IRC Logs for #crux Friday, 2014-01-03

*** dkoby has joined #crux01:58
*** Romster has quit IRC02:07
*** Romster has joined #crux02:12
*** tilman has quit IRC03:04
*** tilman has joined #crux03:06
*** Amnesia has quit IRC03:14
*** mavrick61 has quit IRC03:31
*** mavrick61 has joined #crux03:32
*** nrxtx has quit IRC05:21
*** nullmark has quit IRC05:22
*** nrxtx has joined #crux05:23
*** jue has joined #crux05:25
*** Amnesia has joined #crux05:31
*** nullmark has joined #crux05:32
*** dkoby has quit IRC05:45
*** nullmark has quit IRC07:06
*** Amnesia has quit IRC07:25
*** Amnesia has joined #crux07:25
*** Pingax has joined #crux07:27
*** darfo has quit IRC08:13
*** dkoby has joined #crux08:29
*** jaeger has joined #crux09:08
*** nullmark has joined #crux09:08
*** BitPuffin has joined #crux09:55
*** dkoby has quit IRC10:17
*** c0ck4m0u53 has joined #crux10:39
*** darfo has joined #crux11:12
*** rauz has quit IRC12:40
*** dkoby has joined #crux13:38
*** jdolan has quit IRC14:22
*** jdolan has joined #crux14:32
frinnstjaeger: are you using your openelec/rpi for netflix?16:01
*** dkoby has quit IRC16:17
*** Pingax has quit IRC16:29
nwe does netflix works for linux now? I mean doesnt it need silverlight?16:33
jaegerfrinnst: haven't tried to16:38
jaegernwe: not natively but you can run firefox/silverlight in wine or use pipelight17:49
jaegerso I doubt it would even work on arm17:49
nwejaeger: true.17:52
*** Pingax has joined #crux17:58
*** Pingax has quit IRC17:58
*** Pingax has joined #crux17:58
tilmani thought silverlight was dead18:04
*** BitPuffin has quit IRC18:11
jaegernetflix still uses it. no idea if it's still getting new features or active development18:15
frinnstyeah i thought so. there are some windows program that transcodes to xbmc.. but whats the point in that?18:22
frinnstand i found a blogpost regarding wine but yeah, that wont run on arm18:22
Amnesiafolks, what could trigger pkgmk to remove all the contents of cwd?18:37
*** doomicide has joined #crux18:42
*** c0ck4m0u53 has quit IRC18:44
Amnesiahttps://privatepaste.com/e15df3d58218:47
Amnesiait does work when I add curly braces to $name18:48
Amnesiale bug?18:48
Amnesiahttps://privatepaste.com/e2a048ad1319:00
tilmanshouldn't git to rsync export be instant on crux.nu?19:08
tilmanmmmh19:08
cruxbot[opt.git/3.0]: cryptsetup: 1.6.2 -> 1.6.319:09
Amnesiahttps://privatepaste.com/12460e6bd419:09
cruxbot[opt.git/3.0]: git: 1.8.5.1 -> 1.8.5.219:10
cruxbot[contrib.git/3.0]: gob: 2.0.19 -> 2.0.2019:10
cruxbot[contrib.git/3.0]: nbd: 3.3 -> 3.619:11
cruxbot[contrib.git/3.0]: syslog-ng: 3.4.5 -> 3.4.719:12
cruxbot[contrib.git/3.0]: vlc: 2.1.1 -> 2.1.219:15
jaegerpermissions issue?19:19
Amnesiajaeger: nope19:20
Amnesiathe exit code is 519:20
Amnesiawhich means something goes wrong when unpacking19:20
Amnesiathere isn't a syntax issue right...?19:21
jaegerdid the source file get downloaded properly? can you unpack it manually?19:21
Amnesiayep19:21
Amnesiayep to both questions19:21
jaegerI see a warning that packages should be built as root19:22
Amnesiayeah, running it as root doesn't do the job either19:23
Amnesiaand it's just a warning19:23
Amnesia warning "Packages should be built as root."19:23
jaegerwhat does 'file /usr/ports/dfu-programmer/dfu-programmer-0.6.2.tar.gz' say?19:23
jaegeryes, but if you're not building as root you'll run into permissions issues unless you're handling ports dirs with the right permissions. hence my question about permissions19:24
Amnesiajaeger: it's a tarball19:24
