IRC Logs for #crux Sunday, 2012-06-24

*** mike_k has joined #crux00:23
*** vaddi has joined #crux01:16
Romsterjoe9, and anyone else with cmake not finding pkg-config i've applied a bug report and a patch for it http://crux.nu/bugs/index.php?do=details&task_id=836 it also address some other optimizations02:40
juegood morning03:57
Romstermorning jue03:57
*** Rotwang has joined #crux04:00
jueRomster: your sed will not work for -j6 and is a bit too complicated, how about 's/.*-j *\([0-9]\+\).*/\1/'04:12
Romsterdamn it i thought i had it correct for both -j6 and -j 604:14
Romsterthat was the intention04:14
jueanyway, no sure if we should do such "optimizations"04:16
Romsterjue, that works as well i'll update that patch.04:16
Romstermake jobs that's al it is and use system libs04:16
Romsterall*04:16
Romsterach is using them so it should be safe. and i've also tested it here.04:16
Romsterarch*04:16
Romsterjue, can i also make a sugestion that this be placed in pkgmk.conf by default export MAKEFLAGS="-j $(/usr/bin/getconf _NPROCESSORS_ONLN)"04:19
Rotwang:C04:20
RotwangRomster: why not leave it entirely to the user?04:21
Romsteror that, jsut thought it's a simple and dynamically set to the users system04:22
Romsteror leave ti how it is, it's a suggestion only.04:22
Romsterthat's what a suggestion is Rotwang a wanted by me feature but not required...04:23
Rotwangeven if we did, we could also add -l04:24
jueI'd rather put 'export MAKEFLAGS="-j1"' in it to show the user what can be done, but not what you've suggested04:25
juebut that's only my opinion04:25
RotwangI'm with jue on this04:25
jueRomster: a wiki page with such optimizations would be a good thing04:26
Romsterjue, updated patch on that bug report.04:26
Romsterand it's entirely up to sepen if he want's to apply that or not.04:27
Romsterjue, well i'd be happy with that default so users can see it can be changed. and i can add the optimizations to a wiki page.04:28
Romsteri'll make a feature request for  'export MAKEFLAGS="-j1"' in pkgmk.conf04:29
Romsteri should add too http://crux.nu/Wiki/HowToSpeedUpBuildingPackages or is it best to make a optimizations page04:31
juebasically it fits well there, but you should revise the description to be more general than04:35
Romsteryeah it needs a bit of editing doing that now.04:35
jueand probably add some sub-chapters?04:36
Romsteryep looking at that too.04:37
Romsterit's been awhile since i've touched any wiki page.04:38
Romsterreally I just want to make it aware I'm quite happy if the user sets it them self.04:38
Rotwangwhat is the easiest method to count all (specific) user processes?04:51
RotwangI came up with something like ps -eLf | grep -c <user>04:52
Rotwangis there a better way?04:52
niklasweyay finally can i order my Raspberry PI :D04:56
Romsteredited not sure hwo the contents bit is meant to work.04:56
* niklaswe dancing on the table ^^04:56
* Rotwang is still waiting04:56
niklaswe:(04:58
niklasweI hope this raspberry Pi will works as xbmc-media "box"04:59
*** Rotwang has quit IRC05:02
*** Rotwang has joined #crux05:07
Romsterhmm that'll do for now i'll seperate the ccache from distcc and put it on it's own page later.05:09
niklaswefrinnst: there?06:04
*** jdolan has quit IRC06:30
joe9Romster: regarding the MAKEFLAGS, I read from distcc docs that -j N, where N = 2*cpu's.06:30
joe9is there something different with crux, that it should not be used?06:30
frinnstaye06:31
Rotwang2*cpu is a lot06:31
Rotwangand if you have low memory, some builds may eat it up06:31
niklaswefrinnst: how long did you wait for your raspberry p?06:34
frinnsthm, dunno06:34
frinnstgot it 3 weeks ago06:35
frinnstordered it same day it became available06:35
frinnstif you want one you should sign up straight away06:35
niklaswefrinnst: I have order my today, thats why I wondering :D06:35
Romsterjoe9, i edited that, that was before the kernel had the BKL removed, it's not needed now and it'll most likely run slower with more jobs than cores.06:42
Romsterbefore with the BKL it would iowait on some threads so having more jobs than cores allowed a compile to continue than to wait.06:43
