2014-08-01 06:13:00 <ncopa> morning
2014-08-01 06:13:13 <ncopa> https://gcc.gnu.org/ml/gcc-patches/2014-07/msg02231.html
2014-08-01 06:13:55 <ncopa> fabled: seems like gentoo hardened is trying upstream the default pie support
2014-08-01 06:14:01 <ncopa> in gcc
2014-08-01 06:14:39 <fabled> that's good news
2014-08-01 06:15:04 <ncopa> yes
2014-08-01 06:18:50 <fabled> i'm test building new musl
2014-08-01 06:18:55 <fabled> the local stuff might give some issues
2014-08-01 06:33:03 <shafire> hi
2014-08-01 06:33:21 <shafire> will 3.1 use clang/llvm instead of gcc?
2014-08-01 06:33:33 <ncopa> no
2014-08-01 06:33:48 <fabled> we want musl stabilized first
2014-08-01 06:34:00 <fabled> and not sure if clang is fully up to it, but it's getting mature
2014-08-01 06:34:05 <fabled> so we might consider that later on
2014-08-01 06:39:26 <shafire> at least mac os x/xcode/iphone uses only llvm
2014-08-01 06:41:09 <ncopa> shafire: afaik, we still need gcc, even if we build with clang
2014-08-01 06:41:17 <ncopa> so it would just add complexity
2014-08-01 06:41:36 <ncopa> second, kernel is not yet ready for full clang support
2014-08-01 06:42:23 <ncopa> third, i dont know if clang can do our hardened toolchain by default (enable ssp, pie, z,relro)
2014-08-01 06:42:34 <ncopa> so that needs to be investigated
2014-08-01 06:42:36 <ncopa> so
2014-08-01 06:42:40 <shafire> OK
2014-08-01 06:42:44 <ncopa> cost/benefit
2014-08-01 06:43:12 <ncopa> if you could give me some good arguments that the benefits will be bigger than the cost
2014-08-01 06:43:18 <ncopa> then sure, we might consider it
2014-08-01 06:43:26 <ncopa> but it is a huge project
2014-08-01 06:46:42 <shafire> maybe http://lists.freebsd.org/pipermail/freebsd-questions/2012-June/242709.html
2014-08-01 06:47:56 <ncopa> those are good points
2014-08-01 06:48:04 <ncopa> valid points
2014-08-01 06:48:32 <ncopa> linux thorvalds had a rage attack on gcc recently
2014-08-01 06:48:37 <ncopa> liuns*
2014-08-01 06:48:40 <ncopa> linus*
2014-08-01 06:48:49 <ncopa> for gcc generating bad code
2014-08-01 06:49:12 <ncopa> but please remember that clang/llvm has its own set of issues
2014-08-01 06:49:47 <ncopa> i think at some point we might use clang as the default compiler
2014-08-01 06:49:51 <ncopa> but i dont think we are there yet
2014-08-01 06:49:56 <shafire> looks like kernel 3.16 will support clang
2014-08-01 06:53:23 <shafire> here are the costs, which we will have "maybe" by switching: http://clang.debian.net/
2014-08-01 06:54:22 <barthalion> morning
2014-08-01 06:54:39 <barthalion> ncopa: any idea why does it happen? http://bld1.alpinelinux.org/buildlogs/build-edge-x86/testing/spacefm/spacefm-0.9.4-r1.log
2014-08-01 07:12:14 <ncopa> yeah i think its parallell issue
2014-08-01 07:12:30 <ncopa> there are likely multiple rules that creates the direcotry
2014-08-01 07:12:34 <ncopa> mkdir $somedir
2014-08-01 07:12:42 <ncopa> instead of mkdir -p $somedir
2014-08-01 07:12:58 <ncopa> so when $somedir already exists and you have a race condition
2014-08-01 07:13:10 <ncopa> the mkdir will return failure
2014-08-01 07:13:26 <shafire> maybe this could you guys help more in finding the errors? http://clang.llvm.org/diagnostics.html
2014-08-01 07:13:26 <ncopa> i think you can work around it by doing make -j1 install ...
2014-08-01 07:13:40 <barthalion> thanks, I'll take a look
2014-08-01 07:13:56 <ncopa> shafire: that is a good thing to use for developers yes
2014-08-01 07:14:23 <ncopa> sorry about the libsodium ABI breakage btw...
2014-08-01 07:15:34 <barthalion> don't worry
2014-08-01 07:15:35 <ncopa> barthalion: this solved it on my dev box: http://sprunge.us/TUTC
2014-08-01 07:15:51 <barthalion> although I fixed only testing/tox, no idea if anything more links to it
2014-08-01 07:16:01 <ncopa> i'll chck it now
2014-08-01 07:16:34 <ncopa> testing/toxic
2014-08-01 07:17:11 <barthalion> toxic uses tox, I don't think it makes use of libsodium directly
2014-08-01 07:17:37 <ncopa> i dont have it on x86_64
2014-08-01 07:17:44 <ncopa> apk info -R toxic
2014-08-01 07:17:46 <ncopa> should tell
2014-08-01 07:18:16 <uggedal> in progress OpenBSD GSOC work to implement systemd shim: https://uglyman.kremlin.cc/gitweb/gitweb.cgi?p=systemd-utl.git;a=summary
2014-08-01 07:18:41 <barthalion> apk info -R toxic | grep sodium gives empty output
2014-08-01 07:18:52 <ncopa> here too, on x86_64
2014-08-01 07:19:01 <uggedal> "Provide bsd-licenced, os-agnostic, dbus-api compatible systemd-{hostnamed,timedated,localed,logind} replacements"
2014-08-01 07:19:05 <barthalion> so that's all
2014-08-01 07:19:14 <ncopa> uggedal: thats nice!
2014-08-01 08:07:14 <shafire> is the order in /etc/apk/world important?
2014-08-01 08:09:37 <ncopa> shafire: no
2014-08-01 08:49:20 <uggedal> ncopa: is 3.14.15 coming to edge today? (i need to reboot anyways and don't wan't to do it twice :)
2014-08-01 08:50:50 <ncopa> i didnt plan for it
2014-08-01 08:50:52 <ncopa> but i can
2014-08-01 08:50:56 <ncopa> actually, its a good idea
2014-08-01 08:51:41 <ncopa> heh.. you scared me :)
2014-08-01 08:51:49 <barthalion> :D
2014-08-01 08:51:49 <ncopa> i thought it was me who pushed it :)
2014-08-01 08:52:00 <ncopa> uggedal: i will build new kernel asap
2014-08-01 08:52:06 <ncopa> i'll try get it out for edge before lunch
2014-08-01 08:52:16 <uggedal> ncopa: thanks, no need to rush :)
2014-08-01 08:54:40 <barthalion> ah, right, I need to run autoreconf
2014-08-01 08:55:55 <barthalion> success
2014-08-01 08:55:57 <barthalion> \o/
2014-08-01 09:21:40 <shafire> do we have a package which fills the entropy?
2014-08-01 09:24:46 <uggedal> shafire: haveged
2014-08-01 09:43:32 <fabled> i'm pushing out gcc-4.8.3 to edge soon
2014-08-01 09:47:31 <shafire> bcache has udev rules, alternative is to run "echo /dev/sd(a|b|c) > /sys/fs/bcache/register_quiet". where is the best place to run this in alpine?
2014-08-01 09:48:44 <shafire> Can I use mdev for that?
2014-08-01 09:50:15 <ncopa> shafire: yes, you can have mdev execute script when sd[abc] is plugged if you want
2014-08-01 09:52:48 <shafire> do you know how? I am looking at mdev doc at moment, maybe I find it soon
2014-08-01 10:15:16 <fabled> huh
2014-08-01 10:15:31 <fabled> libstdc++.so seems to have had a soname change
2014-08-01 10:16:22 <fabled> mmm
2014-08-01 10:16:22 <fabled> no
2014-08-01 10:16:45 <shafire> [-]sda[abc]  root:disk 0660 @(echo /dev/$MDEV > /sys/fs/bcache/register_quiet) <- like this?
2014-08-01 10:21:49 <fabled> hum
2014-08-01 10:21:53 <fabled> java-gcj-compat is not rebuilt
2014-08-01 10:21:57 <fabled> that's what is wrong on my box
2014-08-01 10:24:56 <fabled> now itis
2014-08-01 10:24:58 <fabled> ok
2014-08-01 10:53:25 <uggedal> ncopa: would you mind to: `echo section=ncopa > /var/git/posixtz.git/cgitrc`
2014-08-01 10:53:56 <uggedal> the sectionless repo on git.a.o was bugging me ;)
2014-08-01 10:54:28 <ncopa> i think Alpine Projects is better
2014-08-01 10:54:29 <ncopa> but sure
2014-08-01 10:54:45 <uggedal> heh, might have a case of ocd
2014-08-01 10:55:14 <shafire> shit, busybox mdev runs before the modules are loaded :-(
2014-08-01 10:55:35 <shafire> so this will not work, or?
2014-08-01 10:56:00 <uggedal> anyone played with using nldev (http://git.r-36.net/nldev/) btw?
2014-08-01 10:57:19 <ncopa> ah there it is
2014-08-01 10:57:29 <ncopa> i think s6 has something similar
2014-08-01 10:57:35 <ncopa> no i havent played with it
2014-08-01 11:08:10 <ncopa> uggedal: the only thing that thing solves is serialization
2014-08-01 11:08:26 <ncopa> so the mdev scripts are executed in order
2014-08-01 11:08:52 <ncopa> i woudl also like better performance
2014-08-01 11:09:37 <ncopa> for long time solution, i dont like the idea of fork/exec for each event + fork/exec for each modprobe
2014-08-01 11:10:28 <ncopa> at the same time i dont like the idea of having things like libkmod linked in and consuming memory all the time
2014-08-01 11:10:43 <ncopa> since you mostly only need it during boot
2014-08-01 11:11:39 <ncopa> crazy idea:
2014-08-01 11:13:08 <ncopa> something minimal like nldev, but istead of reading/parsing the buffer on poll
2014-08-01 11:13:21 <ncopa> simply splice(2) it to a pipe fd
2014-08-01 11:14:01 <ncopa> if pipefd is -1 then create new pipe, and fork/exec the real handler
2014-08-01 11:14:26 <ncopa> on SIGHUP on the pipefd, set it to -1
2014-08-01 11:14:59 <ncopa> the real handler reads the nlbuffer from like stdin
2014-08-01 11:15:19 <ncopa> and has a timeout on 10 seconds
2014-08-01 11:15:51 <ncopa> so during boot, when there are many many events, it handles those
2014-08-01 11:16:12 <ncopa> after 10sec of silence it exits to free the mem
2014-08-01 11:16:47 <ncopa> that way the real handler can be slighly more heavywieght, and can link in things like libkmod, libmount libblkid etc
2014-08-01 11:17:17 <ncopa> other crazy idea:
2014-08-01 11:17:43 <ncopa> soemthing minimal like nldev, but instead of fork/exec mdev, send the raw buffer to a Lua script
2014-08-01 11:18:06 <ncopa> (or parse the buffer and store it in a lua table first)
2014-08-01 11:19:41 <fabled> a mix? lightweight netlink receiver, and it pipes it to lua based handler in parsed format
2014-08-01 11:20:59 <ncopa> why not forward it to unparsed lua?
2014-08-01 11:21:28 <fabled> depends on the format
2014-08-01 11:21:38 <ncopa> but um, yes
2014-08-01 11:21:40 <ncopa> and also
2014-08-01 11:22:11 <ncopa> if the light nl listener does parsing, then it can also do filtering early
2014-08-01 11:25:19 <ncopa> i dont think we can do splice anyways since we dont know the size of the databuffer
2014-08-01 11:25:59 <ncopa> the idea was to avoid the extra buffer copy
2014-08-01 11:28:35 <fabled> ncopa, it's so low traffic that no one cares
2014-08-01 11:28:43 <fabled> likely in parsed form it's smaller
2014-08-01 11:28:50 <fabled> and parsing in C is faster than lua
2014-08-01 11:28:54 <ncopa> yeah
2014-08-01 11:29:02 <ncopa> makes sense
2014-08-01 11:29:07 <fabled> the slowest part is launching the big handler binary
2014-08-01 11:29:37 <fabled> getting amortized cost of fork/exec is the thing that makes it faster for us
2014-08-01 11:29:48 <ncopa> *nod*
2014-08-01 11:29:48 <fabled> and not parsing the mod deps
2014-08-01 11:29:59 <fabled> or parsing them once per fork
2014-08-01 11:30:29 <ncopa> actually
2014-08-01 11:30:43 <ncopa> i really like the idea of minimal kuevent listener
2014-08-01 11:31:04 <ncopa> kueventd --handler /path/program
2014-08-01 11:31:35 <ncopa> because it makes it relatively easy to have different handlers
2014-08-01 11:31:42 <ncopa> for different stages during boot
2014-08-01 11:32:14 <ncopa> like from initramfs, i'd like a handler that mounts all found devices
2014-08-01 11:32:29 <ncopa> til it finds both an apkovl and a .boot_repository
2014-08-01 11:32:33 <ncopa> and then exit
2014-08-01 11:34:34 <ncopa> i'd also like to have a --trigger-script arg
2014-08-01 11:35:05 <ncopa> so you can run it in foreground and make it block
2014-08-01 11:35:10 <ncopa> set up nl listener socket
2014-08-01 11:35:13 <ncopa> run trigger
2014-08-01 11:35:20 <ncopa> then catch and handle events
2014-08-01 11:35:39 <ncopa> and finally --dont-restart-handler
2014-08-01 11:35:54 <ncopa> or --exit-on-handler-close
2014-08-01 13:33:08 <shafire> uggedal: docker is not that bad!
2014-08-01 13:33:39 <shafire> I think, I have found my combo :)    btrfs raid1, bcache, docker, like a dream
2014-08-01 13:34:11 <shafire> ncopa: I think, I will transform my flashcache disks to bcache
2014-08-01 19:33:11 <alacerda> images are broken on wiki start page from here
2014-08-01 19:37:19 <shafire> alacerda: http://mirror1.hs-esslingen.de/pub/Mirrors/alpine/v3.0/releases/x86_64/
2014-08-01 19:38:41 <alacerda> shafire, sorry... maybe i didnt make my self clear.
2014-08-01 19:38:51 <alacerda> go to http://wiki.alpinelinux.org/wiki/Main_Page
2014-08-01 19:39:00 <alacerda> and you'll see what i am talking about
2014-08-01 19:39:18 <alacerda> s/images/pictures/
2014-08-01 19:47:03 <shafire> its all okay here
2014-08-01 19:47:08 <shafire> send me a screenshot
2014-08-01 19:54:20 <plueschopath> it's okay here too
2014-08-01 22:49:16 <barthalion> algitbot: I wish I could believe you
2014-08-01 22:54:01 <clandmeter> alacerda: they have been broken for me too for some time.
2014-08-03 12:09:40 <moskvax> is there any way to force abuild -r to install masked dependencies?
2014-08-04 08:04:55 <fabled> seems openssl has learned about how to do security fix releases
2014-08-04 08:04:56 <fabled> http://marc.info/?l=openssl-dev&m=140706092626158&w=2
2014-08-04 13:13:44 <fabled> ok. asterisk on edge seems much better after the last commit
2014-08-04 14:54:26 <clandmeter> crow: ^
2014-08-04 15:11:47 <clandmeter> ncopa: why are my pkgs not on mirror?
2014-08-04 18:10:37 <crow> clandmeter nice, thank you for vdr aport :)
2014-08-04 18:11:00 <crow> clandmeter vdr-softhddevice as output device is needed
2014-08-04 18:11:21 <crow> if watching on same machine as the one vdr is running
2014-08-04 19:07:29 <clandmeter> crow: i dont :)
2014-08-05 05:55:08 <uggedal> anyone noticed that bb free no longer reports used memory without the pagecache?
2014-08-05 05:55:19 <uggedal> http://sprunge.us/LGXW
2014-08-05 05:55:42 <uggedal> in the "-/+ buffers" line
2014-08-05 06:07:26 <ncopa> uggedal: i havent noticed. could be related the recent bb top issue
2014-08-05 06:18:53 <uggedal> looks like it never did, it uses sysinfo(2) and does not use /proc/meminfo to determine the page cache like procps does
2014-08-05 07:18:52 <clandmeter> ncopa: something wrong with bb grep? its using a lot of memory on recursive grep.
2014-08-05 07:19:03 <clandmeter> its causing oom
2014-08-05 07:24:04 <fabled> clandmeter, what you grep?
