IRC Logs for #crux Monday, 2015-08-10

*** onodera has quit IRC00:01
*** kori has joined #crux00:13
*** _q00p_ has quit IRC00:18
jaegeryay, having more success with my gpu passthrough machine this time around01:40
teK___nice; VMware or qemu?01:43
teK___also: off for a little nap. it's the final countdown.. ;)01:45
jaegerqemu is nearly the only option for good gpu passthrough01:47
jaegervmware does well with others, though01:47
jaegerhave a nice nap :)01:47
*** Feksclaus has quit IRC01:52
*** ________mavric has quit IRC02:48
*** ________mavric has joined #crux02:49
*** Romster has quit IRC02:57
*** Romster has joined #crux03:04
*** tilman has joined #crux04:42
*** tilman_ has quit IRC04:45
*** druid_droid has joined #crux05:10
*** _q00p_ has joined #crux05:35
*** SiFuh has quit IRC06:52
teK___oh? Didn't know that (qemu+gpu)06:54
*** SiFuh has joined #crux07:05
*** Lukc has joined #crux07:54
*** BitPuffin|osx has quit IRC08:05
*** xeirrr has joined #crux09:35
*** xeirrr has quit IRC09:37
*** xeirrr has joined #crux09:38
*** xeirrr has quit IRC09:42
*** dougl has joined #crux10:23
*** dougl has quit IRC10:27
*** cruxbot has quit IRC10:37
*** _q00p_ has quit IRC10:48
*** Feksclaus has joined #crux11:02
*** _q00p_ has joined #crux11:32
*** Feksclaus has quit IRC12:04
*** LMNOP has quit IRC12:16
jaegerteK___: something like this: http://www.firewing1.com/howtos/fedora-20/create-gaming-virtual-machine-using-vfio-pci-passthrough-kvm12:17
jaegerthere was until very recently a really useful thread on the arch forums about it but the mods threw a fit and deleted the whole thing last week :P12:17
teK___awesome12:20
teK___will click in an hour or two12:20
teK___*hint* *hint* :D12:20
jaegerhttp://vfio.blogspot.com/ is also useful12:20
*** rauz has quit IRC12:24
*** LMNOP has joined #crux12:27
frinnstfyi: http://i.imgur.com/JejW9Um.jpg13:21
frinnstteK___: our cpu doesnt support vga passthrough13:21
frinnstsorry13:21
frinnstif you're still on 2600k13:21
*** BitPuffin has joined #crux13:28
retardmany store-bought eggs are actually fertilized13:30
retardyou can take a pack and hatch a couple of chickens13:31
retardyou can tell by looking closely at the little white string thing attached to the yolk13:31
retardif there's a black dot there, that egg was fertilized and could have been grown into a chick13:31
retardthe cooling of the egg halts the division, but if it only lasts for a couple of days the egg can survive it and hatch just fine13:33
frinnstits a bit late when you get to that stage though13:42
frinnst:>13:42
frinnstor maybe you can tape it back together again?13:42
retardthe trick is to just toss the whole packet into an incubator13:58
retard21 days later you can likely get a couple of hatchlings13:59
frinnstit sounds like you do this often14:14
frinnst:D14:14
retardwhat can i say14:17
retardi am a cat owner14:17
rmullDid you guys see that Arch's AUR is now using a git backend?14:19
rmullThat's something crux could do with ports via the git driver14:20
retardkori already made one14:20
rmullYeah, that's my point14:21
retardif it's not slated for inclusion in 3.2 it should be14:21
retardguess portdb needs some loving though14:21
rmullIt's shipped with the git port, so it's not officially endorsed for ports distribution14:21
rmullBut I vote that it should be :)14:21
retardmeans pushing git into core14:23
retardshould be fine?14:23
frinnstits very buggy though14:23
frinnstit freaks out if you do any local changes in the port tree14:23
rmullThe git driver?14:24
frinnstyes14:24
rmulloh, interesting.14:24
frinnstand it clones everything, not just --depth=1 or whatever14:24
frinnstso not very efficient14:24
