IRC Logs for #crux Saturday, 2014-06-14

jaeger:)00:07
*** BitPuffin has quit IRC00:15
*** sammi` has joined #crux00:23
*** dkoby has joined #crux01:14
cruxbot[contrib.git/3.0]: docker: 0.11.1 -> 1.0.001:29
*** dkoby has quit IRC01:47
fbtwow01:59
fbtrelease!01:59
jaegerlies!02:04
abyxcosGot me a little excited.02:05
fbtdocker release.02:09
fbtNot cru02:09
fbtx02:09
*** mavrick61 has quit IRC02:30
*** mavrick61 has joined #crux02:31
*** felixxthecat has quit IRC02:58
*** vlnx has quit IRC03:25
*** vlnx has joined #crux03:49
prologichehe :)04:15
prologicenjoy04:15
prologicany problems ping me04:15
prologicit's tricky to configure your kernel correctly04:15
prologicbut use the test-config.sh tool04:15
fbtI already have a kernel built for docker (not specifically, but including)04:41
*** sh4rm4 has joined #crux04:45
*** BitPuffin has joined #crux05:51
Romsteri have 5 machines on crux 3.1 wonder if i should docker them all.06:53
Romstercan i like move docker images between hosts?06:53
prologicyes06:54
Romsterif one box dies for example start on another machine. or even migrate off a machine live then shot down a box for maintenance.06:54
Romstergiven i'll move to shared storage between all my boxes eventually.06:55
prologicit's all possible06:58
*** phant0mas has joined #crux06:59
Romsterprologic, have you tried musli and clang? i got a keen interest to try them07:23
BitPuffinRomster: do it07:27
BitPuffinRomster: pretty sure you can move the container, not just the image07:27
prologicmusli?07:30
prologicYeah you can07:30
prologicmigration is not a problem07:30
prologicalthough I know of no builtin docker api(s)/tools yet to do this07:31
prologicbut there are a few hacks using things like tmux and what not07:31
prologicit'll come :)07:31
BitPuffinyeah that's probably on the way07:33
BitPuffinyou could just script a sequence of docker commands though07:33
BitPuffinno big deal07:33
prologic*nods*07:37
nrxtxwhat is musli?07:39
nrxtxprologic: can you send me that link to your port db again?07:42
prologicwe haven't done much yet07:43
prologicnothing really to show :)07:43
prologicwe just have ideas mostly07:43
prologicbut07:43
prologichttps://hg.cruxlinux.org/porthub07:44
RomsterThis Connection is Untrusted07:52
Romsterself signed cert?07:52
*** doomicide has joined #crux07:52
Romsterhmm why cant we gpg sign a cert for trust07:53
BitPuffinthat would be much better07:54
nrxtxhi Romster07:54
BitPuffinthe root cert bullshit is too exploitable07:54
Romsterwhat does porthub do exactly or will do?07:54
Romsterhi nrxtx07:54
RomsterBitPuffin, agreed it's never secure.07:55
BitPuffinRomster: it will give you a web interface for making ports07:55
BitPuffinthat will do test builds07:55
BitPuffinand packages for pkg-get07:55
BitPuffinand I guess easy forking etc of ports07:55
BitPuffinand merge requests etc07:55
BitPuffinbasically easier collaboration for port maintainers07:56
BitPuffin+ builds07:56
nrxtxRomster: what did you mean with musli?07:56
BitPuffinmaybe he's hungry07:56
BitPuffingive the guy a break07:56
Romsterswitch from glibc to musli07:57
nrxtxBitPuffin: something like aur + editing?07:57
nrxtxRomster: it will break a lot :D07:58
Romsternah what could possibly go wrong -_- :D07:58
nrxtxit does not have the glibc legacy bugs which will break a lot of tools08:00
nrxtxjust check sabotages patches to get some basic stuff working08:00
Romsterglibc is meant to be following standards.08:00
nrxtx"meant" :D08:00
nrxtxhttps://github.com/sabotage-linux/sabotage/tree/master/KEEP08:01
prologicaur + editing08:01
prologicyes sort of08:01
prologicaur + github if you will08:01
nrxtxRomster: if you want to try musl i have a port in my repo, which allows you to install musl in parallel to glibc, "musl-gcc" can then be used to compile with musl-libc08:02
Romsterah neat nrxtx08:04