2014-08-05 07:24:12 <fabled> grepping /proc or /sys is generally a bad idea
2014-08-05 07:24:21 <fabled> and can cause that
2014-08-05 07:24:33 <clandmeter> grep /usr
2014-08-05 07:25:12 <clandmeter> i compared it to an ubuntu isntallation, and it uses almost no memory.
2014-08-05 07:36:05 <clandmeter> fabled: a package is looking for atomic_t which it thinks alsa should provide alsa/iatomic.h (which i cannot find)
2014-08-05 07:36:09 <clandmeter> any idea?
2014-08-05 07:48:27 <fabled> atomic_t sound like some compiler generic thingy
2014-08-05 07:48:59 <fabled> i think alsa/iatomic.h is internal alsa header and not supposed to be package
2014-08-05 07:50:21 <clandmeter> it says #include <alsa/iatomic.h>               // portable atomic_t
2014-08-05 07:59:12 <fabled> it's in alsa-lib-dev
2014-08-05 11:44:51 <clandmeter> fabled: i just installed 3.0.2 and when i search for asterisk, it shows asterisk 2x. why is that?
2014-08-05 11:49:19 <fabled> you have it available twice? perhaps you have two mirrors, that are out of sync and there's two versions
2014-08-05 11:49:30 <fabled> or the one in .iso and the one from network
2014-08-05 11:49:59 <clandmeter> i have a single line in repo file
2014-08-05 12:00:16 <fabled> maybe in apk cache then?
2014-08-05 12:00:20 <fabled> try 'apk policy asterisk'
2014-08-05 12:00:25 <fabled> it lists where it finds them
2014-08-05 12:04:54 <clandmeter> fabled: http://sprunge.us/bjCa
2014-08-05 12:05:02 <clandmeter> so it lists extra if its installed?
2014-08-05 12:08:55 <fabled> depends on command
2014-08-05 13:12:43 <fabled> ncopa, you here? i have patch for nfs that you could try...
2014-08-05 13:14:23 <clandmeter> \o/
2014-08-05 13:14:51 <clandmeter> fabled: im just trying asterisk on 3.0.1
2014-08-05 13:15:28 <clandmeter> there are some error messages when i try to build it myself
2014-08-05 13:15:33 <clandmeter> but build continius
2014-08-05 13:15:52 <clandmeter> Makefile:270: recipe for target 'bininstall' failed
2014-08-05 13:15:52 <clandmeter> make[1]: *** [bininstall] Error 1
2014-08-05 13:15:52 <clandmeter> Makefile:566: recipe for target 'main-bininstall' failed
2014-08-05 13:15:53 <clandmeter> make: *** [main-bininstall] Error 2
2014-08-05 13:16:08 <clandmeter> and asterisk-sounds-en is empty
2014-08-05 13:17:25 <fabled> i think it needs internet during build to get files
2014-08-05 13:17:38 <fabled> but that snippet of build output is not enoug
2014-08-05 13:17:51 <clandmeter> it has internet
2014-08-05 13:19:39 <clandmeter> ill rebuild it and get log
2014-08-05 13:20:08 <clandmeter> fabled: asterisk-sounds-en on repo is also 1kb
2014-08-05 13:20:18 <fabled> hmm
2014-08-05 13:20:33 <clandmeter> moh to
2014-08-05 13:21:17 <fabled> yeah, you're right
2014-08-05 13:21:32 <fabled> seems to fail similarly on builders
2014-08-05 13:22:41 <clandmeter> are you ok with me adding cdr_mysql and split it?
2014-08-05 13:23:07 <fabled> i'll fix first the current issue
2014-08-05 13:23:28 <fabled> it's missing rasterisk link and other stuff too
2014-08-05 13:34:30 <fabled> oh
2014-08-05 13:44:12 <fabled> clandmeter, ok, try now
2014-08-05 14:05:43 <clandmeter> fabled: didnt get your reply, but i enabled it anyway.
2014-08-05 14:10:03 <fabled> clandmeter, i don't like the SVN based mp3 playing stuff. it means we have non-deterministic builds now
2014-08-05 14:10:14 <fabled> the mysql stuff is ok
2014-08-05 14:10:59 <clandmeter> nonwhat?
2014-08-05 14:11:27 <fabled> if they update SVN code, and we rebuild, we have rebuilt different code
2014-08-05 14:11:36 <fabled> we should really take SVN snapshots
2014-08-05 14:29:56 <ncopa> finally here
2014-08-05 14:30:14 <clandmeter> hi
2014-08-05 14:30:21 <ncopa> had some really non-fun with trying to run some 16bit app in a windows 7 64 bit...
2014-08-05 14:30:39 <clandmeter> lxc on windows :D
2014-08-05 14:30:52 <ncopa> yeah... its xp mode
2014-08-05 14:30:59 <ncopa> aka virtual-pc
2014-08-05 14:31:02 <ncopa> with windows xp
2014-08-05 14:32:13 <ncopa> i should test the nfs patches
2014-08-05 14:32:24 <ncopa> first look is promising
2014-08-05 14:32:27 <ncopa> less code
2014-08-05 14:32:30 <ncopa> better code
2014-08-05 14:32:33 <clandmeter> i think fabled broke asterisk :)
2014-08-05 14:32:50 <clandmeter> http://sprunge.us/MBZU
2014-08-05 14:32:51 <ncopa> i think asterisk broke it themselves, and fabled fixed it...
2014-08-05 14:33:04 <clandmeter> i think this time its not asterisk
2014-08-05 14:33:25 <clandmeter> prob musl commit releated?
2014-08-05 14:33:29 <ncopa> oh
2014-08-05 14:33:37 <ncopa> i think he tried fix lua module with asterisk
2014-08-05 14:33:59 <ncopa> might be the lua fix broke other modules?
2014-08-05 14:34:03 <fabled> ?
2014-08-05 14:34:21 <clandmeter> fabled: http://sprunge.us/MBZU
2014-08-05 14:34:34 <fabled> i have not tried pjsip module
2014-08-05 14:34:46 <fabled> chan_sip worked for me
2014-08-05 14:35:01 <clandmeter> i just installed it and tried to start it
2014-08-05 14:35:13 <clandmeter> shoudlnt it just run with default config?
2014-08-05 14:35:36 <fabled> sounds like it loads modules in wrong order or something like that
2014-08-05 14:36:12 <fabled> before i started working on it two days ago, it hung on first module load
2014-08-05 14:36:23 <fabled> so it's quite a lot better situation compared to that
2014-08-05 14:36:33 <fabled> but seems not perfect yet
2014-08-05 14:36:53 <clandmeter> dont get me wrong, im not complaining :)
2014-08-05 14:37:08 <ncopa> libvirt is broken too, it looks for "Hugepagesize:" in /proc/meminfo which is not found
2014-08-05 14:37:26 <barthalion> and from completely not important things, udp in umurmur doesn't work
2014-08-05 14:37:39 <barthalion> I'm not complaining either, I'd love some pointers how can I debug it
2014-08-05 14:37:48 <ncopa> strace?
2014-08-05 14:38:11 <barthalion> ok, I'll try
2014-08-05 14:43:21 <fabled> clandmeter, i think those errors are normal
2014-08-05 14:43:45 <fabled> seems asterisk does not recognize dependencies and can try to load them in wrong order, but later reloads them
2014-08-05 14:44:10 <fabled> I see it trying to load chan_pjsip first, and it fails
2014-08-05 14:44:18 <fabled> it then later on load chan_pjsip successfully
2014-08-05 14:44:26 <fabled> seems braindamaged module design
2014-08-05 14:44:34 <fabled> but it's been like that since ages
2014-08-05 14:45:12 <clandmeter> there is no clear msg why it didnt start
2014-08-05 14:46:53 <clandmeter> Assertion failed: mod_evsub.mod.id != -1 (../src/pjsip-simple/evsub.c: pjsip_evsub_register_pkg: 416)
2014-08-05 14:47:48 <fabled> uh
2014-08-05 14:47:49 <fabled> ok
2014-08-05 14:47:54 <fabled> i haven't tried chan_pjsip
2014-08-05 14:51:58 <fabled> but do note, chan_pjsip is new feature in asterisk 12. not available in alpine-2.7
2014-08-05 14:52:07 <fabled> thus it's completely untested both in uclibc and musl
2014-08-05 14:54:23 <clandmeter> yes i know its new
2014-08-05 14:54:49 <clandmeter> I wanted to try new freepbx which has support for it.
2014-08-05 14:55:32 <clandmeter> got a spare APU to try out.
2014-08-05 16:43:24 <alacerda> Technicaly what is the diference between /etc/init.d/<my_service> start and rc-service <my_service> start
2014-08-05 16:43:25 <alacerda> ?
2014-08-05 16:44:53 <alacerda> Is using rc-service better than /etc/init.d/ scripts?
2014-08-05 17:32:52 <uggedal> not really
2014-08-05 17:33:07 <uggedal> it does provide some options though: https://github.com/OpenRC/openrc/blob/master/src/rc/rc-service.c
2014-08-05 18:44:39 <alacerda> uggedal, so there is no any advantage on using rc-service if all that i want is start and stop services?
2014-08-05 18:46:24 <barthalion> yep
2014-08-05 18:46:31 <alacerda> tks
2014-08-06 05:49:07 <zenny> Hi, new to alpine, so bear with me. I would like to build a custom iso which has zfs and snapraid built in. Any suggestion? Could not fine zfs-grsec package. Thanks
2014-08-06 06:21:24 <fabled> bah
2014-08-06 06:21:33 <fabled> huh
2014-08-06 06:21:46 <fabled> ?!
2014-08-06 06:21:55 <fabled> .gitignore masks out -patch.gz ?
2014-08-06 06:22:15 <fabled> looks that way
2014-08-06 06:54:16 <barthalion> zenny: it's in testing
2014-08-06 06:57:07 <zenny> barthalion: how can I use the testing with the 3.0.2?
2014-08-06 07:04:30 <barthalion> zenny: you not really can, maybe pinning can help you
2014-08-06 07:10:15 <zenny> brthalion: Does traditional './configure; make; make install' works in alpine for the programs that is not available from apk?
2014-08-06 07:12:58 <barthalion> assuming you have the deps, yes
2014-08-06 07:13:48 <zenny> btw, there is http://git.alpinelinux.org/cgit/fab/alpine-iso/plain/config-builder.py, but when I try to execute, it reports that there is no 'BeautifulSoup' to import as mentioned in line 28, what is this 'BeautfulSoup' module?
2014-08-06 07:15:50 <barthalion> http://www.crummy.com/software/BeautifulSoup/index.html
2014-08-06 07:21:36 <zenny> Is there a special way to install BeautfulSoup in alpine or 'pip install MODULE' works?
2014-08-06 07:21:52 <barthalion> check the repositories first
2014-08-06 07:46:23 <clandmeter> good morning
2014-08-06 07:53:43 <clandmeter> fabled: Assertion failed: mod_evsub.mod.id != -1, is that a config or code issue?
2014-08-06 07:54:45 <clandmeter> or should i just stay with 2.7?
2014-08-06 08:24:22 <fabled> FYI, new openssl security release is due out today http://marc.info/?l=openssl-dev&m=140706092626158&w=2
2014-08-06 09:00:43 <fabled> clandmeter, seems nfs might work
2014-08-06 09:01:53 <fabled> on edge that is
2014-08-06 09:19:35 <clandmeter> fabled: nice :)
2014-08-06 09:19:44 <clandmeter> i think a few people will be happy
2014-08-06 09:21:55 <crow> fabled thanks for nfs :)
2014-08-06 09:24:47 <f15RyWpgGTO6> ohi, i tested pcsc-lite, and found out, that the musl stacksize and the habbit of the pcsc devs allocating buffers on stack do not harmonize (aka segfault) i came up with this solution, would you please merge this (i know i fucked up the commit msg, bad habbit /o\)
2014-08-06 09:25:16 <f15RyWpgGTO6> this is the complete commit: https://github.com/stef/aports/commit/5887470b8ff74d23bc042efbdc6cf97a55750958
2014-08-06 09:25:29 <f15RyWpgGTO6> also i packaged ccid: https://github.com/stef/aports/commit/d3cb38379dff8f68e1879f487e90dd983eb3aedb
2014-08-06 09:25:53 <f15RyWpgGTO6> which means most openpgp smartcard tokens should now work on alpine.
2014-08-06 09:35:27 <ncopa> cool!
2014-08-06 09:37:11 <f15RyWpgGTO6> re pcsc patch i sent it also upstream to ludovic the dev of pcsc hope it gets merged
2014-08-06 09:37:19 <ncopa> very well
2014-08-06 09:37:24 <ncopa> i was about to ask about that :)
2014-08-06 09:39:10 <ncopa> i'll clean up the patch
2014-08-06 09:39:21 <ncopa> need to increase the pkgrel too
2014-08-06 09:39:35 <f15RyWpgGTO6> oooh, sorry yes
2014-08-06 09:40:01 <ncopa> we have also #3237
2014-08-06 09:40:53 <f15RyWpgGTO6> btw i just saw you merged also my zfs-grsec muslfix patch? awesome!
2014-08-06 09:41:06 <f15RyWpgGTO6> i'll switch over to that kernel again then.
2014-08-06 09:41:20 <ncopa> um, not sure it works atm
2014-08-06 09:41:50 <f15RyWpgGTO6> well it works for me, but from my own apkbuild, will use the official then.
2014-08-06 09:43:29 <f15RyWpgGTO6> re Bug #3237: that is basically a missing user/group i believe which does not get created on install but the pcscd.init script seems to need this combination of owner/group
2014-08-06 09:44:08 <f15RyWpgGTO6> ooh, sorry algitbot not going to provoke you further :)
2014-08-06 09:49:56 <ncopa> f15RyWpgGTO6: i pushed a patch for it, could you please verify that it works now?
2014-08-06 09:51:41 <f15RyWpgGTO6> yes! will try it out in the afternoon.
2014-08-06 13:10:14 <fabled> bah
2014-08-06 13:10:21 <fabled> cloog does not build with new isl
2014-08-06 14:32:04 <f15RyWpgGTO6> ncopa the pcsc-lite patch is included in upstream: http://lists.alioth.debian.org/pipermail/pcsclite-cvs-commit/2014-August/006503.html
2014-08-06 14:33:08 <ncopa> ah
2014-08-06 14:33:09 <ncopa> wonderful
2014-08-06 14:33:38 <ncopa> good job!
2014-08-06 14:34:58 <ncopa> i'll cherry-pick it for 3.0-stable
2014-08-06 14:35:03 <ncopa> before i forget...
2014-08-06 14:41:18 <f15RyWpgGTO6> i'll test the 3237 bug in a bit. but it turns out i can't boot into my system atm, gotta admit i fiddled with some mkfsinit. when i'm in. i'll test it.
2014-08-06 14:42:38 <f15RyWpgGTO6> also there seems to be a problem with gcc --visibility=hidden is there any better way than disabling this in configure(|.ac)?
2014-08-06 14:50:59 <f15RyWpgGTO6> hmm, my system can't find any volumegroups after cryptsetup
2014-08-06 14:54:53 <f15RyWpgGTO6> ooooh. cryptsetup segfaults
2014-08-06 15:14:44 <f15RyWpgGTO6> hmmmm, interesting history for http://git.alpinelinux.org/cgit/aports/log/main/cryptsetup/APKBUILD - why cryptsetup was upgraded to 1.6.4 and then back to 1.6.3?
2014-08-06 15:16:26 <barthalion> commit message is wrong
2014-08-06 15:16:33 <barthalion> it was license fix
2014-08-06 15:21:02 <f15RyWpgGTO6> aaah right
2014-08-06 15:21:23 <f15RyWpgGTO6> thx
2014-08-06 18:58:58 <f15RyWpgGTO6> uhoh. so cryptsetup 1.6.5 segfaults during boot. when downgrading to 1.6.4 it works fine.
2014-08-06 18:59:35 <f15RyWpgGTO6> dunno if it's this: https://bugzilla.redhat.com/show_bug.cgi?id=1115120, but apparently it's been fixed in grsec? https://grsecurity.net/changelog-stable2.txt in commit c7e0864c485249dcd45faa2b1468eecd29544505
2014-08-06 19:59:34 <f15RyWpgGTO6> ncopa: confirmed the user creation fix in pcsc-lite works now for me. (bug: 3237) thx!
2014-08-06 20:01:07 <f15RyWpgGTO6> also: cryptsetup 1.6.5 segfaults in initrd. downgrading to 1.6.4 lets me boot back into the system.