rmullMaybe I should experiment more with using git natively to manage personal port changes rather than using the traditional methods14:24
retardkori: what say you?14:24
retardthe ports i maintain with a friend i do via git, we just have a hook to repgen14:25
frinnstall core/opt/contrib/xorg/whatever ports that are hosted on crux.nu use git14:25
frinnstjust rsync to clone to users14:25
retardi definitely don't think it's time to switch default method for the important core ports14:27
rmullNot right away, but it might be something to put on the roadmap14:27
retardwhat are the actual benefits14:28
rmullFaster sync14:28
retardif it clones the entire repo it will be slower sync in a right jiffy14:29
rmullYes, whole repo is slow, but daily ports -u is much faster14:29
rmullCurious about whether it is more or less costly in terms of computation time for server and clients14:30
retardbeating 10 seconds.. dunno14:30
frinnstcloning contrib takes 2 seconds14:34
retardbugs with local changes definitely needs attention first14:34
frinnstyes, it needs to be foolproof14:35
*** joacim has quit IRC16:09
*** joacim has joined #crux16:09
teK___frinnst:16:32
teK___I still do run it16:32
teK___but Im planning on passing it on to my son, getting a new one. But Im unsure if I should get me that i7 6700K beast16:33
teK___actually, as I'm done doing some work, I planned on getting it this month so no time to wait for cheaper i7s16:33
jaegerI'd recommend sticking with haswell refresh currently and just get the i7-4970k16:33
teK___as an alternative I was thinking about a big haswell-e16:34
teK___I'd have to stick with sandy bridge..16:34
teK___I think it was the i7-4970K, yeah.. still unsure. I gotta buy this month because of tax reasons and I want a new toy for GTA V, I was given for birtday wayyy to long ago ;)16:35
jaegeris gaming the main thing you want to do on it or would you need heavy lifting for scientific analysis or something?16:45
teK___I need cruxentific compilation and more regular gaming16:49
jaegerteK___: I put this list together for a friend a couple days ago: http://pcpartpicker.com/p/P8qkt616:49
teK___I got the VGAs for the latter16:49
jaegerfor gaming and VMs, etc.16:49
jaegerIt's actually very close to what I have, too16:50
teK___so a shame I cannot use th builtin GPU of the 6700K. Still I wonder if I can do th bumblebee with linux already16:50
teK___what? The Haswell?16:50
jaegeryes16:50
jaegerwhy can't you use the onboard gpu on the 6700k?16:50
teK___I can but I want the 970 GTX also16:51
teK___>  Still I wonder if I can do th bumblebee with linux already16:51
jaegerI don't think that works with non-optimus hardware but I could be wrong16:51
teK___how unfortunate :)16:51
teK___it's really a pitty16:51
teK___on the other hand, I don't know the piwer consumption per VGA "mode" ..16:52
jaegerpcpartpicker will estimate that for you if you put in parts16:52
teK___nice, didnt know that one. Thanks16:53
jaegernp, it's a neat feature16:54
jaegermaybe relevant: http://forums.anandtech.com/showpost.php?p=37617058&postcount=3255 and http://forums.anandtech.com/showpost.php?p=37617084&postcount=325616:55
jaegersome benchmarks of the skylake i5/i716:55
teK___will check in a heart beat16:55
teK___:)(16:56
jaegerthe first reviews seemed a bit lackluster, these benchmarks make it look a little better16:56
jaegerthough if the cost isn't pretty close to haswell I'd probably go with haswell16:56
teK___I saw some marks mentioning a 12% gain or so16:56
teK___what sucks are the expensive boards and RAM16:57
teK___usually I go with <<100€ MSI ones16:57
jaegeryeah, and that there aren't a ton of 170 boards out yet16:57