BitPuffinprologic: yeah, except the AUR and github doesn't do continuous integration08:04
Romsterneed musl-clang too <<08:04
BitPuffinRomster: link to musli?08:04
Romstermusl sorry i added the i08:05
nrxtxRomster: since clang can be used as gcc drop in replacement it might simply work when copying musl-gcc and replace the command08:05
Romsteri would think so.08:05
BitPuffinhttp://www.musl-libc.org/08:05
nrxtxmusl-gcc is a shell script calling gcc omitting glibc and adding musl-libc08:05
Romsternot doing it on my desktop but i have a second desktop pc i can try it on08:06
prologicnot bug aur + gitub + travisci08:06
prologic:)08:06
BitPuffinprologic: plus ppa08:06
BitPuffinor something08:06
BitPuffinlol08:06
BitPuffinXD08:06
BitPuffinRomster: you could try it in a docker :P08:06
nrxtxRomster: in addition musl is on a nice way and it is worth testing/using it but keep in mind not all syscalls are implemented08:07
BitPuffingreat first impression when the have a CVE on their fucking front page08:07
BitPuffinhowever in a shallow way it is a bad impression, but it's good that they are upfront about it08:07
prologicsorry what is musl?08:08
BitPuffinit's just a libc08:08
prologicoh08:08
Romsteractually am hungry but i was making a coffee.08:08
nrxtxhehe08:08
nrxtxwhat currently is keeping me away is the complete lack of scheduler/priority support for processes08:09
nrxtxthey will just end up in nothing08:09
nrxtxhttp://git.musl-libc.org/cgit/musl/tree/src/process/posix_spawnattr_sched.c08:09
BitPuffinnice08:10
BitPuffinMIT license08:10
BitPuffinnrxtx: is the goal to be portable?08:11
Romsternrxtx, which is yoru repo again on portsdb? i only see kris and teatime08:11
nrxtxmine is "kris"08:11
Romsterah08:12
Romsterwell i know it's beng worked on and probably is missing stuff still08:12
nrxtxBitPuffin: i guess but i think currently they are working on implementing all the basic stuff08:12
Romsterthey might add in some define for some programs that depend on glibc bugs who knows?08:13
Romsteror bugger glibc bugs stick to the standards.08:13
nrxtxRomster: no they wont08:13
Romsterforce programs to either die or fix thee bad behavior08:13
nrxtxthey say the programs have to be fixed upstream08:13
nrxtxthey will not implement any legacy layers or glibc bugs programs are depending on08:14
Romsterglibc is really full of legacy code08:14
nrxtxi'm also really interested in the musl + clang combination08:15
Romstergcc is so bloated08:15
Romsterand there plugin system ugh08:15
Romsterllvm compiles really fast when it has access to 14 cores :D08:17
nrxtx:D08:17
*** Rotwang has joined #crux08:27
*** Rotwang has quit IRC08:46
*** pidsley has quit IRC08:52
*** Rotwang has joined #crux08:59
Romstercompiling core with clang over distcc :D08:59
nrxtxplease tell me if it worked :)09:02
Romsteryou using any optimizer flags?09:02
nrxtxonly tried clang with some ports without special flags09:10
Romsterjust keeps to using -O2 -pipe -march=amdfam1009:11
Romsternot certain if clang is respecting MAKEFLAGS09:17
Romsterbut that's for make itself...09:17
Romsterah it is jsut alot of ports are just too small to take advantage of multiple make jobs09:18
nrxtxdamn it nettle broke gnutls09:19
Romsterah09:20
Romsterthat's why looop couldn't.09:20
nrxtxthey changed a struct in the headers of nettle09:21
Romsterwell 2.7.1 -> 3.009:22
Romsterusually such breakage happens when the major number is changed.09:22
nrxtxupdating ports using gnutls is now a problem :D09:24
Romstergreat...09:24
Romsterwhen people upgrade stuff they should see if it breaks anything that depends on them....09:25
Romsterespecially major stuff like gnutls09:25
nrxtxcan someone verify gnutls 3.2.15 not building with nettle 3.0?09:26
tilmansounds like a trap ;)09:27
nrxtxjust be careful :D09:27
tilmanpkgexport ftw :)09:28
Romstertesting09:30