2014-08-06 21:32:22 <f15RyWpgGTO6> oh, stacksize tuned down after tests: http://lists.alioth.debian.org/pipermail/pcsclite-cvs-commit/2014-August/006505.html
2014-08-06 21:34:49 <zenny> Can anyone help me to build an alpine iso with zfs and snapraid. The one here ( https://github.com/zfsonlinux/zfs/pull/2380) does not seem to work with latest 3.0.2 and zfs-0.6.3. Any input appreciated!
2014-08-06 22:01:12 <f15RyWpgGTO6> i need to check that takes some time though.
2014-08-06 22:02:05 <f15RyWpgGTO6> there are merge conflicts
2014-08-06 22:02:07 <f15RyWpgGTO6> hmmm
2014-08-07 05:18:49 <fabled> urgh
2014-08-07 05:18:58 <fabled> i messed up the man-pages
2014-08-07 05:19:01 <fabled> sigh
2014-08-07 05:19:03 <fabled> i need more coffee
2014-08-08 00:22:15 <f15RyWpgGTO6> i found the solution to Bug #3281, but a better patch and upstream notification is still pending
2014-08-08 06:50:35 <ncopa> f15RyWpgGTO6: do you have link to upstream discussion
2014-08-08 08:06:11 <ncopa> ok
2014-08-08 08:06:15 <ncopa> im looking at nfs again
2014-08-08 08:06:25 <ncopa> status is: basic server works
2014-08-08 08:06:31 <ncopa> nfsvers=2 is broke
2014-08-08 08:06:48 <ncopa> rpc.idmap is broke due to dnotify not enabled in kernel
2014-08-08 08:07:06 <ncopa> i wonder if we should just enable it?
2014-08-08 08:07:32 <ncopa> rpcbind appears to segfault once in a while? i dont know how to reliable reproduce though
2014-08-08 08:28:36 <f15RyWpgGTO6> ncopa: no upstream contact yet. sorry was to late in the morning when i fixed this
2014-08-08 08:29:04 <ncopa> good work
2014-08-08 08:29:05 <f15RyWpgGTO6> but basically the problem is i use detached luks header, which is a simple file and not a device
2014-08-08 08:29:21 <ncopa> and the file neds to be on tmpfs i suppose
2014-08-08 08:29:28 <f15RyWpgGTO6> does it?
2014-08-08 08:29:30 <f15RyWpgGTO6> nah
2014-08-08 08:29:35 <f15RyWpgGTO6> it's in my initramfs
2014-08-08 08:29:45 <f15RyWpgGTO6> it's ok if it's read only
2014-08-08 08:30:00 <ncopa> i mean to reproduce the issue
2014-08-08 08:30:08 <ncopa> i think tmpfs has no support for O_DIRECT
2014-08-08 08:30:09 <f15RyWpgGTO6> but initramfs (whatever the underlying fs, haven't checked) does not support directio
2014-08-08 08:30:18 <f15RyWpgGTO6> ooh. ok, the maybe this is a way
2014-08-08 08:30:22 <ncopa> thats tmpfs
2014-08-08 08:30:35 <ncopa> initramfs uses tmpfs
2014-08-08 08:30:37 <f15RyWpgGTO6> ok.
2014-08-08 08:30:50 <ncopa> so mount a tmpfs, but a file there
2014-08-08 08:30:53 <ncopa> and cryptsetup
2014-08-08 08:30:56 <ncopa> should reproduce it
2014-08-08 08:30:59 <f15RyWpgGTO6> so then you can create a detached luks header like this:
2014-08-08 08:32:11 <ncopa> i sent an email in the upstream direction: http://www.saout.de/pipermail/dm-crypt/2014-August/004228.html
2014-08-08 08:32:34 <f15RyWpgGTO6> dd if=/dev/zero of=test.img bs=16M count=1 # for the crypt device
2014-08-08 08:32:42 <f15RyWpgGTO6> dd if=/dev/zero of=test.luks bs=4M count=1 # for the crypt device
2014-08-08 08:32:59 <f15RyWpgGTO6> ^^^ for the luks header, needs 4M of space, see: https://code.google.com/p/cryptsetup/wiki/Cryptsetup140
2014-08-08 08:33:31 <f15RyWpgGTO6> and then cryptsetup luksFormat test.img --header test.luks
2014-08-08 08:33:42 <f15RyWpgGTO6> and then luksOpen
2014-08-08 08:33:54 <f15RyWpgGTO6> maybe the luksformat will already fail if you put it directly on tmpfs
2014-08-08 08:34:07 <ncopa> i can try
2014-08-08 08:35:53 <ncopa> hm
2014-08-08 08:36:06 <ncopa> seems like i have flush error
2014-08-08 08:36:31 <ncopa> looks like there is missing fflush()
2014-08-08 08:36:52 <ncopa> the Are you sure? (Type uppercase yes):
2014-08-08 08:36:56 <ncopa> is not fflushed
2014-08-08 08:37:32 <ncopa> what the syntax for luksOpen?
2014-08-08 08:38:23 <f15RyWpgGTO6>  cryptsetup luksOpen --debug --header test.luks test.img test
2014-08-08 08:40:32 <ncopa> https://gist.github.com/ncopa/b3560e7da2c6e6183f16
2014-08-08 08:40:38 <ncopa> Command failed with code 22: Device /run/test.luks is not a valid LUKS device.
2014-08-08 08:41:19 <f15RyWpgGTO6> huhwtf
2014-08-08 08:42:26 <f15RyWpgGTO6> ok i'm booting in my dev vm to check
2014-08-08 08:51:25 <f15RyWpgGTO6> why is test.luks considered a luks device?
2014-08-08 08:53:04 <f15RyWpgGTO6> how about --header=test.luks when doing luksformat?
2014-08-08 09:09:25 <f15RyWpgGTO6> so i was able to reproduce it in a dev vm: http://www.ctrlc.hu/~stef/cryptsetup.txt
2014-08-08 09:09:56 <f15RyWpgGTO6> but now i gotta go afk for a few hours sorry.
2014-08-08 09:16:09 <f15RyWpgGTO6> testing took so long because i did not recognize that i had a patched cryptsetup installed which failed to rail
2014-08-08 09:16:13 <f15RyWpgGTO6> s/rail/fail/
2014-08-08 11:42:20 <f15RyWpgGTO6> ncopa: have you been able to reproduce according to my script?
2014-08-08 14:35:40 <f15RyWpgGTO6> oooh, nice there is already a fix upstream, and much nicer https://code.google.com/p/cryptsetup/source/detail?r=89f795d7b453acf347c0b293d09746a8c0d48d12#
2014-08-08 15:15:59 <f15RyWpgGTO6> \o/
2014-08-08 15:16:42 <f15RyWpgGTO6> <3
2014-08-08 15:17:11 <ncopa> been out of office half day
2014-08-08 15:17:16 <ncopa> and on my way home now
2014-08-08 15:17:31 <ncopa> thanks for helping me fix that issue. good job!
2014-08-08 15:17:41 <ncopa> or more like, thanks for helping upstream fixing it :)
2014-08-08 15:17:45 <ncopa> see u monday
2014-08-08 15:19:06 <f15RyWpgGTO6> good job too, thanks!!!
2014-08-08 18:54:13 <grimeton> ehm, why is chrony listening on udp4 on port 323?
2014-08-08 19:00:28 <grimeton> ah yeah, pebcak
2014-08-09 13:09:46 <darkfader> anyone here who could take a stab at http://bugs.alpinelinux.org/issues/2531 ? i'll gladly donate a pizza even just for a working client to mount filesystems :)
2014-08-09 13:14:39 <f15RyWpgGTO6> is that "pizza even" or a typo and you mean "pizza oven"
2014-08-09 13:29:46 <darkfader> good question, something  happeneded between "think" and "keyboard"
2014-08-09 13:29:57 <darkfader> definitely not oven, home pizza ovens are stupid
2014-08-09 13:34:40 <f15RyWpgGTO6> naah. you can use them to solder smds for example, highly useful.
2014-08-09 13:41:30 <darkfader> hehe
2014-08-11 02:35:57 <Mp5shooter> is it just me or is there a very slow mirror right now
2014-08-11 06:52:11 <ncopa> kaniini: can i just purge encahceio?
2014-08-11 07:32:51 <clandmeter> ncopa: regarding nfs, its safe to use now?
2014-08-11 07:33:50 <ncopa> we barely got it running
2014-08-11 07:33:54 <ncopa> might still be issues
2014-08-11 09:38:32 <f15RyWpgGTO6> clandmeter: ohi i simplified the musl patch for spl-grsec https://github.com/stef/aports/blob/master/testing/spl-grsec/pagesize.patch
2014-08-11 09:39:07 <f15RyWpgGTO6> also i'm working (and getting contribs) to make the zfs-grsec patch also better. some of the changes have been already merged in mainline, but not all
2014-08-11 09:39:26 <f15RyWpgGTO6> but zfs-grsec will take some more work.
2014-08-11 13:49:08 <vkrishn> new release - mosquitto v1.3.4
2014-08-11 13:49:55 <vkrishn> this could be nice to have, http://www.themukt.com/2014/08/08/move-running-program-another-terminal-using-reptyr/
2014-08-11 13:50:06 <vkrishn> http://sooda.dy.fi/foo/screenify
2014-08-11 13:50:12 <vkrishn> https://github.com/nelhage/reptyr
2014-08-11 13:51:04 <barthalion> reptyr won't work on grsec
2014-08-11 14:06:04 <vkrishn> ok
2014-08-11 14:22:28 <barthalion> in fact, I doubt it would on any modern server
2014-08-11 14:22:33 <barthalion> s/server/kernel/
2014-08-11 14:22:57 <barthalion> maybe with root
2014-08-11 14:23:08 <barthalion> but it renders it quite useless
2014-08-11 14:25:03 <barthalion> yep, I'm right
2014-08-11 14:25:24 <barthalion> yama disallows ptracing non-child processes by regular users
2014-08-11 14:29:03 <vkrishn> the author is 'security hacker'
2014-08-11 14:29:28 <barthalion> well, it doesn't change the fact it won't work on grsec or without root on vanilla kernels
2014-08-11 14:31:35 <vkrishn> ok, just found it interesting
2014-08-11 14:31:48 <vkrishn> currently reading more of his work
2014-08-11 14:31:50 <vkrishn> https://nelhage.com/talks/kvm-defcon-2011.pdf
2014-08-11 18:25:24 <kaniini> ncopa: yes
2014-08-11 18:25:28 <kaniini> ncopa: go for it
2014-08-12 09:48:48 <ncopa> bah
2014-08-12 09:48:56 <ncopa> php-5.5.15 is buggy
2014-08-12 09:49:02 <ncopa> the crypt implementation
2014-08-12 09:51:07 <darkfader> hurray
2014-08-13 04:58:02 <kaniini> ncopa: have you seen http://lkml.iu.edu//hypermail/linux/kernel/1408.1/02496.html yet?
2014-08-13 04:58:22 <kaniini> ncopa: i think this might be a good opportunity to highlight that alpine does not, and does not have plans to, ship systemd
2014-08-13 06:08:16 <ncopa> morning
2014-08-13 06:08:24 <ncopa> *reading*
2014-08-13 06:12:35 <ncopa> lol
2014-08-13 06:15:23 <ncopa> he has some good points
2014-08-13 06:15:44 <ncopa> "In a 'One Linux' world, what would distros actually be? Deprecated. No longer relevant."
2014-08-13 06:15:48 <ncopa> thats kinda my point
2014-08-13 06:16:20 <ncopa> why no just use redhat/fedora since it just does the same thing anyways
2014-08-13 06:16:57 <ncopa> i bet systemd will come up with a "One Linux" standard package manager in the future
2014-08-13 06:22:35 <rnalrd> OneLinux == zero-choice
2014-08-13 06:22:42 <rnalrd> how true
2014-08-13 07:09:59 <kaniini> ncopa: there is PackageKit already.
2014-08-13 07:10:42 <kaniini> anyway, my point here is
2014-08-13 07:11:00 <kaniini> i think we could seriously take advantage of alpine's unique position on systemd, to gain usershare
2014-08-13 07:15:08 <ncopa> feel free to comment on it, but i cannot promise that we will *never* switch to systemd
2014-08-13 07:15:26 <ncopa> i higly doubt we ever will, but i dont want give any promises
2014-08-13 07:16:11 <ncopa> i found his CameCase comment fun :)
2014-08-13 07:16:38 <ncopa> PackageKit is just a frontend to other packagemanagers
2014-08-13 07:16:52 <ncopa> i think systemd have something in mind for rolling out software
2014-08-13 07:17:03 <ncopa> atleast they have a solution for system reset
2014-08-13 07:17:07 <ncopa> 'factory default'
2014-08-13 07:17:30 <ncopa> so the natural next step, once most distros are onboard with systemd
2014-08-13 07:17:47 <ncopa> will be a solution for rolling out software (eg packagemanager replacement)
2014-08-13 07:17:57 <kaniini> my understanding is that our position is that systemd would have to eliminate numerous design flaws *and* kdbus would have to become a reality
2014-08-13 07:18:10 <ncopa> correct
2014-08-13 07:18:49 <kaniini> alpine has a very unique position here though.
2014-08-13 07:19:00 <ncopa> *and* upstream support for musl libc
2014-08-13 07:19:16 <ncopa> upstream support for musl libc is critical
2014-08-13 07:19:32 <ncopa> their current position is 'every libc should mimic glibc'
2014-08-13 07:19:41 <kaniini> not to mention any other thing we do
2014-08-13 07:19:55 <kaniini> for example porting the alpine userland to run on bsd
2014-08-13 07:20:06 <kaniini> openrc would work on freebsd right now :P
2014-08-13 07:20:12 <ncopa> yeah
2014-08-13 07:20:33 <ncopa> that kind of stuff gentoo and debian does (or used to do?)
2014-08-13 07:20:56 <ncopa> that kind of things that systemd thinks is "pointless"
2014-08-13 07:21:06 <kaniini> my understanding is that the main driver of debian gnu/kfreebsd gave up
2014-08-13 07:21:16 <ncopa> i dont blame him
2014-08-13 07:21:26 <ncopa> thats what sysmted wanted i suppose
2014-08-13 07:21:30 <ncopa> systemd devs
2014-08-13 07:21:47 <kaniini> i think what the lkml post says is right
2014-08-13 07:21:51 <kaniini> systemd is a trojan horse
2014-08-13 07:22:04 <ncopa> they are talking about "unifying" linux
2014-08-13 07:22:11 <kaniini> redhat and microsoft (i.e. suse) wish to eliminate distributions
2014-08-13 07:22:26 <kaniini> instead they would have one distribution, with different support options
2014-08-13 07:22:41 <kaniini> and it would become more about 'middleware'
2014-08-13 07:22:44 <ncopa> that fits into what systemd is slowly doing yes
2014-08-13 07:23:01 <ncopa> making it pointless to roll your own distro
2014-08-13 07:23:15 <kaniini> while alpine is the opposite
2014-08-13 07:23:37 <kaniini> most of the technology in alpine is explicitly about 'rolling your own system images'
2014-08-13 07:23:40 <kaniini> i.e. lbu
2014-08-13 07:24:24 <kaniini> thusly, i would say that the systemd agenda and our agenda are mutually incompatible, making it highly unlikely (effectively an option of 'last resort') for us to adopt systemd
2014-08-13 07:24:45 <ncopa> *nod*
2014-08-13 07:24:53 <kaniini> but, back to my original point: alpine in a lot of ways still has debian influences
2014-08-13 07:25:08 <kaniini> a debian sysadmin can pick up an alpine install in a few minutes
2014-08-13 07:25:37 <kaniini> and i do not think many debian users are happy about this systemd thing
2014-08-13 07:25:47 <kaniini> or developers
2014-08-13 07:25:53 <ncopa> so you want to invite those over to alpine...
