Eddi|zuHause> as long as one script doesn't call the other script, there's no overlap
20:30:46 <Eddi|zuHause> and if you'd use the makefile for calling scripts like normal people...
20:31:30 <andythenorth> this is the makefil
20:32:04 <andythenorth> if I call python3 nml with a python2 virtualenv active, it bails
20:32:09 <andythenorth> as you’d expect :P
20:32:17 <andythenorth> how is this not the expected result?
20:32:58 <Eddi|zuHause> in the makefile you have a line "PYTHON=blah", just replace that with two lines "PYTHON2=blah" and "PYTHON3=blub". then replace every $(PYTHON) with either $(PYTHON2) or $(PYTHON3)
20:33:43 <keoz> Changing the Makefile.in did the trick for me
20:33:46 <Eddi|zuHause> of course your "virtualenv" needs to have both python2 and python3 installed
20:34:11 <andythenorth> this seems like a lot of string
20:38:33 <Eddi|zuHause> oh, also every literal "python" needs to be changed into either of these
20:38:37 <andythenorth> I don’t understand how to install 2 pythons into a virtualenv
20:38:43 <andythenorth> that seems to be anti-virtualenv
20:39:09 <andythenorth> I think if I go into #python or whatever and ask how to switch pythons in a virtualenv, I’ll get kbanned
20:39:09 <Eddi|zuHause> your virtualenvs seem to be a terrible concept
20:39:13 <andythenorth> they’re not mine :O
20:39:21 <andythenorth> they’re the standard solution
20:39:35 <andythenorth> I have no other fricking idea how you do work
20:39:39 <Eddi|zuHause> is that like a chroot?
20:39:56 <andythenorth> honestly how do you get anything done without virtualenv?
20:40:15 <andythenorth> in any given day I am working on python 2.4, python 2.6, python 2.7 and now, python 3
20:40:23 <Eddi|zuHause> who says i get anything done?
20:40:24 <andythenorth> plus Apple have their own ideas about python
20:40:44 <Alberth> it has its own ideas about everything :)
20:41:32 <andythenorth> it used to have one idea
20:41:41 <andythenorth> one man, one idea
20:42:00 <andythenorth> does that magically run multiple pythons together?
20:42:13 <andythenorth> do your scripts just ‘know’ which one to use?
20:42:19 *** kais58__ has joined #openttd
20:42:19 <keoz> well at least here, I have python 2 and 3. Without problems.
20:42:23 <andythenorth> how do they know?
20:42:25 <Eddi|zuHause> they say that in the first line
20:44:18 <Alberth> #!/usr/bin/env python3
20:44:23 <Eddi|zuHause> this first line gets evaluated by your shell, when you type "./script.py"
20:44:57 <andythenorth> changing the shebangs is definitely seen as good practice
20:45:09 <Alberth> Eddi|zuHause: sure it's the shell, I think it deeper, in the kernel somewhere
20:45:25 <andythenorth> changing the shebangs on every file definitely wouldn’t get your commits reverted instantly
20:45:44 <andythenorth> anyway, I can’t see how to make progress
20:45:50 <Eddi|zuHause> and then you get the weird distributions who think /bin/python should be mapped to /bin/python3
20:46:04 <Eddi|zuHause> where every legacy python program will break
20:46:14 <andythenorth> or they just get a virtualenv...
20:46:22 <andythenorth> anyway the virtualenv is misleading
20:46:56 <andythenorth> even with virtualenv deactivated, nml is still borked
20:47:12 <andythenorth> it only works if I run it in a python3 virtualenv
20:47:33 <andythenorth> maybe it’s just missing deps
20:47:53 <andythenorth> so I have to install deps systemwide to get this to work?
20:48:02 <andythenorth> that is just not the right solution
20:48:08 <andythenorth> installing deps systemwide is always a problem
20:48:17 <andythenorth> always / often /s
20:49:06 <Eddi|zuHause> you can easily install the deps into the nml directory
20:49:16 <andythenorth> that sounds good
20:49:31 <andythenorth> add it to the path?
20:49:51 <Eddi|zuHause> the pythonpath variable?
20:50:08 <Eddi|zuHause> that should work
20:50:34 <andythenorth> I suspect that might be all a virtualenv really does :P I didn’t check
20:50:44 <Eddi|zuHause> there should be a Makefile.local where you can cram that stuff into
20:50:59 <Eddi|zuHause> or maybe Makefile.config
20:54:00 <andythenorth> I think if I just migrated my scripts to python 3 this would go away as a problem
20:55:33 <Eddi|zuHause> problems that are made to "go away" have a tendency to reappear
20:55:50 <Eddi|zuHause> like you solve it for FISH, now, next week you have to do it for FIRS as well
20:56:01 <Eddi|zuHause> and then for iron horse, and then...