teK___and then? .. you have to buy the most expensive ones to get ONE - 1 -  3.1 Type C USB connector16:58
teK___what a fiasco16:58
jaegeranother reason I might stick with haswell, there's a lot of new stuff for skylake :)17:03
jaegerearly adopter tax, heh17:03
jaegeron the other side of the coin, if you need nvme SSDs you need skylake17:03
retardnew hardware is spooky17:04
jaegertechnically you can use them with haswell but it depends on the motherboard manufacturer to actually make a motherboard with it (and boot support)17:04
retardv-tech and amt17:04
retardscariest shit i know17:04
koriwe copied the rsync driver behavior to the extent of our git knowledge17:05
koriif that makes any sense17:05
retardkori: why does it misbehave when local changes are made?17:06
kori┐('~'; )┌17:06
koriit's supposed to pull and reset hard17:07
koriaka, no local changes are supposed to remain if they're tracked17:07
teK___where has cruxbot gone :)17:57
korihe ded17:57
*** cruxbot has joined #crux17:59
cruxbot[contrib.git/3.1]: vala: 0.27.2 -> 0.29.218:00
cruxbot[contrib.git/3.1]: vlc: 2.1.5 -> 2.2.118:00
cruxbot[contrib.git/3.1]: spice-gtk: 0.28 -> 0.2918:00
cruxbot[contrib.git/3.1]: p5-net-dns: 0.83 -> 1.0118:00
cruxbot[contrib.git/3.1]: spamassassin: 3.4.0 -> 3.4.118:00
cruxbot[contrib.git/3.1]: cacti: 0.8.8c -> 0.8.8f18:00
cruxbot[contrib.git/3.1]: celt51: 0.5.1.3 -> 0.5.218:00
cruxbot[contrib.git/3.1]: scribus: 1.4.4 -> 1.4.518:00
cruxbot[contrib.git/3.1]: keychain: 2.7.1 -> 2.8.118:00
cruxbot[contrib.git/3.1]: getmail: 4.47.0 -> 4.48.018:00
cruxbot[contrib.git/3.1]: iw: 4.0 -> 4.118:00
cruxbot[contrib.git/3.1]: glade3: 3.8.4 -> 3.8.518:00
rmullThe setcap portion of the opt/wireshark post-install is failing for me: Note <filename> must be a regular (non-symlink) file.18:27
rmullAnyone else?18:27
rmullmaybe I don't have kernel support for file caps, I will check18:28
rmullNot seeing anything related to capabilities. I don't think it's related to xattrs18:30
rmulloh, maybe it is xattrs. I will set those up.18:30
rmull(eventually)18:33
*** BitPuffin has quit IRC18:47
*** timcowchip has joined #crux19:14
*** kori has quit IRC19:18
timcowchipdoes anyone have a package at https://aur4.archlinux.org/ ?19:24
timcowchipI was just wondering why anyone thinks it the way crux should go :)19:30
z3braisn't it just a port tree managed via git?19:36
z3braso to sync ports you just run "git pull"19:36
timcowchipby "sync ports" you mean upload?19:45
z3bradownload19:46
z3brathough "uploading" would be as simple as "git push"19:46
z3brain fact, that's how alpine works19:46
z3bra(and void linux iirc)19:47
z3brathough both provide binary packages19:47
timcowchipthe aur provide a snapshot to download19:48
timcowchipof a pkgbuild19:48
timcowchipnot a binary19:48
z3brayeah binaries are totally different sthings19:49
timcowchipI guess a wrapper for prt-get like yaourt might be a way to sink crux user ports19:56
timcowchipwithout adding the entire user repo to /usr/ports19:56
z3brawhy a wrapper for prt-get ?20:00
teK___/1/120:00
z3brajust create a new repo, and add it to your prt-get config20:01
teK___rmull: I will try to look into it :)20:01
timcowchipI have the gparted port from Romster's repo installed20:01
timcowchipbut I don't have Romster's entire repo in /usr/ports20:02
timcowchipnor will I20:02
timcowchipI don't think there's enough room on my drive20:02
timcowchipmaybe something like http://slackbuilds.org/repository/14.1/system/sbotools/ would work20:05
*** maldoror has joined #crux20:21
*** BitPuffin|osx has joined #crux20:33
*** quasinoxen has quit IRC20:40
*** onodera has joined #crux20:46
*** quasinoxen has joined #crux20:51
*** maldoror has quit IRC20:55
*** joacim has quit IRC21:18
*** joacim has joined #crux21:18
frinnsttimcowchip: you can use opt/mpup to "subscribe" to single ports from a collection21:20