2014-08-13 07:25:59 <kaniini> yes
2014-08-13 07:26:29 <ncopa> feel free to do so
2014-08-13 07:26:47 <ncopa> i dont think i will put much energy in that
2014-08-13 07:26:53 <kaniini> oh of course not
2014-08-13 07:27:09 <kaniini> i was mostly just wondering if that is something you felt worth pursuing
2014-08-13 07:27:10 <ncopa> but please dont invite the ubuntu crowd... :-)
2014-08-13 07:27:14 <ncopa> yes
2014-08-13 07:27:26 <kaniini> the ubuntu crowd is the type of crowd which likes systemd
2014-08-13 07:27:32 <ncopa> right
2014-08-13 07:27:42 <kaniini> a lot of the alpine base comes from debian already
2014-08-13 07:27:51 <ncopa> and arch
2014-08-13 07:27:57 <kaniini> we are the people, who basically got forced out by the ubuntu people
2014-08-13 07:27:58 <kaniini> yes
2014-08-13 07:28:40 <ncopa> i think you have good points
2014-08-13 07:28:48 <kaniini> for example, when i evaluated alpine 2.0.0-rc1 to begin with
2014-08-13 07:28:55 <kaniini> (wow that was a long time ago :))
2014-08-13 07:29:07 <ncopa> ha you have become a longtimer!
2014-08-13 07:29:14 <kaniini> i was looking for something like debian, but more modern, and without the influences of canonical
2014-08-13 07:30:16 <ncopa> yes, i think the kinda people we want invite is
2014-08-13 07:30:37 <ncopa> people who know to appreciate effective code
2014-08-13 07:30:39 <ncopa> small
2014-08-13 07:30:41 <ncopa> simple
2014-08-13 07:30:42 <ncopa> secure
2014-08-13 07:31:25 <ncopa> "this distro can do this in 10MB? wow!"
2014-08-13 07:31:39 <ncopa> if you appreciate that
2014-08-13 07:31:43 <ncopa> then alpine is for you
2014-08-13 07:32:06 <ncopa> if you say "2 or 10G who cares?" then maybe not so much
2014-08-13 07:32:15 <kaniini> because, if you've been watching debian over the years
2014-08-13 07:32:19 <kaniini> it's gotten a lot more bloated
2014-08-13 07:32:51 <ncopa> i havent followed really, but i can clrearly see that there are traces of wanting avoid bloat
2014-08-13 07:32:54 <kaniini> debian used to be a distribution known for being lightweight
2014-08-13 07:33:10 <ncopa> yes i can see that it atleast have been lightweight
2014-08-13 07:33:20 <ncopa> the split
2014-08-13 07:33:25 <ncopa> they split almost every library
2014-08-13 07:33:32 <ncopa> into a libsomething package
2014-08-13 07:33:46 <ncopa> so you can avoid installing stuff you will not use
2014-08-13 07:34:03 <ncopa> = clear sign of wanting to be lightweight
2014-08-13 07:34:18 <kaniini> i always like to kind of think of alpine as what debian would have been if it were started in the 2000s instead of the 1990s
2014-08-13 07:35:02 <kaniini> so i definitely think there's certainly a conversation worth having with at least some of the more server-oriented/devops areas of debian
2014-08-13 07:35:47 <ncopa> yes
2014-08-13 07:36:03 <kaniini> and with partial glibc binary compatibility being a goal of musl, we can do more to bring people over now
2014-08-13 07:36:15 <ncopa> *nod*
2014-08-13 07:37:22 <ncopa> i'm ok with inviting them
2014-08-13 07:38:57 <kaniini> i will kind of work on that then
2014-08-13 07:41:57 <ncopa> btw, how is it going with the supervised idea you talked about?
2014-08-13 07:42:50 <barthalion> I'll change the topic a bit
2014-08-13 07:43:01 <ncopa> i have been thinking a bit of it, and i think one of the major challenges is message passing
2014-08-13 07:43:10 <barthalion> have we ever considered becoming an organisation for google summer of code?
2014-08-13 07:43:24 <ncopa> not really
2014-08-13 07:43:30 <ncopa> i'm afraid of the paperwork :)
2014-08-13 07:43:34 <barthalion> I see
2014-08-13 07:44:10 <ncopa> or more like, i am afraid of paperwork. i want hack :)
2014-08-13 07:46:58 <kaniini> well, i recently went to work for a media firm, so i haven't had much time to work out the details of it yet. :/
2014-08-13 07:51:45 <ncopa> maybe we should make a wiki page or soemthing
2014-08-13 07:51:49 <ncopa> with the current ideas
2014-08-13 07:52:10 <ncopa> and write down the objective
2014-08-13 08:09:54 <uggedal> regarding systemd I tried voidlinux.eu yesterday. they are glibc based but recently switched to runit. it worked surprisingly well
2014-08-13 08:33:01 <ncopa> runit is a good candidate for pid 1
2014-08-13 08:35:32 <ncopa> i'm also thinking for boot process/service management the full dbus experience is probably overkill
2014-08-13 08:36:27 <ncopa> ubus might be an alternative
2014-08-13 09:31:23 <f15RyWpgGTO6> ohai, so let me confess: i'm a longterm debian user, and was forced out exactly because of systemd, that is the reason why i came to alpine a few months ago. i love alpine for all it is not debian: musl, no-systemd, default grsec, even zfs. alpine is in a unique position, and i'll hope to convert some more people who can maintain a few packages.
2014-08-13 09:32:09 <f15RyWpgGTO6> i was quite shocked when the tone re systemd was not completely rejecting when it was mentioned two times in the last month.
2014-08-13 09:38:25 <ncopa> f15RyWpgGTO6: I am trying to be balanced re systemd
2014-08-13 09:38:29 <ncopa> but its difficult :)
2014-08-13 09:38:47 <ncopa> i would *hate* to implement systemd for alpine linux
2014-08-13 09:40:13 <ncopa> but i don't want to burn that bridge
2014-08-13 09:40:28 <ncopa> currently we *can* not do systemd due to absolutely no support for musl
2014-08-13 09:40:40 <ncopa> not even attempting to support it from upstream
2014-08-13 09:40:55 <ncopa> they say: "your libc should mimic GNU libc. EOD"
2014-08-13 09:41:02 <ncopa> that clearly does not work for us
2014-08-13 09:41:41 <ncopa> what is the point with any competing libc if all they do is copying what GNU libc does?
2014-08-13 09:42:09 <ncopa> and i think thats kind of what is on their agenda, as kaniini mentioned earlier
2014-08-13 09:43:06 <f15RyWpgGTO6> thank the elder gods, for the arrogance of the systemd devs \o/
2014-08-13 09:43:33 <ncopa> yeah, thats the part of the problem too
2014-08-13 09:43:48 <ncopa> their narsissistic management style
2014-08-13 09:43:53 <ncopa> does not work for us
2014-08-13 09:45:18 <f15RyWpgGTO6> oh, so one of my goals is to become poettering and drepper clean
2014-08-13 09:45:35 <f15RyWpgGTO6> no pulse, ahavi, systemd, and all the drepper crap in glibc
2014-08-13 09:46:06 <f15RyWpgGTO6> i mean why do they task someone with such a central component, who has proven incompetent previously with pulse and avahi?
2014-08-13 09:46:24 <f15RyWpgGTO6> those are already crimes against humanity
2014-08-13 09:46:30 <ncopa> :)
2014-08-13 09:46:42 <ncopa> yes, i am not comfortable with that indeed
2014-08-13 09:48:05 <ncopa> i dont like php
2014-08-13 09:48:17 <ncopa> just in case anyone wondered :)
2014-08-13 09:49:09 <f15RyWpgGTO6> !!!
2014-08-13 09:49:16 <f15RyWpgGTO6> i'm with you.
2014-08-13 09:49:40 <f15RyWpgGTO6> i heard it's borken? :P
2014-08-13 09:50:50 <ncopa> well
2014-08-13 09:51:20 <ncopa> php configure script tells me: crypt supports sha512: yes
2014-08-13 09:51:31 <ncopa> our system crypt(3) has support for it
2014-08-13 09:51:35 <ncopa> good, he finds it
2014-08-13 09:51:47 <ncopa> but php -r 'echo CRYPT_SHA512."\n";'
2014-08-13 09:51:49 <ncopa> say 0
2014-08-13 09:52:24 <f15RyWpgGTO6> heh
2014-08-13 09:52:45 <ncopa> so php apps that check if system supports CRYPT_SHA512 believes, oh no sha512, lets use des/sha1/<insert other weak hash>
2014-08-13 09:52:51 <ncopa> https://bugs.php.net/bug.php?id=67827
2014-08-13 09:54:03 <f15RyWpgGTO6> ooh, but only upstream fix is needed? kool
2014-08-13 09:54:41 <f15RyWpgGTO6> anyway my only two gripes i have with alpine is: 1/ cannot use 1-char usernames, 2/ no real OS^W^Wemacs
2014-08-13 09:54:56 <f15RyWpgGTO6> (i'm working on both)
2014-08-13 10:03:54 <ncopa> problem with emacs is that it has its own malloc
2014-08-13 10:03:58 <ncopa> iirc
2014-08-13 10:04:49 <ncopa> i find it funny, because the saying that emacs is an OS is not completely wrong :)
2014-08-13 10:09:36 <f15RyWpgGTO6> hehe maybe i'll switch to vim when they include systemd in emacs ;)
2014-08-13 15:07:52 <vkrishn> ncopa, do check http://security.debian.org/debian-security/pool/updates/main/p/php5/php5_5.4.4-14+deb7u12.diff.gz
2014-08-13 15:08:17 <vkrishn> line 9253
2014-08-13 15:08:49 <vkrishn> hmmm.. 9263
2014-08-13 15:10:25 <vkrishn> could not find if https://bugs.php.net/bug.php?id=67827 is fixed in php 5.6x yet
2014-08-13 20:28:12 <kaniini> f15RyWpgGTO6: pulse is not so bad anymore now that poettering has little involvement
2014-08-13 20:29:58 <f15RyWpgGTO6> dunno, i avoid it still. luckily i can do all my ham stuff using alsa
2014-08-13 20:32:15 <kaniini> improving alpine's ham radio support would be nice
2014-08-13 20:32:32 <kaniini> i think debian adopting systemd was a huge mistake for debian
2014-08-13 20:32:47 <kaniini> i honestly could see alpine gaining a serious share of debian's userbase as a result
2014-08-13 20:32:58 <kaniini> since it's very familiar to debian users already
2014-08-14 06:22:09 <uggedal> http://www.codekid.net/grafix/systemd_monster_small.gif
2014-08-14 06:25:06 <barthalion> lol
2014-08-14 07:43:36 <barthalion> while I really like systemd, this gif is so true
2014-08-14 07:57:12 <kaniini> i have mixed feelings on systemd
2014-08-14 07:57:27 <kaniini> the core goals of systemd are good
2014-08-14 07:57:35 <kaniini> the result is another lennartp product
2014-08-14 07:57:59 <kaniini> the fact that it is being effectively *forced* onto the linux community is also something that is bad
2014-08-14 07:58:31 <kaniini> systemd goal ultimately is to destroy distributions
2014-08-14 08:01:04 <ncopa> morning
2014-08-14 08:01:48 <kaniini> i should put that a different way.
2014-08-14 08:01:55 <ncopa> i share kaniinis feelings about systemd
2014-08-14 08:02:03 <ncopa> some of the ideas there are good
2014-08-14 08:02:06 <ncopa> i like some of the ideas
2014-08-14 08:02:17 <ncopa> there are some technically good stuff
2014-08-14 08:02:22 <kaniini> systemd (or at least the commercial sponsors of it) goal ultimately is to blur the differences about distributions to the point where redhat and suse are the only ones left standing
2014-08-14 08:02:52 <barthalion> not that distributions were so different before, more or less screwed init scripts, various package managers and philosophies
2014-08-14 08:03:06 <ncopa> but i dont like the management style / politics
2014-08-14 08:03:07 <barthalion> still same glibc and usually the same software, more or less old
2014-08-14 08:03:16 <kaniini> yes, the goal is to blur it to the point where it's literally, where you get your support contract from
2014-08-14 08:03:17 <kaniini> :P
2014-08-14 08:03:58 <barthalion> it's just one element down here, and not having to deal with sysvinit is a big plus
2014-08-14 08:04:06 <barthalion> devouring a lot of stuff is not
2014-08-14 08:04:09 <ncopa> they want blur the different philosophies and want everyone follow whatever philosophy they have this week
2014-08-14 08:06:32 <barthalion> and I could see the point of merging consolekit and so on, but the dark magic you need to use to build udev without systemd… ugh
2014-08-14 08:09:49 <kaniini> i think ultimately i have an idea on how to solve the problem
2014-08-14 08:11:51 <kaniini> the solution is quite obvious really
2014-08-14 08:12:05 <kaniini> we just need to do the same thing with systemd that we did with pkg-config
2014-08-14 08:12:13 <kaniini> build a compatibility framework around it which behaves more sanely
2014-08-14 08:25:18 <shafire> hi
2014-08-14 08:25:31 <shafire> kaniini: do you have a 3.x openvz/virtuozzo template?
2014-08-14 08:26:22 <kaniini> i do not, sorry.
2014-08-14 08:26:42 <kaniini> all you have to do to make one is
2014-08-14 08:26:51 <kaniini> apk add --root somewhere/ --initdb
2014-08-14 08:27:16 <kaniini> apk update --root somewhere/ -X http://nl.alpinelinux.org/alpine/v3.0/main
2014-08-14 08:27:24 <kaniini> apk add --root somewhere/ -X http://nl.alpinelinux.org/alpine/v3.0/main alpine-base
2014-08-14 08:27:30 <kaniini> apk add --root somewhere/ -X http://nl.alpinelinux.org/alpine/v3.0/main alpine-keys
2014-08-14 08:27:38 <kaniini> then tar up the result
2014-08-14 08:27:41 <kaniini> it should work
2014-08-14 08:27:56 <kaniini> i guess add openssh or something if you want that
2014-08-14 08:28:49 <shafire> thanks
2014-08-14 13:55:52 <k0ri0n> hello all
2014-08-14 13:57:57 <k0ri0n> has somebody tried to compile kde on alpine ?
2014-08-14 13:59:15 <barthalion> I afraid not
2014-08-14 13:59:25 <barthalion> but it should be doable
2014-08-14 13:59:38 <barthalion> (as long as it doesn't pull systemd like gnome :c))
2014-08-14 14:00:02 <k0ri0n> i compiled almost full base package without systemd
2014-08-14 14:00:18 <k0ri0n> but had some troubles with kdebase-workspace
2014-08-14 14:00:23 <k0ri0n> TEXTREL problem
2014-08-14 14:01:16 <barthalion> yeah, it has to be fixed in code
2014-08-14 14:01:41 <barthalion> there are some chances that gentoo hardened has a patch
2014-08-14 14:01:56 <k0ri0n> i tried to patch with -fPIC option, but no luck.
2014-08-14 14:02:25 <k0ri0n> already tried to search on gentoo site, but no path has been found....
2014-08-14 14:02:32 <k0ri0n> *patch
2014-08-14 14:03:09 <barthalion> you need to dive in the code then
2014-08-14 14:03:18 <k0ri0n> i tried to use scanelf, but i don't have experience with ASM
2014-08-14 14:03:23 <barthalion> I have no experience in fixing textrels and failed shamefully last time
2014-08-14 14:03:29 <barthalion> http://wiki.gentoo.org/wiki/Hardened/HOWTO_locate_and_fix_textrels this was helpful though
2014-08-14 14:04:13 <k0ri0n> thank you for link, i tried to avoid it, but seems i have to go deeper to code
2014-08-14 14:18:22 <shafire> kaniini: I tried it, but I always get:
2014-08-14 14:18:22 <shafire> ERROR: Unable to lock database: No such file or directory ERROR: Failed to open apk database: No such file or directory
2014-08-14 14:18:42 <shafire> oups wait
2014-08-14 14:20:57 <shafire> localhost:~# apk update --root template/ -X http://nl.alpinelinux.org/alpine/v3.0/main fetch http://nl.alpinelinux.org/alpine/v3.0/main/x86_64/APKINDEX.tar.gz ERROR: http://nl.alpinelinux.org/alpine/v3.0/main: UNTRUSTED signature WARNING: Ignoring APKINDEX.9ac7f32e.tar.gz: No such file or directory OK: 0 distinct packages available
2014-08-14 14:23:42 <shafire> got it
2014-08-14 16:55:08 <shafire> hi
2014-08-14 16:55:27 <shafire> df: /dev/mqueue: No such file or directory <- why do I have this, when it's not specified in /etc/fstab?
2014-08-14 16:56:19 <shafire> when I login over openssh, I get error "PTY allocation request failed on channel 0", because /dev/pts is not mounted, thought it's in /etc/fstab specified
2014-08-14 16:56:20 <shafire> whats wrong?