jaegerIf it won't build as root, though, something else is odd19:25
Amnesiayup19:25
AmnesiaI'm checking the source/strace atm19:25
jaegerYeah, I can see that... the point is more does the system handle it properly? Wondering if there's FS corruption or something like that19:25
Amnesianope19:25
Amnesiaeverything else's working fine19:26
*** doomicide has quit IRC19:26
Amnesiaother ports I made are still working fine19:27
jaegerdon't know what to tell you, pkgmk definitely thinks the file isn't there19:27
Amnesiait isn't there anynmore..19:29
Amnesiastuff gets removed for some reason:)19:29
Amnesiaeven the dir the Pkgfile resides in19:29
jaegerwhat's your PKGMK_WORK_DIR?19:30
Amnesiah wait19:30
Amnesiapebcak19:30
Amnesiaforget it-.-19:30
AmnesiaHIT MEEEE19:30
Amnesia:p19:30
jaegerheh19:30
Amnesiaffs, thank god it's weekend19:30
Amnesiasorry for the spam^^19:31
Amnesiashould thought about this myself-.-19:31
jaegerno worries19:31
Amnesialinux compatible?19:32
jaegerunfortunately not. still a great game19:42
Amnesia:)19:44
nwels21:01
*** Pingax has quit IRC21:11
tilmanteK_: does valgrind work for you? it's complaining about illegal instructions even for the most simple programs23:07
teK_yeah23:08
teK_wait..23:08
tilmanand i believe we need to s/lib/lib32/ in glibc-32's .nostrip23:08
teK_last update was two-three weeks ago, right?23:08
tilmandunno23:08
tilmani haven't valground anything in quite some time23:08
teK_Nov 6th23:08
teK_I wrote a daemon with libpcap and libasyncns.. it worked with that just fine23:09
jaegertilman: good catch, I'll fix that23:17
tilmanthanks23:18
teK_just reran valgrind.. works for me [tm]23:19
tilmanthat --enable-gdb thing is worthless btw23:20
tilmanas configure itself will tell you ;)23:20
teK_> /dev/null helps with ignoring configure :P23:20
tilmanlet's add more pointless configure arguments all over the place then \o/23:21
teK_pfff23:21
jaeger--whatisthisiamnotgoodwithcomputer23:21
jaegertilman: I'm thinking a release bump isn't necessary for this but if you want me to I'll do that23:22
tilmanteK_: does valgrind work for you on 32 bit binaries?23:24
tilmanjaeger: ^ the libc change should fix that23:24
jaegertilman: I mean a bump of release= in the Pkgfile23:24
tilmani know, but i don't know the answer23:25
tilman;d23:25
jaeger:D23:25
jaegerI'm guessing most people don't care if their glibc-32 is stripped23:25
tilmantrue23:25
jaegerI'll just tell anyone else who asks to rebuild it, you already know about it :)23:25
teK_not really.23:26
cruxbot[core.git/3.0]: glibc-32: fixed .nostrip, thanks Tilman23:26
teK_00:08 <@tilman> and i believe we need to s/lib/lib32/ in glibc-32's .nostrip23:27
teK_great conclusion from valgrind's error msg :p23:27
teK_i.e. I get that, too23:27
jaegertilman: incidentally I love "valground" :D23:27
tilmantease me more while i'm fixing your crap23:27
tilmanjaeger: =)23:27
teK_crap? That's a bit harsh buddy :p23:28
cruxbot[opt.git/3.0]: valgrind: remove enormous ./configure-bloat; thanks tilman23:28
teK_rebuilding glibc-32.23:29
tilmanyeah, it might be23:29
jaegereither of you awesome debugglers mind guessing what the hell is happening in http://jaeger.morpheus.net/misc/gnome-shell.log.xz ? :D23:30
teK_gotta code smarty, sorry :p23:30
tilmani want to debug proprietary stuff23:31
tilmanjaeger: what's the problem?23:32
tilmanhigh cpu usage?23:32
jaegeryeah, 100% of a core at all times23:33
jaegerIt runs and works, I'm just trying to figure out what it's trying to do in the background23:33