Romsterwhich happened to be faster than waiting on a locked thread.06:43
Romsterbut now doing more jobs than cpu cores will only increase time spent compiling.06:44
joe9romster, just to get it into my thick skull, use the below even for distcc? export MAKEFLAGS="-j$(/usr/bin/getconf _NPROCESSORS_ONLN)"06:44
Romsteryou can use export MAKEFLAGS="/usr/bin/distcc -j 2> /dev/null"06:44
Romsterfor distcc06:44
Romsterwhich i've added to http://crux.nu/Wiki/Distcc06:44
Romsterjoe9, be sure to set DISTCC_HOSTS='localhost/2 foo/2 bar/2' to the / number of cpus on each node.06:47
joe9ok, thanks.06:47
Romsterthat way you wont bog any one node down with more jobs than it can handle.06:47
Romsterand that distcc -j will report the total for you.06:47
Romstergetconf _NPROCESSORS_ONLN will only report the number of online cores of the local host only.06:50
cruxbot[opt-x86_64.git/2.7]: lua: updated to 5.2.107:53
cruxbot[opt-x86_64.git/2.7]: emacs: removed unmaintained port07:54
cruxbot[opt-x86_64.git/2.7]: netpbm: removed redundant port07:59
cruxbot[opt.git/2.7]: lua: add PIC08:28
Romsterwasn't aware of that lua in opt-x86_64 builds fine with -fPIC on both 32bit and 64bit now.08:29
Romsterfrinnst, ^08:33
Rotwangany good movies?08:46
* Romster shrugs08:48
Romsteri haven't been following the new releases much.08:48
Rotwang:C08:55
cruxbot[opt.git/2.7]: fontforge: 20110222 -> 2012011909:26
rmullAnybody here have lirc all figured out?09:50
rmullHaving a hard time getting real data from my IR receiver09:51
Rotwangcan I get order in which prt-get sysup is going to update ports?10:02
Rotwangeasily10:03
Rotwangquickdiff sorts in alphabetical order10:03
Rotwang:C10:03
rmullsysup --test?10:14
*** joacim has quit IRC10:25
jaegerniklaswe: raspbmc or whatever it's called still needs some work from what I've read10:35
jaegerhaven't used it myself, though, I don't have one yet10:35
tilmanfrinnst: can 'tjat' be translated as talk/chatter/... in some context?10:41
*** nthwyatt has quit IRC10:46
niklaswejaeger: oh, okey I will try and see, how it works..10:59
tilmanniklaswe: can you help with that question?11:01
frinnsttilman: no11:08
frinnsttjat translates to nagging11:09
tilmanaha, that makes sense as well11:09
frinnst"frugan bara tjatar"11:09
tilmanthe dictionary of my choice translated it as 'deed'11:09
frinnst:)11:09
frinnsthah, what?11:09
tilman"jag är trött pa ditt tjatÃ"11:10
*** joacim has joined #crux11:10
tilman"i'm tired of your nagging"?11:10
frinnstyep11:10
tilmank, thanks11:10
*** joacim has quit IRC12:18
nogagplz"[23:45:53] <Rotwang> any good movies?" have you seen god bless america14:14
Rotwangno14:30
*** vaddi has quit IRC16:03
*** deep42thought has joined #crux16:26
deep42thoughtHi there, does anyone have some experiences with mdadm and "handmade" RAID XY's? I made a RAID5 including some RAID0s as devices and don't get it started.16:28
deep42thought... I mean, it worked when I assembled it, but after reboot I cant tell mdadm to start the raid5 (the raid0s seem to run)16:29
deep42thoughtnever mind, I got it: you need to _stop_ the incomplete array before assembling it ^^16:46
*** deep42thought has left #crux16:49
*** mike_k has quit IRC16:51
*** joacim has joined #crux17:27
*** Rotwang has quit IRC17:33
*** jdolan has joined #crux17:36
*** ChanServ sets mode: +o jdolan17:36
*** horrorStruck has joined #crux19:27
joe9i just bought a motherboard which seems to have raid-1 built in. which is better software raid or hardware raid?20:17
*** joe9 has quit IRC20:22
*** nogagplz__ has joined #crux21:25
*** mavrick61 has quit IRC21:29
*** mavrick61 has joined #crux21:30
Romsterjoe9, most likely fakeraid best to use mdadm unless it's real raid. which i haven't seen inbuilt on a motherboard before.21:38
jaegerdepends on how the metadata and data are stored, and if you will transplant those drives to a new motherboard, etc.22:02
jaegerif you think you'll change hardware, use mdadm22:02
jaegerif you don't, kinda a toss-up22:02
*** nogagplz__ has quit IRC22:13
*** deus_ex has quit IRC23:13
*** tilman has quit IRC23:46
*** tilman has joined #crux23:47
*** ChanServ sets mode: +o tilman23:47

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