2014-08-14 17:07:06 <fabian_a> please git pull git://git.alpinelinux.org/fab/aports
2014-08-14 17:07:15 <fabian_a> there are a couple of updates
2014-08-14 17:07:17 <fabian_a> thanks
2014-08-14 17:24:50 <barthalion> fabian_a: thanks
2014-08-14 17:27:03 <barthalion> hm
2014-08-14 17:27:18 <barthalion> git magicians, shouldn't git display me as the commiter?
2014-08-14 18:14:06 <shafire> running alpine 3 on virtuozzo
2014-08-14 18:14:11 <shafire> :-) successfully
2014-08-14 20:11:22 <f15RyWpgGTO6> re systemd: https://lh3.googleusercontent.com/-bZId5j2jREQ/U-vlysklvCI/AAAAAAAACrA/B4JggkVJi38/w426-h284/bd0fb252416206158627fb0b1bff9b4779dca13f.gif
2014-08-14 20:43:39 <shafire> Can I not compile java files with openjdk7?
2014-08-14 20:43:45 <shafire> there is no javac there...
2014-08-14 20:43:53 <fabian_a> ccccccddchcgjlltrcrknfcjddiduinihbhhriburdkr
2014-08-14 20:44:19 <shafire> ?
2014-08-14 20:45:03 <kaniini> openjdk7-jre maybe
2014-08-14 20:45:13 <fabian_a> sorry
2014-08-14 20:45:21 <kaniini> hmm, weird
2014-08-14 20:45:51 <shafire> kaniini: jre is only the runtime, I can run jar files, but I can not compile any java files :(
2014-08-14 20:46:07 <kaniini> hmmph
2014-08-14 20:46:30 <Meths> fabian_a: yubikey?
2014-08-14 20:46:53 <fabian_a> Meths: yes
2014-08-14 20:48:07 <fabian_a> not a good idea to walk around with a connected yubikey
2014-08-15 04:39:51 <shafire> hi
2014-08-15 04:40:57 <shafire> I think, the openjdk-devel package is missing
2014-08-15 04:42:41 <shafire> https://www.archlinux.org/packages/?sort=&q=openjdk&maintainer=&flagged= <- has jdk7-openjdk jre7-openjdk
2014-08-15 04:42:48 <shafire> and -headless
2014-08-15 04:46:39 <shafire> I think, we should pack like arch too
2014-08-15 04:47:00 <shafire> ubuntu does the same: http://wiki.ubuntuusers.de/Java/Installation#OpenJDK-7
2014-08-15 09:44:12 <clandmeter> fabled: your latest nfs commit, should it resolve most issues?
2014-08-15 10:03:25 <fabled> clandmeter, well, we did have alpine as nfs server, and ubuntu was able to mount a share.
2014-08-15 10:03:31 <fabled> but it's probably not perfect yet
2014-08-15 10:03:49 <ncopa> idmap does not work
2014-08-15 10:04:16 <ncopa> i think rpcbind segfaults too once in a while, but i am not able to reliable reproduce that
2014-08-15 10:04:41 <ncopa> infact i'm not able to reproduece at all when i try
2014-08-15 10:15:30 <clandmeter> i think ill make a backup of my tmpfs install and upgrade to musl and check to see if it works.
2014-08-15 10:48:33 <clandmeter> zenny: ^
2014-08-15 10:55:24 <clandmeter> f15RyWpgGTO6: is zfs currently working on musl?
2014-08-15 10:55:56 <f15RyWpgGTO6> no
2014-08-15 10:56:02 <f15RyWpgGTO6> i'm still fiddling on a new patch
2014-08-15 10:56:06 <f15RyWpgGTO6> i got some contributions
2014-08-15 10:56:19 <f15RyWpgGTO6> and the zfs guys cherrypicked some of my patches
2014-08-15 10:56:44 <f15RyWpgGTO6> i have two enthusiastic contributors sending me improvements to my patchset
2014-08-15 10:56:58 <clandmeter> nice
2014-08-15 10:57:02 <f15RyWpgGTO6> i'll try to make a releaseable version in the next 1-2 days
2014-08-15 10:57:17 <clandmeter> there is somebody in alpine-linux asking about it
2014-08-15 10:57:18 <f15RyWpgGTO6> but i have trouble testing. my vm is damn slow dunno why
2014-08-15 10:58:05 <f15RyWpgGTO6> makes test-compilations very demotivating.
2014-08-15 10:58:29 <clandmeter> i can provide some infra if needed.
2014-08-15 10:58:38 <f15RyWpgGTO6> it's like the xkcd, where dev is slacking and saying, it's compiling. but in my case it's that but in slomo
2014-08-15 11:00:39 <f15RyWpgGTO6> thanks alot clandmeter, i might come back to you, but i think i should rather investigate why my vms are degraded and fix them.
2014-08-15 11:00:58 <clandmeter> most proabaly :)
2014-08-15 11:10:00 <zenny> clandmeter: here I am ;-)
2014-08-15 11:10:56 <f15RyWpgGTO6> zenny: i summarized the zfs situation here earlier, but now also on alpine-linux
2014-08-15 11:11:42 <clandmeter> looks like more and more linux users are jumping on the zfs-wagon
2014-08-15 11:13:54 <zenny> clandmeter: with openzfs project going, more are interested in.
2014-08-15 11:15:01 <f15RyWpgGTO6> openzfs? is that separate from zfsonlinux?
2014-08-15 11:15:36 <JT-EC> It's the umbrella helping coordinate between illumos/FBSD/ZoL devs
2014-08-15 11:15:52 <shafire> ncopa: you can throw away flashcache, I am using bcache now
2014-08-15 11:16:00 <shafire> ncopa: maybe put zfs in :o
2014-08-15 11:25:28 <zenny>  f15RyWpgGTO6: It is a collective effort which uses v5000 of zfs. open-zfs.org (note the hyphen between the words).
2014-08-15 12:24:42 <shafire> can someone open #2224 please
2014-08-15 12:37:39 <fabled> shafire, commented #2224. if the -jdk split is desirable, please file a new bug.
2014-08-15 12:38:20 <shafire> mh
2014-08-15 12:38:22 <ncopa> shafire: export PATH=$PATH:/usr/lib/jvm/java-1.7-openjdk/bin
2014-08-15 12:39:54 <fabled> ncopa, not needed
2014-08-15 12:40:06 <fabled> or
2014-08-15 12:40:08 <fabled> hum
2014-08-15 12:40:11 <fabled> maybe it is
2014-08-15 12:40:24 <fabled> java-common maybe ships 'java' only
2014-08-15 12:43:00 <ncopa> sigh
2014-08-15 12:43:12 <ncopa> lighdm does not work with consolekit apparently
2014-08-15 13:10:02 <shafire> fabled: you only link some java files... heh http://git.alpinelinux.org/cgit/aports/tree/main/java-common/APKBUILD?id=6217d4883d8ef4db4665e119bb5d49066a1d285d
2014-08-15 13:10:32 <fabled> yeah, seems that way
2014-08-15 13:10:57 <shafire> did you already try to compile openjdk8?
2014-08-15 13:11:08 <shafire> I would like to have a headless version without the webapplet/gui stuff
2014-08-15 14:38:47 <shafire> I want to try to build openjdk8 tomorrow
2014-08-15 14:38:54 <shafire> clandmeter: can I get infra too? :)
2014-08-15 14:45:04 <clandmeter> shafire: didnt i gave you access already?
2014-08-15 14:45:26 <shafire> if yes, than I don't know how to use
2014-08-15 14:45:32 <clandmeter> oh no
2014-08-15 14:45:36 <clandmeter> it was somebody else
2014-08-15 14:46:47 <clandmeter> shafire: which repo?
2014-08-15 14:48:20 <clandmeter> i guess edge is ok?
2014-08-16 23:26:36 <f15RyWpgGTO6> i updated spl-grsec and zfs-grsec to work against 3.14.17: https://github.com/stef/aports also i updated my patchset against zfsonlinux:master and sent a pull request: https://github.com/zfsonlinux/zfs/pull/2604
2014-08-17 12:51:13 <barthalion> f15RyWpgGTO6: good job
2014-08-17 12:58:44 <f15RyWpgGTO6> meh apparently not: http://buildbot.zfsonlinux.org/console?branch=refs/pull/2604/head
2014-08-17 12:59:41 <barthalion> well, each one is about style, not actual build errors
2014-08-17 13:01:32 <f15RyWpgGTO6> nah, not only style :/
2014-08-17 13:02:31 <Meths> Is that auto builds from the pull request?  How are you supposed to fix the Fed20 dep in a pull request?
2014-08-17 13:04:01 <barthalion> well, not your fault that buildbot is missing dependencies
2014-08-17 13:04:33 <f15RyWpgGTO6> yes, but the __asssert fix does seem to cause problems
2014-08-17 13:04:48 <f15RyWpgGTO6> http://buildbot.zfsonlinux.org/builders/centos-6.5-x86_64-builder/builds/1209/steps/shell_4/logs/stdio
2014-08-17 13:05:16 <barthalion> ah, I see
2014-08-17 14:02:18 <shafire> where do you put single files packages?
2014-08-17 14:02:25 <shafire> into /usr/bin directly?
2014-08-17 14:03:12 <shafire> I want to pack https://code.google.com/p/tb-tun/
2014-08-17 14:06:50 <barthalion> yeah, /usr/bin
2014-08-17 15:13:09 <shafire> barthalion: when do I know if I need to move it in /usr/bin, /usr/sbin or /sbin
2014-08-17 15:13:58 <barthalion> the rule of thumb is to put executables requiring root access to /usr/sbin
2014-08-17 15:14:16 <barthalion> or /sbin, if the tool is crucial for boot and general usage
2014-08-17 15:14:35 <barthalion> it doesn't really matter unless you're a purist
2014-08-17 15:14:39 <barthalion> just put it in /usr/bin
2014-08-18 05:55:34 <shafire> morning
2014-08-18 06:30:55 <ncopa> morning
2014-08-18 08:30:04 <shafire> http://git.alpinelinux.org/cgit/aports/tree/main/cmph/APKBUILD?id=1fca3de10730ab711d2b8151b3116fdffd491eae
2014-08-18 08:30:09 <shafire> is _builddir here okay?
2014-08-18 08:30:35 <barthalion> yeah, why not?
2014-08-18 08:30:53 <barthalion> $pkgname and $pkgver doesn't contain spaces
2014-08-18 08:31:04 <barthalion> and I think abuild sanity check would find out anyway
2014-08-18 08:31:24 <barthalion> so you can quote just $srcdir
2014-08-18 08:31:34 <barthalion> or whole string, whathever looks better to you
2014-08-18 08:31:58 <barthalion> same about curly brackets
2014-08-18 08:35:58 <shafire> OK
2014-08-18 08:36:41 <barthalion> rule of thumb is to quote anything that can break with whitespice inside
2014-08-18 08:36:46 <barthalion> whitespace even
2014-08-18 08:51:35 <shafire> $pkgname/$pkgname_$pkgver.tar.gz <- do I need to escape _?
2014-08-18 08:52:35 <barthalion> not espace, this is where you need curly brackets
2014-08-18 08:52:40 <barthalion> as _ can be used in variable name
2014-08-18 08:52:56 <barthalion> so it will evaluate as variable $pkgname_
2014-08-18 08:53:11 <barthalion> use ${pkgname}_${pkgver}
2014-08-18 08:53:30 <shafire> OK
2014-08-18 09:00:21 <shafire> what can I do, when it is not unpacked in its own dir?
2014-08-18 09:00:35 <shafire> delete cd "$_builddir"?
2014-08-18 09:01:19 <shafire> ah
2014-08-18 09:07:02 <clandmeter> shafire: you can adjust builddir to whatever dir the archive unpacks to.
2014-08-18 09:08:19 <clandmeter> most sane projects unpack to pkgname-pkgver
2014-08-18 09:13:35 <shafire> could you check my apkbuild, I don't understand the error
2014-08-18 09:14:20 <clandmeter> sure
2014-08-18 09:14:26 <shafire> http://pastie.org/private/j8e3jakpgbbo75wulvosa
2014-08-18 09:14:54 <clandmeter> can you show me the contents of src dir?
2014-08-18 09:15:41 <shafire> $ ls src/ tb-tun_r18.tar.gz  tb_userspace       tb_userspace.c
2014-08-18 09:15:41 <clandmeter> ls tb-tun/src |sprunge
2014-08-18 09:16:53 <shafire> http://sprunge.us/UAZZ
2014-08-18 09:18:33 <clandmeter> you need to move it to $pkgdir/usr/bin
2014-08-18 09:20:35 <shafire> now it says mv: can't rename 'tb_userspace': No such file or directory
2014-08-18 09:20:58 <shafire> I changed the line to mv tb_userspace "$pkgdir/usr/sbin" || return 1
2014-08-18 09:22:25 <clandmeter> mkdir $pkgdir
2014-08-18 09:22:31 <clandmeter> or use install
2014-08-18 09:28:16 <clandmeter> hmm thats freaky, my /usr/bin/install was from coreutils even it was not installed.
2014-08-18 09:29:04 <clandmeter> shafire: i guess it works now?
2014-08-18 09:29:47 <shafire> with install
2014-08-18 09:29:57 <shafire> but need to check if my package is "correct"
2014-08-18 09:30:59 <clandmeter> what could be incorrect?
2014-08-18 09:32:05 <shafire> I installed the package and I can not find the file :D
2014-08-18 09:39:04 <shafire> I found it
2014-08-18 09:39:11 <shafire> it installed as sbin.apk-new
2014-08-18 09:39:12 <shafire> lol
2014-08-18 09:44:25 <shafire> now :-)
2014-08-18 09:48:47 <shafire> http://sprunge.us/jcQL
2014-08-18 09:50:00 <shafire> git: 'send-email' is not a git command. See 'git --help'.
2014-08-18 09:50:23 <shafire> is this a plugin?
2014-08-18 09:51:23 <clandmeter> apk search git |grep mail
2014-08-18 09:52:15 <shafire> I looked already, its empty
2014-08-18 09:52:35 <clandmeter> what is empty?
2014-08-18 09:53:34 <shafire> apk search git |grep mail :)
2014-08-18 09:53:41 <shafire> shows nothing
2014-08-18 09:53:44 <shafire> there is only git
2014-08-18 09:53:58 <clandmeter> apk search git |grep mail
2014-08-18 09:53:59 <clandmeter> git-email-2.0.4-r0
2014-08-18 09:54:21 <shafire> mh
2014-08-18 09:54:55 <ncopa> i think it earlier was shipped with git-perl
2014-08-18 09:55:15 <ncopa> so with older alpine its probably apk add git-perl
2014-08-18 09:59:42 <clandmeter> [aports-police] submitting patches to edge which are confirmed building on something different then edge is not allowed! [/aports-police]
2014-08-18 10:02:32 <clandmeter> shafire: looks like the infa box i gave you is 2.7...
2014-08-18 10:02:41 <shafire> yes
2014-08-18 10:03:07 <clandmeter> i forgot to add edge switch
2014-08-18 10:03:27 <clandmeter> but you can upgrade it if you like
2014-08-18 10:03:46 <shafire> yeah, please
2014-08-18 10:03:59 <shafire> but don't loose my keys in /home/dp
2014-08-18 10:04:08 <clandmeter> you can do it
2014-08-18 10:04:46 <shafire> ok
2014-08-18 10:04:50 <shafire> I read wrong, heh
2014-08-18 10:05:34 <clandmeter> just follow http://wiki.alpinelinux.org/wiki/Upgrading_Alpine
2014-08-18 10:05:46 <shafire> I wanted to send it by email: Died at /usr/libexec/git-core/git-send-email line 1207, <FIN> line 1.
2014-08-18 10:06:41 <clandmeter> but dont forget to exchange 3.0 with edge :)
2014-08-18 10:12:48 <shafire> :( I did something wrong. I upgraded as wiki said. simulation was okay...
2014-08-18 10:13:23 <shafire> I have many errors in musl-dev: (29/55) Installing musl-dev (1.1.4-r3) ERROR: musl-dev-1.1.4-r3: trying to overwrite usr/lib/libdl.a owned by uclibc-dev-0.9.33.2-r27.
2014-08-18 10:13:51 <clandmeter> remove uclibc-dev
2014-08-18 10:15:52 <shafire> I think, it's okay now
2014-08-18 10:16:36 <clandmeter> looks like it
2014-08-18 10:17:56 <shafire> I can't send mails from this box
2014-08-18 10:18:05 <shafire> sendmail: can't connect to remote host (127.0.0.1): Connection refused Died at /usr/libexec/git-core/git-send-email line 1227, <FIN> line 1.