nrxtxtilman: not in this case09:30
nrxtxi guess, since it is a struct changed and not the exports09:30
tilmani mean:09:31
tilmanmake a backup of the "good"/old nettle before upgrading09:31
nrxtxhehe that's how i rescued myself :D09:32
Romsteralways09:33
Romsterbut it'll break depinst for others09:33
RomsterBuilding '/var/ports/packages/gnutls#3.2.15-1.pkg.tar.gz' succeeded09:40
Romsternot sure what you did nrxtx09:41
Romsterah for gods sake seppen didn't push nettle to 3.1 branch...09:41
nrxtxso you are still on nettle 2.7?09:43
Romsteryeah on 3.1 branch i compiled. but i'll do a 3.0 branch compile now.09:43
Romsteri don't fancy pushing breakage to 3.1 branch09:43
tilmanyes09:51
*** dkoby has joined #crux10:04
Romsteryou can't say much tilman -_-10:05
Romstercontributed in the past so now you have the right to say witty remarks so it seems.10:07
Romsternrxtx, be awhile i ran a sysup on that chroot, but i did see gentoo is reling on pre 3.0 nettle10:08
Romstermay need to downgrade nettle or provide nettle210:09
Romsteronly looks like gnutls depends on nettle?10:10
Romsterprt-get dependent nettle --all10:10
Romsternettle-3210:10
Romstergnutls10:10
Romsteri'm sure there are more ports10:10
nrxtxhttps://git.lysator.liu.se/nettle/nettle/commit/cb83bd331e5774027bcab965ccf9bcbe798ad167#diff-910:11
Romsterwonder if upstream gnutls has a patch10:12
nrxtxuhm no10:16
nrxtxat least not in master branch10:16
Romsterthough crux poicy is to downgrade nettle than to patch gnutls10:16
Romsterone of us needs to file a bug to sepen10:17
nrxtxi'll do, but i think its urgent so if you see him on irc maybe you can give him a hint if i'm not available?10:17
nrxtx#102310:23
*** markuman has quit IRC10:53
*** markuman has joined #crux10:54
Romstererror: macro "GCM_SET_KEY" passed 5 arguments, but takes just 411:02
Romsternrxtx, http://pastebin.com/5sPz8T7h with clang. so need to look though the breakage.11:05
nrxtxRomster: got the prtverify output in my portdb11:10
nrxtxonly need to add it into the frontend11:11
nrxtxnice excepted more packages having problems with clang11:12
*** SiFuh has joined #crux11:14
*** Rotwang has quit IRC11:14
*** BitPuffin has quit IRC11:35
*** BitPuffin has joined #crux11:48
*** Romster has quit IRC11:52
*** Romster has joined #crux11:52
nrxtxuh that killed my update history12:00
*** phant0mas has quit IRC12:21
nrxtxRomster: done with the prtverify part12:28
nrxtx25% of all ports got output by prtverify :D12:29
nrxtxthx for the hint with it12:31
Romstercool12:36
nrxtxhehe i should fix my ports :D12:37
nrxtxlot of "junk" files12:37
Romsternot looked though mine in awhile12:41
nrxtxi think the worst is ogmrip ;)12:43
Romsterthis is awesome goes though ports.12:44
nrxtx?12:47
Romsterprtverify i don't have to manaully do it.12:47
Romsterand i get a icon saying what port needs looking at.12:47
Romstergong though my romster repo12:47
nrxtxah :)12:48
nrxtxfor your repo i need to add list.js or pagination :D12:49
Romsterlol12:49
nrxtxadded a table hover effect, might be handier when going through a log list12:54
*** sh4rm4 has quit IRC12:55
*** sammi` has quit IRC13:08
*** sammi` has joined #crux13:09
*** sh4rm4 has joined #crux13:26
*** lnds has joined #crux14:03
*** phant0mas has joined #crux14:26
tornainbowhey has anyone had an issue upgrading gnutls?14:31
tornainbownvm, read through the backlog :)14:33
cruxbot[contrib.git/3.0]: tor: 0.2.4.20 -> 0.2.4.2214:39
cruxbot[contrib.git/3.0]: asciidoc: 8.6.8 -> 8.6.914:39
*** rexich has joined #crux14:49
*** dkoby has quit IRC14:59
*** dkoby has joined #crux15:02
Romstertornainbow, yeah bug downgrade nettle15:07
*** Rotwang has joined #crux15:10
nrxtxthat was my last update for my portdb today :)15:25
*** dkoby has quit IRC15:35
*** sammi` has quit IRC16:08
*** horrorSt1uck has joined #crux16:21
*** horrorStruck has quit IRC16:21
cruxbot[core.git/3.0]: gcc: update to 4.7.416:25