20:57:47 *** Myhorta has joined #openttd
21:01:28 * andythenorth could just give in and add the deps systemwide
21:01:35 <andythenorth> I’ll get yelled at for it…later
21:09:23 <Eddi|zuHause> the key element to check for adding things systemwide is that you don't break stuff that is already working by doing so
21:11:30 <andythenorth> I’ve got a magic get out for most cases now
21:11:37 <andythenorth> which is that policy is to use virtualenv
21:11:48 <andythenorth> that only leaves non-work stuff to go wrong
21:12:01 <andythenorth> ply fails to install
21:12:03 <andythenorth> pillow installed ok
21:12:07 <andythenorth> but python3 can’t find it
21:12:34 <andythenorth> it’s in the site-packages dir that python3 is running from
21:13:35 *** Progman has joined #openttd
21:14:39 * andythenorth rtfm on PYTHONPATH
21:21:23 <Eddi|zuHause> somewhen i'm gonna add a manual called "m.rtf"
21:25:48 <andythenorth> tfm doesn’t explain why my python 2.7 is now importing from the python 3.3 site-packages
21:25:52 <andythenorth> so PIL is broken
21:25:59 <andythenorth> amongst other fun things
21:28:21 <andythenorth> I broke everything, time to stop
21:28:44 <andythenorth> I now have broken nml, broken Iron Horse, broken FISH, broken bash_profile, and broken system-wide python
21:28:47 <andythenorth> doing pretty well
21:29:38 <Eddi|zuHause> just a normal day at the office
21:30:10 <andythenorth> a good day’s work
21:30:12 <Eddi|zuHause> that's why whenever someone asks what i'm doing, i'm answering "i'm breaking everything"
21:31:12 <Eddi|zuHause> with the bonus that if it's actually broken afterwards i can say "i told you so"
21:33:22 <andythenorth> I will come back to this tomorrow, or another day entirely :P
21:36:01 <andythenorth> it’s not helpful that the version of nmlc created by make install appears to be borked
21:39:22 * andythenorth tests other tags
21:40:25 <andythenorth> 0.3.0 tests pass
21:43:59 <andythenorth> the effects patch works on nml 0.3.0 too
21:44:10 <andythenorth> although that solves not a lot :)
21:47:16 <Zuu> Or just stick to python 2.x?
21:47:30 <Zuu> That seems to be what most people do.
21:48:25 <andythenorth> nml has moved to python3
21:48:40 <andythenorth> the version that will be shipped to bundle server will be python3
22:00:10 <andythenorth> PYTHONPATH=../../lib/python2.6/site-packages && python2.7 src/build_fish.py
22:00:17 <andythenorth> in a python3 virtualenv
22:00:32 <andythenorth> and yes, 2.6 and 2.7 are not a typo :(
22:01:00 <andythenorth> definitely bedtime
22:11:31 *** Zmastershabeeb has quit IRC
22:26:22 <berndj-blackout> i've always wondered about openttd's splash screen - is it just an animation or is it running a simulation in the background?
22:32:05 <Zuu> berndj-blackout: It is a savegame running in the background
22:32:34 <Zuu> Copy it to opentile.sav and you can open it in the game.
22:33:36 <berndj-blackout> heh, cool, i'm totally going to do that right now
22:42:23 <berndj-blackout> *opntitle.dat btw; money everywhere!
22:42:59 *** Prof_Frink has joined #openttd
22:44:48 <planetmaker> probably cheated money, berndj-blackout
22:44:58 <planetmaker> check the cheat dialogue (ctrl+alt+c)
22:45:03 *** Zmastershabeeb has joined #openttd
22:46:27 <berndj-blackout> it doesn't show any operating cheats for me...
22:47:17 <planetmaker> then it's indeed well-earned money
22:51:00 <berndj-blackout> i'm not referring to the total amount, just to income popping up everywhere
22:54:12 <planetmaker> ah yeah. Pretty busy savegame for the 1.4 series :)
23:06:57 <berndj-blackout> in general, if i want to "up my game", what's a good idea to do?
23:11:45 <planetmaker> what do you understand under "up your game"?
23:12:19 <planetmaker> The best way is to just play. Maybe join a multiplayer server where you can play with other people and get direct feedback
23:12:27 <planetmaker> e.g. the #openttdcoop or reddit servers
23:13:56 <berndj-blackout> yeah i tried reddit's servers but apparently i need their custom client. i don't have the courage to compile openttd right now
23:14:29 <planetmaker> you only need that on some of their servers. Or join the #openttdcoop Welcome server
23:16:11 <Sylf> awww, that spaghetti juction is half way beautiful
23:20:19 *** ChanServ sets mode: +v planetmaker
23:21:11 *** Prof_Frink is now known as Guest3
23:43:16 <berndj-blackout> Sylf, care to elaborate? do you mean in an aesthetic sense or technically? i built it organically. originally just one line running between lake and city
23:44:47 <Sylf> There's nothing wrong with rails going everywhere in my play style, as long as they are all built with intention
23:48:19 <berndj-blackout> i find a lot of my rail chaos results from trying to gradually convert between rail types. i end up with duplicated paths
continue to next day ⏵