2014-08-18 10:18:59 <clandmeter> you need to specify an smtp server i guess
2014-08-18 10:29:16 <shafire> do you know where? I can not find any config file for sendmail
2014-08-18 10:31:21 <clandmeter> http://git-scm.com/docs/git-send-email
2014-08-18 10:31:52 <shafire> sry
2014-08-18 10:32:58 <clandmeter> shafire: you can also install ssmtp
2014-08-18 10:33:07 <clandmeter> and configure it to use something like gmail smtp
2014-08-18 10:51:41 <shafire> i think now, finall
2014-08-18 10:51:44 <shafire> y
2014-08-18 13:16:13 <shafire> Can I add some modules in nginx?
2014-08-18 13:17:07 <fabled> shafire, internal or external?
2014-08-18 13:17:23 <fabled> internal should be ok. if it needs patching or additional source downloads, then it depends a bit.
2014-08-18 13:17:31 <fabled> i ended up making separate ngingx-lua
2014-08-18 13:17:51 <shafire> external...
2014-08-18 13:17:54 <shafire> https://code.google.com/p/modpagespeed/
2014-08-18 13:18:11 <shafire> oups wrong link
2014-08-18 13:18:18 <shafire> http://ngxpagespeed.com/ngx_pagespeed_example/
2014-08-18 13:22:07 <uggedal> that module is quite large and c++
2014-08-18 13:25:59 <shafire> and the second one is https://github.com/nginx-clojure/nginx-clojure
2014-08-18 13:26:56 <fabled> would be nice if nginx had dynamic modules support
2014-08-18 13:27:08 <fabled> but now all that is built statically
2014-08-18 13:27:33 <shafire> I see
2014-08-18 13:31:21 <shafire> http://tengine.taobao.org/document/dso.html
2014-08-18 13:32:09 <fabled> yes, tengine is an interesting fork
2014-08-18 13:32:35 <fabled> upstream nginx does not support that. it's tengine feature.
2014-08-18 13:33:41 <fabled> ncopa, i wonder if we should go with tengine instead and make ngx-lua and other stuff as dso modules
2014-08-18 13:34:11 <shafire> that would be good for me two
2014-08-18 13:34:13 <shafire> too
2014-08-18 13:35:23 <shafire> how long do you know tengine, fabled?
2014-08-18 13:35:47 <fabled> have not used. but bumped into it when doing ngingx-lua package.
2014-08-18 13:35:49 <ncopa> fabled: i think that sounds like a good idea
2014-08-18 13:35:51 <fabled> exactly on this subject
2014-08-18 13:36:00 <fabled> it has dso support, that i very much wanted for ngingx-lua
2014-08-18 13:37:19 <shafire> would you go with 1.x or 2.0?
2014-08-18 13:37:47 <fabled> as said, i never tried tengine
2014-08-18 13:38:37 <shafire> the latest (2.0.3) is with nginx 1.4.7 compatible (Change: merged changes between nginx-1.4.6 and nginx-1.4.7. (chobits))
2014-08-18 13:40:44 <shafire> https://github.com/pagespeed/ngx_pagespeed/wiki/Using-ngx_pagespeed-with-Tengine <-- looks good
2014-08-18 16:28:24 <shafire> oh, we have already tengine
2014-08-18 16:29:08 <shafire> should I create a new aports for tengine modules or include it in tengine apkbuild?
2014-08-18 16:39:31 <fabled> preferably separate module aports if possible
2014-08-18 17:31:24 <shafire> is alpine more like debian or ubuntu? in terms of /etc/network/interfaces and so on
2014-08-18 18:40:46 <uggedal> afaik ifup/ifdown from bb is based on debian
2014-08-18 18:41:10 <uggedal> (have no clue how ubuntu does it though)
2014-08-18 19:48:14 <f15RyWpgGTO6> ohshit: "libtirpc is GPL licensed. We cannot link to it. We are going to need to port the Illumos Transport-Independent RPC library."
2014-08-18 19:52:22 <shafire> whats that?
2014-08-19 03:43:13 <vkrishn> kaniini, around
2014-08-19 03:58:34 <vkrishn> I remember you mentioning packaging 'kmscon', was it added to aports ?
2014-08-19 07:49:47 <ncopa> fabled: i am checking why some packages unexpectedly are trying to get built on arm builder
2014-08-19 07:49:52 <ncopa> it seems that
2014-08-19 07:49:57 <ncopa> options="!libc_musl"
2014-08-19 07:50:19 <ncopa> does not match the 'muslgnueabihf' used on arm builder
2014-08-19 07:51:05 <barthalion> they want more attention
2014-08-19 07:51:19 <barthalion> and we need a tool for behavioral analysis of packages :c)
2014-08-19 07:51:30 <ncopa> what is the proper fix? 1) change the aports to:  options="!libc_musl !libc_muslgnueabihf"
2014-08-19 07:51:58 <fabled> ncopa, huh?
2014-08-19 07:52:01 <ncopa> implement some wildcard, eg. options="!libc_musl*"
2014-08-19 07:52:02 <fabled> i'd fix abuild
2014-08-19 07:52:26 <fabled> the configuration triplet last part != c-library name
2014-08-19 07:52:42 <fabled> abuild should map muslgnueabihf to libc_musl
2014-08-19 07:52:42 <ncopa> ok
2014-08-19 07:52:55 <fabled> gnueabihf == use hard-float ABI
2014-08-19 07:53:05 <ncopa> so how do we detect libc?
2014-08-19 07:53:17 <fabled> it can be used to detect it
2014-08-19 07:53:20 <fabled> but it needs mapping
2014-08-19 07:53:26 <fabled> musl* -> musl
2014-08-19 07:53:30 <fabled> uclibc* -> uclibc
2014-08-19 07:53:31 <fabled> etc
2014-08-19 07:53:50 <ncopa> so we need a static list of known libcs?
2014-08-19 07:53:59 <fabled> well
2014-08-19 07:54:08 <fabled> we only support musl (+ previously uclibc)
2014-08-19 07:54:25 <fabled> might be nice to keep glibc there too, but i don't think it's a must to do that
2014-08-19 07:54:50 <ncopa> what if some else wants to play with bsd kernel and use aports for bsd?
2014-08-19 07:55:16 <ncopa> they'd need add the bsdlibc to the list?
2014-08-19 07:55:31 <ncopa> should we keep that list in /etc/abuild.conf?
2014-08-19 07:55:47 <ncopa> i'd prefer not hardcode it
2014-08-19 07:56:10 <ncopa> since there are various scripts needing it, shell scripts that parses APKBUILDs, lua scripts, etc
2014-08-19 07:57:26 <ncopa> we could have a KNOWN_LIBCS variable that is default "musl uclibc glibc"
2014-08-19 07:57:36 <ncopa> so you could override it in /etc/abuild.conf
2014-08-19 07:59:54 <fabled> well, i would not go there until we need it. but sure.
2014-08-19 08:00:19 <ncopa> hum
2014-08-19 08:00:27 <ncopa> its a triplet
2014-08-19 08:00:34 <ncopa> machine-vendor-operatingsystem
2014-08-19 08:01:01 <ncopa> linux-gnu
2014-08-19 08:01:07 <ncopa> is for glibc
2014-08-19 08:01:26 <shafire> Can I install glibc parallel to musl?
2014-08-19 08:01:31 <ncopa> so for disabling glibc we should use options="!libc_gnu"
2014-08-19 08:01:54 <fabled> ncopa, libc name is not detectable reliably from the triplet
2014-08-19 08:02:11 <fabled> we really need "triplet" -> { arch, libc } mapping table
2014-08-19 08:02:36 <fabled> and often, the triplet is faked in some places
2014-08-19 08:03:03 <ncopa> shafire: in theory yes, but you need do special things if you want both libc headers installed in parallel
2014-08-19 08:03:33 <fabled> well
2014-08-19 08:03:39 <ncopa> http://wiki.osdev.org/Target_Triplet
2014-08-19 08:03:47 <fabled> you end up in trouble after building different libs against different libcs
2014-08-19 08:04:03 <ncopa> there is no way to use the target triplet in a sane way
2014-08-19 08:04:11 <ncopa> x86_64-linux-gnu
2014-08-19 08:04:12 <fabled> ncopa, correct. and the wiki page you refer is outdated
2014-08-19 08:04:46 <ncopa> do you have a more updated doc on the subject?
2014-08-19 08:07:58 <ncopa> what debian does: https://wiki.debian.org/Multiarch/Tuples
2014-08-19 08:12:10 <fabled> right. that's actually fairly complete. wrt, glibc
2014-08-19 08:12:32 <ncopa> but we use a vendor too
2014-08-19 08:12:47 <fabled> and annoyingly it's 'muslgnueabihf' for us, and not just 'musleabihf'. iirc, gnueabihf was hard coded in multiple gcc configure places
2014-08-19 08:13:21 <ncopa> that sounds like a gcc bug
2014-08-19 08:13:35 <ncopa> musl != gnu
2014-08-19 08:29:57 <ncopa> debian's multiarch thingy is kinda nice
2014-08-19 08:30:06 <ncopa> https://www.mail-archive.com/systemd-devel@lists.freedesktop.org/msg20806.html
2014-08-19 09:00:44 <shafire> hi
2014-08-19 09:00:51 <ncopa> hi
2014-08-19 09:00:55 <shafire> could someone update coreutils to the newest version?
2014-08-19 09:01:04 <shafire> maybe there is the users bug fixed
2014-08-19 09:01:57 <ncopa> which is the latest version? 8.24?
2014-08-19 09:02:21 <ncopa> its 8.23
2014-08-19 09:02:26 <ncopa> we have it in edge already
2014-08-19 09:02:46 <shafire> I see, okay
2014-08-19 09:02:55 <ncopa> do you have a specific bug you want fixed?
2014-08-19 09:03:29 <ncopa> i would prefer not update stable branches unless there are a specific issue you want fix
2014-08-19 09:04:31 <shafire> # users users: /dev/null/utmp: Not a directory
2014-08-19 09:04:48 <shafire> users from coreutils do not work
2014-08-19 09:05:02 <shafire> does*
2014-08-19 09:05:06 <ncopa> that is not fixed in latest coreutils
2014-08-19 09:05:15 <ncopa> and i think it requires some patching
2014-08-19 09:11:25 <f15RyWpgGTO6> re zfs-grsec, my patchset has been scrutinized: https://github.com/zfsonlinux/zfs/pull/2604
2014-08-19 09:11:49 <f15RyWpgGTO6> and a repost with bad news: 21:47:38    f15RyWpgGTO6 | ohshit: "libtirpc is GPL licensed. We cannot link to it. We are going to need to port the Illumos Transport-Independent RPC library."
2014-08-19 09:12:20 <f15RyWpgGTO6> unfortunately i'm participating at an event this week and will have no time to respond.
2014-08-19 09:13:53 <clandmeter> f15RyWpgGTO6: isnt that the same reason it cant be mainline?
2014-08-19 09:14:31 <f15RyWpgGTO6> oh you mean zfs itself?
2014-08-19 09:14:39 <clandmeter> yes
2014-08-19 09:14:42 <barthalion> yeah
2014-08-19 09:14:48 <clandmeter> compared with rpc
2014-08-19 09:24:03 <ncopa> f15RyWpgGTO6: i think you have gotten pretty good feedback on your pull request
2014-08-19 09:24:15 <ncopa> some of the commits needs fixing
2014-08-19 09:24:49 <ncopa> and it seems that upstream is interested in fixing them properly
2014-08-19 09:25:00 <f15RyWpgGTO6> yes! i'm very glad myself
2014-08-19 09:25:15 <f15RyWpgGTO6> the only problem is i'm stressed out because of the event i'm having.
2014-08-19 09:25:20 <ncopa> :)
2014-08-19 09:25:41 <ncopa> i dont think they expect immediate attention
2014-08-19 11:56:04 <ncopa> I think I'll mess a bit with the build servers
2014-08-19 11:56:23 <ncopa> i will replace the /etc/init.d/aports-build init script with /etc/init.d/mqtt-exec.aports-build
2014-08-19 11:56:37 <ncopa> which is a more general purpose
2014-08-19 12:12:47 <ncopa> build servers are currently broken
2014-08-19 12:12:59 <ncopa> i'm working on fixing them...
2014-08-19 12:33:16 <ncopa> build servers should be online again
2014-08-19 14:25:58 <shafire> backports in 2.7 is quite new, isn't it?
2014-08-19 14:28:38 <ncopa> yeah
2014-08-19 14:28:40 <ncopa> and i hate it :)
2014-08-19 14:38:34 <shafire> acutally I can also use lbu on alpine hard drive installs
2014-08-19 14:39:40 <shafire> "lbu st -a" shows me all files that have been modified
2014-08-19 14:39:50 <shafire> pretty cool to find garbage
2014-08-19 16:05:40 <shafire> http://alpinelinux.org/forum/general-discussion/running-alpine-ram-installed-hard-drive#comment-170
2014-08-19 16:05:45 <shafire> will my solution work or not?
2014-08-19 19:52:16 <kaniini> 04:17:11    f15RyWpgGTO6 | and a repost with bad news: 21:47:38    f15RyWpgGTO6 | ohshit: "libtirpc is GPL licensed. We cannot link to it. We are going to need to port the Illumos Transport-Independent RPC library."
2014-08-19 19:52:17 <kaniini> ouch
2014-08-19 19:52:27 <kaniini> that is indeed bad news
2014-08-20 05:06:02 <shafire> morning
2014-08-20 05:06:26 <shafire> I build a package/module for tengine, but I do not know if I did it right
2014-08-20 05:41:42 <shafire> ok, tengine dso modules are saved in /usr/modules
2014-08-20 05:45:32 <shafire> looks like we should move it to /usr/lib/tengine?
2014-08-20 05:59:18 <fabled> shafire, yes, i'd probably go with that. make sure tengine main binary looks for the modules there too
2014-08-20 06:02:01 <shafire> should I create extra packages for the modules?
2014-08-20 06:03:24 <shafire> but its too much: http://pastie.org/private/uwli6xv4suiolgxpumv5iq
2014-08-20 06:06:23 <shafire> Ok, some are build anyway
2014-08-20 08:01:49 <uggedal> "A new library systemd-terminal has been added that implements full TTY stream parsing and rendering. This library is supposed to be used later on for implementing a full userspace VT subsystem, replacing the  current kernel implementation." http://lists.freedesktop.org/archives/systemd-devel/2014-August/022295.html
2014-08-20 08:05:23 <ncopa> tty in systemd...
2014-08-20 08:05:48 <ncopa> i suppose that means you will not be able to get a tty without systemd in future?
2014-08-20 08:12:43 <fabled> i wonder when they implement a VM in systemd to run rest of the OS inside it
2014-08-20 08:16:33 <f15RyWpgGTO6> i doubt the kernel devs will allow to be replaced
2014-08-20 08:34:05 <StarWarsFan> hi@all
2014-08-20 08:34:14 <StarWarsFan> any news regarding the nfs-problem?
2014-08-20 08:40:07 <clandmeter> StarWarsFan: did you try edge?
2014-08-20 08:44:54 <StarWarsFan> actually not, using v3.0
2014-08-20 08:45:06 <StarWarsFan> are there changes on edge?
2014-08-20 08:48:16 <ncopa> i think i pushed the nfs changes to 3.0-stable too
2014-08-20 08:48:38 <ncopa> yeah
2014-08-20 08:48:50 <ncopa> nfs fixes should be in 3.0-stable too
2014-08-20 08:49:07 <ncopa> i think there might be one issue with rpcbind still
2014-08-20 08:49:12 <shafire>  /dev/shm is normally 777, or?
2014-08-20 08:49:34 <ncopa> shafire: i think so yes
2014-08-20 08:49:48 <StarWarsFan> ncopa: nfs still not working on 64bit...
2014-08-20 08:53:17 <StarWarsFan> using nfs-utils-1.3.1_rc3-r0
2014-08-20 08:53:29 <ncopa> that should be the same fixes as in edge
2014-08-20 08:53:35 <ncopa> what is not working?
2014-08-20 08:53:41 <ncopa> server or client?