cruxbot[core.git/3.0]: libusb: update to 1.0.1916:25
*** BitPuffin has quit IRC16:32
*** phant0mas has quit IRC16:35
*** rexich has quit IRC16:47
*** linXea has quit IRC16:57
*** Amnesia has quit IRC16:57
*** Amnesia has joined #crux17:08
*** linXea has joined #crux17:09
*** sh4rm4 has quit IRC17:20
*** phant0mas has joined #crux17:22
*** sh4rm4 has joined #crux17:23
*** sh4rm4 has quit IRC17:40
*** sh4rm4 has joined #crux17:40
cruxbot[xorg.git/3.0]: xorg-xcb-util-renderutil: update to 0.3.917:47
*** BitPuffin has joined #crux18:14
*** phant0mas has quit IRC18:16
*** Tezu has joined #crux18:22
*** darfo has quit IRC18:27
*** doomicide has quit IRC19:03
*** phant0mas has joined #crux19:31
*** Tezu has quit IRC20:04
*** darfo has joined #crux20:10
*** phant0mas has quit IRC20:13
pedjagcc update fails ('no space left on device').build dir is tmpfs, 2gb.what magic incantation am I missing to make it use swap?20:21
*** lnds has quit IRC20:25
frinnstincrease the size of the ramdisk?21:06
frinnst2gb isnt that much these days - firefox require more than 821:07
pedjafrinnst: that is why I happily use firefox-bin :)21:10
pedjaafter increase to 2.5gb, it still fails, with error 1.wth21:11
pedjastage2 bootstrap21:11
frinnsthow much ram do you have?21:11
pedja4gb21:11
frinnstand how much swap?21:12
pedja2gb21:12
frinnsti really doubt its enough21:12
frinnstand it certanly wont be any faster than just setting $PKGMK_WORK_DIR to you disk21:13
jaegerIt will link in 4GB but it won't *build* in that much21:13
jaegerit takes at least 7.7GB to build in a tmpfs21:13
pedjadamn21:13
jaegerI just ran into this very recently. Building it entirely in tmpfs failed for me on a machine with 16GB RAM (7.9GB tmpfs) because something else was using it as well21:14
jaegerI rebuilt it with nothing else happening at the time and it grew to 7.7GB before finishing21:15
jaegerI had 212MB free21:15
pedjathe joys of using source based distro :)21:16
jaegerindeed21:16
pedjaI think I'll ditch the whole build as tmpfs idea for now :)21:24
pedjaeven if ddr2 was not that damn expensive, 16gb is m/b limit, 4 slots.21:28
frinnstyeah upgrading old hardware is pointless21:32
frinnstyou can probably find cheap amd based motherboards, and amd cpus are not that pricey21:33
frinnstwill still give you a shitload more performance than you have now21:34
pedjait will have to do, for now, money is tight, with tendency to get even tighter :)21:35
pedjabut this channel is hardly the place for that story.apologies.21:39
pedjaone of these days I really should start playing with docker, instead of just building it from git snapshot...21:41
pedjapostgres, docker,qgis, the list of my interests is endless21:43
pedjaI just love when FF crashes at random21:45
pedjajue: json-c is at 0.12 now.21:46
prologicprt-get depinst docker21:55
prologic:)21:55
pedjaprologic: living on the wild side, build from git master :)21:56
pedjaprt-get lock nettle :)21:58
pedjaafter downgrade to 2.7.1, of course21:59
prologiclet me update my docker-git port then22:00
prologicbut you'll have to go get my ports httpup or hgup :)22:00
pedjaI made my own, but thanks22:00
prologicfine then22:00
prologic:)22:00
pedjas/made/stolen and modified yours/ :)22:01
prologicI assume you just updated the sha hash :)22:02
*** tilman has quit IRC22:04
*** tilman has joined #crux22:05
pedjaprologic: changed version to 1.0.0-dev,install bash completion and check-config.sh from tarball exported from git.22:06
*** sh4rm4 has quit IRC22:14
pedjaand gcc builds.22:15
pedja\me rejoices22:16
pedjai suck at this irc thingie22:16
abyxcosIt's a /command, not an escape :D22:20
pedjayesss :)22:21
pedjareflex from psql22:22
*** Rotwang has quit IRC23:21

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