jaegerOr what's missing that it's trying to contact23:33
jaegerI figure it's something that some of the DEs have in common because one process does this in gnome3, cinnamon, AND mate23:34
jaegergnome3: gnome-shell, cinnamon: cinnamon, mate: caja23:34
teK_/c/23:34
jaegerwhat is that?23:35
tilmanjaeger: see how it calls recvmsg(6) after each poll? you should check what that fd is23:35
jaegerI've been trying to figure that out, actually :) Glad I'm not completely nuts23:36
teK_.nostrip fixed things with valgrind btw..23:36
tilmanmost likely it's a socket23:36
jaegerlrwx------ 1 jaeger users 64 Jan  3 16:07 6 -> socket:[103001]23:36
jaegerfrom /proc/<PID>/fds/23:37
tilmanlook for: socket.* = 623:37
tilmanin the strace log23:37
jaegersocket doesn't even show up in the strace log23:37
tilmanyou are attaching strace to a live process, yes?23:38
jaegeryes23:38
tilmanthat socket has been created when you started the session probably23:39
tilmanyou could mv gnome-session{,old}23:39
jaegerso I need to get a trace from the start23:39
tilmanand write a wrapper script23:39
tilmanthat runs strace -o/tmp/huge.log gnome-session.old23:39
jaegerhrmm, will try that. thanks :)23:39
tilmanwait23:39
tilmantry strace -o/tmp/huge.log -esocket,recvmsg23:40
jaegerok23:40
tilmanto only get calls to socket and recvmsg23:40
jaegernice, will try it23:40
jaegeronce my glibc-32 rebuild finishes, that is23:41
jaegershould have run that in tmux23:41
tilmani once read an article on how to move stuff into tmux after the fact23:42
tilmanpretty sure it was close to black magic23:42
tilman;d23:42
jaegerinteresting idea, I must search on that :D23:42
teK_21:26 < solardiz> RT @climagic: longcmd ; [Ctrl-Z] ; bg ; disown ; screen ; reptyr $( pidof longcmd ) # Suspend and reattach a process to screen. Thx @KirilsSolovjovs23:43
jaegerjust found disown and reptyr :)23:43
tilmandisown is great23:43
tilmannever used reptyr myself23:43
jaegernot available on my system but definitely worth looking into23:44
Amnesiatilman: reptyr can do the same for screen23:45
Amnesiadunno whether it does the job with tmux though23:45
Amnesiaoh lol23:45
jaegerhttps://github.com/nelhage/reptyr23:45
Amnesianelhage++23:46
*** BitPuffin has joined #crux23:47
jaegertilman: mind taking another look at the log? I uploaded one with the new strace to http://jaeger.morpheus.net/misc/gnome-shell-strace.log.xz23:51
jaegerI see the socket = 6 call but not to what it's attached. maybe filtering for socket and recvmsg is too draconian23:51
tilmanyes, i suck23:52
tilmantry strace -e trace=network23:52
tilmanas a replacement for the -e arg i gave ealier23:52
jaegerhrmm, maybe inotify-related23:53
jaegerwill try that23:53
jaeger-e trace=network is equivalent to -e network unless I'm reading the manpage wrong23:54
tilmani never ran neither23:56
tilmanwas just quoting from the manual too23:56
jaegeruploaded a new log to the same url23:56
jaegerfair enough :)23:56
tilmanconnect(6, {sa_family=AF_LOCAL, sun_path=@"/tmp/.X11-unix/X0"}, 20) = 023:56
tilmanso it's X23:57
jaegerhrmm, weird23:57
tilmanare you sure the frequent recvmsg calls are the problem?23:57
jaegerI guess not. Just seemed like that was the case because a test on a fedora system shows far fewer lines of log generated and no load23:58
jaegerSo I figured I was missing something on the crux system that gnome-shell wants23:58
jaegerit seems like /tmp/.X11-unix/X0 is first, then it connects /dev/log later23:59

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!