2014-08-20 08:53:53 <ncopa> i think I only tried server
2014-08-20 08:54:04 <ncopa> but once i experienced rpcbind segfault
2014-08-20 08:54:08 <ncopa> when retry it worked
2014-08-20 08:54:23 <ncopa> and i have not tried nfs server inside a lxc
2014-08-20 08:56:21 <StarWarsFan> client
2014-08-20 08:56:43 <StarWarsFan> trying to use an nfs-share on an ubuntu server
2014-08-20 08:57:19 <StarWarsFan> so nfs-client on 64bit system is not working
2014-08-20 08:57:40 <StarWarsFan> but on this machien on an 32bit lxc container it works
2014-08-20 08:57:52 <StarWarsFan> so host is alpine 3.0 64bit
2014-08-20 08:58:03 <StarWarsFan> and lxc-container is alpine 3.0 32bit
2014-08-20 09:18:31 <shafire> third_party/chromium/src/base/debug/stack_trace_posix.cc:8:22: fatal error: execinfo.h: No such file or directory
2014-08-20 09:18:36 <shafire> is this glibc specific?
2014-08-20 09:18:49 <ncopa> shafire: yes
2014-08-20 09:19:23 <ncopa> StarWarsFan: and it is rpcbind that segfaults?
2014-08-20 09:22:09 <StarWarsFan> ncopa: http://sprunge.us/DdaZ
2014-08-20 09:22:16 <StarWarsFan> dmesg output
2014-08-20 09:27:33 <shafire> what can I do against that?
2014-08-20 09:34:37 <ncopa> #ifdef __GLIBC__
2014-08-20 09:34:44 <ncopa> #include <execinfo.h>
2014-08-20 09:34:47 <ncopa> #endif
2014-08-20 09:35:42 <ncopa> or maybe __GLIBCXX__
2014-08-20 09:37:03 <ncopa> and you need fix the file to fallback to not do any backtrace at all
2014-08-20 09:39:24 <ncopa> shafire what happens if you dont try mount it from within a container?
2014-08-20 09:39:53 <ncopa> [    8.105633] rpcbind[1073]: segfault at d6ee07b0 ip 00006b52d6ed15cd sp 00007833284c3228 error 4 in ld-musl-x86_64.so.1[6b52d6e82000+85000]
2014-08-20 09:39:53 <ncopa> [    8.105658] grsec: Segmentation fault occurred at 00000000d6ee07b0 in /sbin/rpcbind[rpcbind:1073] uid/euid:103/103 gid/egid:103/103, parent /bin/busybox[init:1]
2014-08-20 09:40:01 <ncopa> that is what we need to fix
2014-08-20 09:40:02 <ncopa> i think
2014-08-20 09:40:04 <shafire> StarWarsFan: ^
2014-08-20 09:41:54 <StarWarsFan> the link above shows dmesg output from the 64bit-host
2014-08-20 09:42:06 <StarWarsFan> not from the 32bit-container
2014-08-20 09:42:41 <StarWarsFan> but to complete the whole picture: the 64bit host itself is an xen-domU on a ubuntu xen host
2014-08-20 10:10:15 <ncopa> fabled: will apk add --root /foo ... use the apk cache relative /foo ?
2014-08-20 10:10:57 <shafire> I fixed tengine today :-)
2014-08-20 10:10:59 <ncopa> or more like: if i have apk add --root /foo ... && apk add --root /bar
2014-08-20 10:11:22 <ncopa> how do i get them to share the cache so we avoid downloading double?
2014-08-20 10:55:52 <shafire> how can I check, if a library (.a) uses glibc or not?
2014-08-20 11:33:47 <shafire> alpine is the only distro that runs from tmpfs (ram) and ships something like lbu?
2014-08-20 11:37:05 <fabled> shafire, LEAF might be similar
2014-08-20 12:36:30 <manjaro_user> hi
2014-08-20 12:40:01 <manjaro_user> which kernel is to use for alpine xen to create alpine pv domu
2014-08-20 12:41:46 <manjaro_user> how igot these files? location?
2014-08-20 12:46:55 <manjaro_user> to create a hvm domu xl want xenbr0. i have only br0. what can i do?
2014-08-20 12:53:41 <clandmeter> manjaro_user: http://nl.alpinelinux.org/alpine/v2.7/releases/x86_64/alpine-xen-2.7.9-x86_64.iso is the latest afaik
2014-08-20 12:55:27 <manjaro_user> thank you
2014-08-20 13:07:53 <shafire> http://alpinelinux.org/forum/installation/asterisk-erro
2014-08-20 13:15:13 <clandmeter> shafire: i think he followed the wiki
2014-08-20 13:15:20 <clandmeter> to install freepbx
2014-08-20 13:17:32 <clandmeter> he probably is using stable repo, which doesnt have mysql addon.
2014-08-20 13:17:42 <clandmeter> i think someobdy removed it in time
2014-08-20 13:18:37 <clandmeter> i think freepbx ppl use odbc driver now.
2014-08-20 14:17:43 <ncopa> i think we have an alpine-xen-3.0
2014-08-20 14:17:50 <ncopa> but the hvmloader is broken
2014-08-20 14:18:44 <ncopa> i need spank or encourage kaniini to fix it, or do it myself...
2014-08-20 14:23:37 <clandmeter> i would try to spank first
2014-08-20 14:26:33 <ncopa> clandmeter: good if you do it - i feel bad to spank someone to do unpaid work :)
2014-08-20 14:28:23 <clandmeter> you suggested it first!
2014-08-20 14:37:40 <ncopa> :)
2014-08-20 14:39:07 <ncopa> kaniini: have you been able to look at building xen hvmloader with musl? i think many people would be happy when it gets fixed ;)
2014-08-20 16:10:27 <shafire> mh, what can I do, when the compiler does not find the apr-dev header files?
2014-08-20 16:10:41 <shafire> they are in /usr/include/apr-1
2014-08-20 16:10:57 <shafire> but it looks like it wants to have it in /usr/include
2014-08-20 16:48:21 <shafire> ngx_pagespeed error: http://pastie.org/private/hdga8obrytmcrfauywt3fq
2014-08-20 16:48:31 <shafire> are these errors in libpng-dev?
2014-08-20 21:25:55 <kaniini> ncopa: well at my current job we use vmware instead of xen
2014-08-20 21:25:56 <kaniini> lol
2014-08-21 05:21:36 <shafire> hi
2014-08-21 05:21:41 <shafire> I got this error: pagespeed/kernel/thread/pthread_rw_lock.cc:36:33: error: 'PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP' was not declared in this scope                                  PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP);
2014-08-21 05:21:43 <shafire> what is this?
2014-08-21 05:22:59 <shafire> I have pthread-libs installed
2014-08-21 05:23:00 <fabled> it's some non-portable pthreads extensions
2014-08-21 05:23:10 <fabled> that's what _NP means, Non Portable
2014-08-21 05:23:22 <fabled> generally those can be just #ifdeffed out
2014-08-21 05:23:28 <fabled> #ifdef PTHREAD_RWLOCK_...
2014-08-21 05:23:50 <shafire> ok
2014-08-21 05:33:16 <shafire> thanks
2014-08-21 05:49:52 <ncopa> morning
2014-08-21 05:50:15 <ncopa> somethimes there are some PTHERAD_something_NP that has got standardized to PTHREAD_something
2014-08-21 05:50:23 <ncopa> so it might help just remove the _NP
2014-08-21 06:30:40 <shafire> ncopa: how do you decide, if to unpack directly to /usr/include or subfolder?
2014-08-21 06:31:12 <ncopa> i dont, upstream does
2014-08-21 07:26:19 <shafire> how does alpine_dev need to look like, when I want to boot from the disk first partition?
2014-08-21 07:26:34 <shafire> alpine_dev=usbdisk:vfat is for the usb stick
2014-08-21 07:35:04 <ncopa> alpine_dev=sda1:vfat
2014-08-21 07:44:05 <shafire> ncopa: can I say that it should mount as /media/disk and not /media/sda1?
2014-08-21 07:44:17 <ncopa> not really
2014-08-21 07:44:27 <ncopa> you can also use the UUID
2014-08-21 07:44:39 <ncopa> alpine_dev=UUID=xx-yy
2014-08-21 07:44:54 <ncopa> hm
2014-08-21 07:45:08 <ncopa> might be that you can add it to fstab
2014-08-21 07:45:31 <ncopa> but then you'll have to build your own initramfs i think
2014-08-21 07:55:40 <fabian_a> please git pull git://git.alpinelinux.org/fab/aports
2014-08-21 07:55:58 <fabian_a> some updates and new aports incl. deps
2014-08-21 07:56:06 <fabian_a> thanks
2014-08-21 07:56:45 <shafire> it does work without own initramfs
2014-08-21 07:57:08 <shafire> mounted on /media/disk :)
2014-08-21 07:57:13 <shafire> just using /etc/fstab
2014-08-21 07:57:23 <ncopa> ah
2014-08-21 07:57:24 <ncopa> right
2014-08-21 07:57:40 <ncopa> i think i do mount move before switch root
2014-08-21 07:57:44 <ncopa> nice :)
2014-08-21 07:58:35 <shafire> now, I am running tmpfs from disk
2014-08-21 07:58:50 <shafire> now, I am running tmpfs from disk mounted on /media/disk
2014-08-21 08:01:20 <ncopa> fabian_a: done. Thanks!
2014-08-21 11:29:45 <vkrishn> ncopa, mosquito v1.3.4 released
2014-08-21 11:30:45 <vkrishn> would this work under grsec, http://www.netfort.gr.jp/~dancer/software/dsh.html.en ?
2014-08-21 11:33:05 <ncopa> why wouldnt it run under grsec?
2014-08-21 11:35:26 <vkrishn> not sure, logically it should
2014-08-21 11:36:46 <vkrishn> unlike https://github.com/nelhage/reptyr , that should also if same user does the re-attaching
2014-08-21 11:57:09 <ScrumpyJack> hello all
2014-08-21 11:58:55 <vkrishn> ncopa, in kernel 3.17 lustre has improved, chance of getting in AL ?
2014-08-21 11:59:46 <vkrishn> found http://lustre.org and http://lustre.opensfs.org/ but not sure if they are same
2014-08-21 12:01:44 <vkrishn> nice to read, http://marc.merlins.org/perso/btrfs/post_2014-05-21_My-Btrfs-Talk-at-Linuxcon-JP-2014.html
2014-08-21 13:29:47 <clandmeter> ncopa: does alpine have support for building go pkgs?
2014-08-21 13:31:59 <ncopa> paritially
2014-08-21 13:32:04 <ncopa> only for x86_64 afaik
2014-08-21 13:32:15 <ncopa> uggedal has packaged go
2014-08-21 13:32:51 <clandmeter> im trying to build something but it complains about godep
2014-08-21 13:33:18 <clandmeter> i can find it on github, but it looks like another dep resolver without proper versioning :|
2014-08-21 13:40:13 <uggedal> godep is not supported upstream
2014-08-21 13:40:22 <uggedal> so, you need to install it yourself
2014-08-21 13:40:33 <uggedal> and yes, versioning in go is a pita
2014-08-21 13:40:54 <uggedal> we should have go for x86 also
2014-08-21 13:41:13 <ncopa> kaniini: i cannot find any bug on bugs.a.o that mentions hvmloader issue
2014-08-21 13:41:18 <ncopa> and i dont remember the details
2014-08-21 13:41:45 <ncopa> iirc it was soemthing with building 32bit binary pulled in 64bit structs from libc
2014-08-21 13:41:59 <ncopa> or 64bit something from libc
2014-08-21 13:42:08 <ncopa> so we should try build with -nostdinc
2014-08-21 13:42:29 <ncopa> build fails due to they use stdint.h and stdarg.h
2014-08-21 14:06:40 <ncopa> kaniini: do you think you could file a bug on bugs.a.o saying hvmloader on 3.0 is broken and write down the details
2014-08-21 14:06:46 <kaniini> sure
2014-08-21 14:47:15 <clandmeter> ncopa: looks like x86 builder has bash installed
2014-08-21 14:48:10 <ncopa> edge?
2014-08-21 14:48:22 <clandmeter> yes i think so
2014-08-21 14:48:25 <ncopa> yup it had
2014-08-21 14:48:27 <ncopa> fixed that now
2014-08-21 14:48:44 <clandmeter> where are the logs for arm?
2014-08-21 14:49:02 <ncopa> i dont think they are public any place
2014-08-21 14:49:48 <clandmeter> go is new, and its pretty quiet when building. does it actually compile anything?
2014-08-21 14:49:55 <clandmeter> new for me
2014-08-21 14:50:51 <clandmeter> and the bins are nice and big. must be very optimized
2014-08-21 16:41:27 <clandmeter> rnalrd: ok if i upgrade owncloud to rc1?
2014-08-21 18:39:51 <barthalion> joys of static linking
2014-08-22 07:49:06 <rnalrd> clandmeter: owncloud edge is already at latest version afaik
2014-08-22 07:55:17 <clandmeter> rnalrd: yes only oc core is updated to rc1
2014-08-22 07:55:32 <clandmeter> i thought core == owncloud
2014-08-22 07:55:56 <clandmeter> the picture viewer is not really working in 7.0.1
2014-08-22 07:58:14 <clandmeter> barthalion: arch has a headless version of libreoffice?
2014-08-22 07:59:28 <rnalrd> clandmeter, picture has a bug in 7.x iirc
2014-08-22 07:59:57 <clandmeter> should be fixed now. thats why i watned to try rc1. but ill wait for 702
2014-08-22 08:00:56 <clandmeter> rnalrd: i think there are also some deps missing in oc
2014-08-22 08:01:08 <clandmeter> php-xml and i think another one.
2014-08-22 08:01:38 <clandmeter> and some subpkgs dont install the needed php modules.
2014-08-22 08:01:38 <rnalrd> weird, i did a setup from scratch when pushing version 7
2014-08-22 08:01:50 <clandmeter> maybe i did it wrong
2014-08-22 08:02:05 <rnalrd> i just tried to install and login though
2014-08-22 08:02:08 <rnalrd> nothing more
2014-08-22 08:02:31 <rnalrd> but install should do a php modules check
2014-08-22 08:02:38 <rnalrd> and complains if anything is missing
2014-08-22 08:03:19 <clandmeter> from what i remember, it was missing 2 php modules
2014-08-22 08:03:34 <clandmeter> and ldap doesnt install php-ldap
2014-08-22 08:03:44 <rnalrd> 
ACTION setups an OC brand new instance and tests

2014-08-22 08:03:52 <clandmeter> :)
2014-08-22 08:04:07 <rnalrd> subpkg may miss a dep, yes
2014-08-22 08:04:31 <rnalrd> 
ACTION never used OC w/ LDAP

2014-08-22 08:04:33 <clandmeter> same goes for external support is missing ftp
2014-08-22 08:05:05 <clandmeter> but will also complain about smbclient (but i guess ppl can isntall that themselves)
2014-08-22 08:05:20 <clandmeter> huge dep list on samba-client
2014-08-22 08:06:02 <rnalrd> clandmeter, the says to contact your adminstrator for that give storage stype :)
2014-08-22 08:06:07 <rnalrd> the page says*
2014-08-22 08:06:37 <clandmeter> yes, i called my office phone from my mobile ;-)
2014-08-22 08:06:54 <rnalrd> :D
2014-08-22 08:07:43 <clandmeter> there is also an issue with the apps directory when using lighttpd
2014-08-22 08:07:48 <clandmeter> had to chown it.
2014-08-22 08:08:26 <clandmeter> but i think lighttp is member of www-data group, so with correct permissions it should work.
2014-08-22 08:09:11 <rnalrd> clandmeter, the provided config.php makes the apps directory read-only
2014-08-22 08:09:24 <rnalrd> which is correct for run-from-ram installations
2014-08-22 08:09:34 <rnalrd> that should prevent installer to complain
2014-08-22 08:09:42 <clandmeter> after i login it works
2014-08-22 08:09:49 <rnalrd> however, during installation default config its overwritten
2014-08-22 08:09:51 <clandmeter> i logout and get an error about apps permissions
2014-08-22 08:10:10 <rnalrd> and apps directory becomes read-write
2014-08-22 08:10:31 <rnalrd> but w/o correct ownership it fails
2014-08-22 08:10:48 <rnalrd> btw i'd suggest using nginx instead of lighttpd
2014-08-22 08:11:13 <clandmeter> because?
2014-08-22 08:11:14 <rnalrd> when dealing with large files, lighty takes lot of RAM
2014-08-22 08:11:25 <clandmeter> oh ok
2014-08-22 08:11:31 <clandmeter> thats for the headsup
2014-08-22 08:11:34 <clandmeter> thanks
2014-08-22 08:11:39 <rnalrd> clandmeter http://wiki.alpinelinux.org/wiki/Owncloud#Webserver
2014-08-22 08:15:48 <clandmeter> rnalrd: have you tried OO or libre with owncloud?