frinnstif thats what you were requesting21:23
*** crash_ has quit IRC21:30
teK___rmull21:36
rmullteK___: Yo yo21:36
*** quasinoxen has quit IRC21:37
teK___did you try to enable capabilities and re-run post-install?21:38
rmullNo, can't reboot until later tonight21:38
teK___I think you are right and it is the problem21:38
teK___so atm they are not enabled?21:39
rmullextended attributes / xattrs are not - I didn't see anything specifically concerning capabilities in the kernel, so I'm assuming they are equivalent21:39
rmullSo my plan is to enable xattrs for my root filesystem21:39
rmullwhich are currently disabled21:39
teK___the only sound route would be to check if the setcap failed and reset file permissions for the binary21:41
teK___which is ugly..21:41
frinnsthttps://www.sunfrog.com/SYSADMIN--HEROES.html?2101121:41
frinnstIm fucking getting one of those21:41
teK___sounds like a great idea21:42
teK___could be wearing that on my first day at my new job :D21:42
rmullteK___: fwiw getcap returns Failed to get capabilities of file `/usr/bin/dumpcap' (Operation not supported)21:42
teK___yeah, sounds reasonable21:43
rmullI'll let you know, thanks for the thoughts21:43
teK___will use that, thanks21:43
rmulloh, I see you're concerned about the state of the file after the failure I'm experiencing - I hadn't even thought about that21:43
*** crash_ has joined #crux21:44
teK___mind trying? http://dpaste.com/2P2PS9C21:45
teK___it should bitch and then exit with 0 (for you)21:45
rmullyep, one sec21:46
teK___it does not. duh.21:47
teK___jsut a sec :)21:47
rmullNo, doesn't seem to work. Maybe because getcap doesn't return an error when it fails21:47
frinnstteK___: btw there were a cve request for wireshark today21:47
teK___no update yet. But happy to be rebuilding again21:48
teK___rmull: please check by running getcap /bin/ls; echo $?21:49
rmullIt's 021:50
rmullwheras setcap returns 1 on failure21:50
rmullso that's annoying21:50
teK___it is21:51
teK___but we should be safe using setcap for checks, too21:51
rmulluse -v?21:51
teK___then what?21:51
rmullWell, it doesn't change anything, and it returns 1 (fail) for me21:52
teK___wtf21:52
rmullDoes it return 1 if the match fails for you, or 0?21:52
rmullIf it returns 0 but prints the mismatch, then we can use it as a check21:53
teK___does not work on Gentoo21:53
teK___i.e. it always returns 021:53
teK___I am not checkking fo certain output of the tool21:53
teK___then I could switch to Windows, too21:53
rmullno, that's good, right? If it returns 0 (success) it means capabilities are supported21:54
rmullwhereas for me, I get a 1 (fail) meaning they aren't supported21:54
rmullas long as the return value doesn't depend on whether or not the verification matches21:54
teK___http://dpaste.com/2M75KMK on gentoo21:54
rmull-v is for setcap21:55
rmullfor verificaiton that the caps match the args21:55
rmullsorry, I wasn't clear earlier21:55
rmullsee man setcap for the details21:56
rmullbrb a few minutes21:56
teK___ok21:56
rmullActually I'm gonna head home22:02
rmullLet me know if you need any other tests done and I'll get to them later on22:02
teK___kk22:03
*** dougl has joined #crux22:09
teK___rmull: please test http://dpaste.com/2XZE8KA22:24
*** Lukc has quit IRC22:48
*** Lukc has joined #crux22:49
*** onodera has quit IRC22:54
rmullteK___: That properly errors out on my system23:21
rmulla successful failure, if you will :D23:22
*** timcowchip has quit IRC23:45
*** kori has joined #crux23:55
*** _q00p_ has quit IRC23:58

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