2014-08-22 08:15:58 <rnalrd> nope
2014-08-22 08:16:20 <clandmeter> i wonder if its worth the hassle
2014-08-22 08:16:35 <rnalrd> php-xml is pulled in
2014-08-22 08:17:03 <rnalrd> test3:~# apk add owncloud-pgsql
2014-08-22 08:17:08 <rnalrd> (20/26) Installing php-xml (5.5.15-r1)
2014-08-22 08:23:06 <clandmeter> rnalrd: i guess you are usging edge?
2014-08-22 08:23:14 <rnalrd> y
2014-08-22 08:26:12 <clandmeter> rnalrd: http://sprunge.us/QhEH
2014-08-22 08:28:04 <rnalrd> so installer complains that's missing php-xml?
2014-08-22 08:29:25 <clandmeter> yep
2014-08-22 08:29:29 <clandmeter> i just installed base
2014-08-22 08:29:33 <rnalrd> k
2014-08-22 08:29:38 <clandmeter> PHP module libxml not installed.
2014-08-22 08:29:38 <clandmeter> Please ask your server administrator to install the module.
2014-08-22 08:29:52 <rnalrd> php-xml added
2014-08-22 08:30:17 <rnalrd> since I installed only the pgsql version, i always had that module installed
2014-08-22 08:30:54 <clandmeter> looks like its the only one missing'
2014-08-22 08:40:18 <rnalrd> added php-ldap and php-ftp for respective subpks
2014-08-22 08:41:38 <barthalion> clandmeter: no
2014-08-22 08:42:00 <barthalion> well, depends what you mean by headless
2014-08-22 08:42:07 <barthalion> unoconv is standalone package yet
2014-08-22 08:42:50 <barthalion> but upstream has dropped file lists, do we won't be splitting libreoffice to respective subpackages anymore, too much burden given the mess in libreoffice
2014-08-22 08:43:04 <barthalion> like draw doesn't start without calc and so on
2014-08-22 08:44:07 <clandmeter> barthalion: https://wiki.documentfoundation.org/Development/HeadlessBuild
2014-08-22 08:44:46 <barthalion> interesting, I didn't know about it
2014-08-22 08:45:00 <barthalion> then no, we don't
2014-08-22 08:46:43 <clandmeter> build already complains aobut musl not supported :)
2014-08-22 08:47:09 <barthalion> yeah, I'm not surprised
2014-08-22 08:47:36 <clandmeter> i saw your pkg on arch... it makes me cry :)
2014-08-22 08:47:51 <barthalion> good to know I'm not alone
2014-08-22 08:48:33 <barthalion> ncopa: your thoughts on mips port of alpine?
2014-08-22 08:49:19 <clandmeter> i think he will mention maintenance :)
2014-08-22 08:51:03 <barthalion> sure, not big deal
2014-08-22 08:55:49 <clandmeter> haha: checking whether libc is >= 2.1.1... configure: error: no, upgrade libc
2014-08-22 08:57:36 <clandmeter> bug detected --without-glibc doesnt work!
2014-08-22 08:58:23 <rnalrd> clandmeter lol
2014-08-22 11:51:10 <shafire> hi
2014-08-22 11:51:24 <shafire> how can I see the difference between package paxctl and pax-utils?
2014-08-22 11:51:34 <shafire> which files each package contains?
2014-08-22 11:55:47 <clandmeter> apk info --help
2014-08-22 11:56:23 <clandmeter> -L, --contents          List contents of the PACKAGE
2014-08-22 12:49:17 <ScrumpyJack> hello
2014-08-22 13:13:01 <rnalrd> algitbot: build 2.5-stable
2014-08-22 13:13:08 <rnalrd> algitbot:build 2.5-stable
2014-08-22 13:16:21 <rnalrd> algitbot: build master
2014-08-22 20:53:55 <tdtrask> ncopa: I just pushed acf-freeradius3 to git.alpinelinux.org:/home/tdtrask/. Can you please move it to /gitroot/ and add the hook for creating the source package?
2014-08-22 20:53:57 <tdtrask> thanks
2014-08-22 22:40:53 <fabled> seems diff breakage is caused by ncopa's --no-deref patch to bb
2014-08-22 22:45:41 <fabled> ha
2014-08-23 11:41:59 <fabian_a> please git pull git://git.alpinelinux.org/fab/aports
2014-08-23 11:42:23 <fabian_a> there are a couple of update
2014-08-23 11:42:25 <fabian_a> thanks
2014-08-25 06:37:42 <kaniini> hi
2014-08-25 06:39:16 <barthalion> morning
2014-08-25 06:42:47 <ncopa> morning
2014-08-25 06:56:22 <kaniini> ncopa: looking in more detail at the hvmloader stuff, i think the better plan is to just include a stdint.h with the types declared correctly.
2014-08-25 06:59:09 <ncopa> kaniini: that was my general impression too
2014-08-25 06:59:50 <ncopa> ship a stdint.h and add -I
2014-08-25 07:05:44 <kaniini> i was going to go a step further and change it to like stdint_local.h
2014-08-25 07:07:23 <ncopa> that would work too
2014-08-25 07:08:49 <ncopa> i suppose the longterm fix woudl be to install 32 bit headers for x86_64 in separate subdir
2014-08-25 07:09:03 <ncopa> and fix the gcc spec when -m32 is used
2014-08-25 07:10:13 <ncopa> we might have the same issue with other packages like qemu bios images, dosbox, etc
2014-08-25 07:12:54 <zenny> Hi, what can be done when grsec prevents restricts openvswich. See http://imgur.com/fArJkb1?
2014-08-25 07:14:29 <ncopa> zenny: it is likely not grsec that causes it. It just catches it
2014-08-25 07:14:43 <zenny> openswitch from testing repo works perfect except the grsec seems to be restricting from getting onto the network as seen on http://imgur.com/fArJkb1
2014-08-25 07:15:41 <ncopa> i cannot see anything that indicates that grsec prevents it getting onto the network
2014-08-25 07:15:47 <ncopa> it looks that it segfaults
2014-08-25 07:15:53 <ncopa> so it is a bug
2014-08-25 07:16:05 <zenny> ncopa: there is everything setup including network and ovs bridges, but no network. Routing does also works. Even I removed bridge and ebtables to check hurdles.
2014-08-25 07:16:16 <zenny> I see, but with what?
2014-08-25 07:16:49 <ncopa> is it from edge/testing repo?
2014-08-25 07:16:59 <ncopa> i dont know why it segfaults/craches
2014-08-25 07:17:03 <zenny> openvswitch is from testing
2014-08-25 07:17:06 <ncopa> it is likely a bug some place
2014-08-25 07:17:38 <ncopa> you dont mix edge/testing with v2.x repos, right?
2014-08-25 07:18:00 <ncopa> i can enable a -dbg package
2014-08-25 07:18:09 <ncopa> so you can generate a core dump and backtrace
2014-08-25 07:18:09 <zenny> Nope, just pulled openvswitch-2.3.0 using @tag
2014-08-25 07:18:25 <ncopa> what other repos do you have there?
2014-08-25 07:19:40 <zenny> v3.0, edge and testing enabled, but the latter two are tagged.
2014-08-25 07:20:17 <ncopa> ok
2014-08-25 07:20:23 <ncopa> that should work
2014-08-25 07:20:35 <zenny> Alpine Release is at 3.0.3
2014-08-25 07:21:44 <zenny> I am here because the network does not work no matter what I do. Even tried by enabling and disabling libvirtd with openvswitch.
2014-08-25 07:21:55 <zenny> But no go. :-(
2014-08-25 07:22:10 <ncopa> well, openvswtich crashes
2014-08-25 07:23:02 <ncopa> do you think you could create a coredump?
2014-08-25 07:23:14 <ncopa> and a backtrace?
2014-08-25 07:23:57 <ncopa> first add rc_ulimit="-c unlimited" to the conf.d file(s)
2014-08-25 07:23:59 <zenny> Is there any pointer how it is done in alpine?
2014-08-25 07:24:22 <ncopa> that should enable the coredump
2014-08-25 07:24:47 <ncopa> it will create the 'core' file in the working dir, which might depend on the daemon
2014-08-25 07:25:45 <ncopa> you can force the core dump to a given location with: echo '/tmp/core' > /proc/sys/kernel/core_pattern
2014-08-25 07:25:54 <ncopa> then the coredump will be saved as /tmp/core
2014-08-25 07:26:13 <zenny> do you want me to coredump for all init processes or just libvirtd, openvswitch or anything else?
2014-08-25 07:26:24 <ncopa> openvswitch
2014-08-25 07:26:37 <zenny> okay, shall do.
2014-08-25 07:26:40 <ncopa> it is ovs-vswitchd that segfaults
2014-08-25 07:26:49 <ncopa> so that is the core dump we need
2014-08-25 07:27:05 <ncopa> upgrade to 2.3.0-r1 and apk add openvswitch-dbg
2014-08-25 07:27:21 <ncopa> do that before you generate the core dump
2014-08-25 07:27:28 <ncopa> apk add gdb too while at it
2014-08-25 07:27:46 <ncopa> once you have the 'core' file, you can analyze it with gdb
2014-08-25 07:28:29 <ncopa> gdb --core /tmp/core /usr/sbin/ovs-vswitchd
2014-08-25 07:28:39 <ncopa> and withing gdb you run: bt
2014-08-25 07:28:42 <ncopa> for 'backtrace'
2014-08-25 07:29:10 <ncopa> that should tell where in the code the segfault happens
2014-08-25 07:30:14 <ncopa> create an issue on bugs.alpinelinux.org with the info you have: (what repos you have, the dmesg output image and with the backtrace if possible)
2014-08-25 07:31:06 <clandmeter> ncopa: have you used bb tar with xz?
2014-08-25 07:31:28 <ncopa> probably only extract
2014-08-25 07:31:48 <clandmeter> xz should be better then gz right?
2014-08-25 07:31:55 <clandmeter> compression wise
2014-08-25 07:32:03 <ncopa> dependson how you define 'better' :)
2014-08-25 07:32:04 <ncopa> yes
2014-08-25 07:32:11 <kaniini> fabled: hi, now that it is many weeks later, what is your viewpoint on openssl vs libressl
2014-08-25 07:32:11 <clandmeter> i think there is somethign wrong
2014-08-25 07:32:14 <clandmeter> size is same as gz
2014-08-25 07:32:26 <clandmeter> makes me believe it just using gz
2014-08-25 07:32:29 <fabled> morning kaniini
2014-08-25 07:32:34 <kaniini> use file on it
2014-08-25 07:32:39 <kaniini> it will tell you if it's just gz
2014-08-25 07:32:40 <ncopa> clandmeter: I suppose it depends on the data you want compress
2014-08-25 07:32:41 <fabled> i have not looked at libressl for a while
2014-08-25 07:32:55 <fabled> should probably take a look at it
2014-08-25 07:34:34 <clandmeter> file says its a tar archive.
2014-08-25 07:35:01 <kaniini> that would indicate not compressed at all
2014-08-25 07:35:35 <kaniini> otherwise it would say
2014-08-25 07:35:37 <kaniini> gzip archive
2014-08-25 07:35:41 <kaniini> or 'data'
2014-08-25 07:36:19 <clandmeter> correct. its just a tar archive
2014-08-25 07:36:45 <zenny> ncopa: there is no openvswitch-dbg package
2014-08-25 07:36:45 <clandmeter> so it just ignores J switch
2014-08-25 07:37:23 <ncopa> zenny: I just added it, it should be on nl.alpinelinux.org or on the mirrors in an hour or so
2014-08-25 07:37:51 <kaniini> clandmeter: j would be to use bzip2
2014-08-25 07:38:07 <clandmeter> j works
2014-08-25 07:38:19 <kaniini> oh, wow
2014-08-25 07:38:22 <clandmeter> J       (De)compress using xz
2014-08-25 07:38:23 <kaniini> they made capital J use xz
2014-08-25 07:38:30 <kaniini> that's really kind of a usability fail
2014-08-25 07:38:40 <ncopa> clandmeter: yes, I think bb does not support -J (yet?)
2014-08-25 07:38:54 <clandmeter> the help says it does
2014-08-25 07:38:54 <ncopa> it might been fixed in edge
2014-08-25 07:39:03 <ncopa> then its likely a bug
2014-08-25 07:39:08 <clandmeter> im on edge
2014-08-25 07:40:06 <ncopa> kaniini: i'm trying to use a stdint_local.h
2014-08-25 07:40:07 <ncopa> its messy
2014-08-25 07:40:33 <ncopa> there is a hvmloader/acpi file that uses HOSTCC, for a tool used compile time only
2014-08-25 07:40:43 <ncopa> then there is util.h which does #include <xen/xen.h>
2014-08-25 07:40:43 <ncopa> #include <xen/hvm/hvm_info_table.h>
2014-08-25 07:41:08 <kaniini> ncopa: plan c: bitch at royger
2014-08-25 07:41:16 <ncopa> which will give conflicting type uint64_t
2014-08-25 07:41:29 <ncopa> :)
2014-08-25 07:41:31 <kaniini> he works for those people
2014-08-25 07:41:36 <kaniini> this is now his problem(tm)
2014-08-25 07:42:00 <ncopa> right
2014-08-25 07:42:17 <ncopa> i'm jusst amused that this works at all...
2014-08-25 07:42:35 <kaniini> i've found worse
2014-08-25 07:42:37 <kaniini> in xen before
2014-08-25 07:42:41 <ncopa> true
2014-08-25 07:43:07 <kaniini> a friend of mine wants me to help him build a cloud solution, i think i will just use kvm this time
2014-08-25 07:43:30 <ncopa> i am sceptic to kvm
2014-08-25 07:43:39 <ncopa> they have relatively short support time
2014-08-25 07:43:56 <kaniini> hmm, true
2014-08-25 07:43:58 <ncopa> there are a number of sec bugs that is not fixable in v2.6 and older i think
2014-08-25 07:44:13 <ncopa> when asking the devs, they said: they dont support older for a reason
2014-08-25 07:44:28 <ncopa> its very timeconsuming to fix those bugs
2014-08-25 07:44:46 <ncopa> the fix was higly intrusive patch set of 10+ patches
2014-08-25 07:45:43 <ncopa> xen otoh, does a good jbo supporting the old stuf long time
2014-08-25 07:45:47 <ncopa> job*
2014-08-25 07:47:38 <kaniini> hmm true
2014-08-25 07:50:01 <kaniini> i think xen
2014-08-25 07:50:04 <kaniini> has a more coherent design
2014-08-25 07:50:35 <kaniini> but the code is in some places done by people who dont really know what they are doing, or don't think about what-if's
2014-08-25 07:51:29 <ncopa> i know what you mean
2014-08-25 07:52:34 <zenny> ncopa: after appending the file to conf.d file related to ovs-vswitchd and then apk add gdb and openvswithc-dbg packages, rebooted again, but no core was dumped to neither to the working directory nor to /tmp/ which is specified in the /proc/sys/kernel/core_pattern
2014-08-25 07:52:43 <zenny> did I miss something?
2014-08-25 07:53:14 <ncopa> the /proc/sys/kernel/core_patter does not survice reboots
2014-08-25 07:53:21 <ncopa> survive*
2014-08-25 07:54:00 <ncopa> i think you could add kerne.core_pattern = /tmp/core in /etc/sysctl.conf, but i dont think that should be necessary
2014-08-25 07:54:08 <zenny> but didn't find anything in the working directory too
2014-08-25 07:54:22 <ncopa> the working directory depends on the daemon
2014-08-25 07:54:38 <ncopa> the daemon could do chdir(workdir)
2014-08-25 07:54:56 <ncopa> it coudl be in /var/lib some place or it coudl be in /
2014-08-25 07:54:59 <zenny> ncopa: okay found in /core
2014-08-25 07:55:25 <ncopa> then try: gdb --core /core /usr/sbin/ovs-vswitchd
2014-08-25 07:55:34 <clandmeter> ncopa: isnt it possible to let algitbot shout errors?
2014-08-25 07:55:44 <ncopa> and when in gdb do: bt
2014-08-25 07:55:53 <ncopa> clandmeter: everythign is possible
2014-08-25 07:56:35 <ncopa> i think i enable continue on errors for edge for now
2014-08-25 07:56:41 <ncopa> til we have cleaned up testing repo
2014-08-25 07:56:49 <ncopa> but
2014-08-25 07:56:50 <ncopa> hum
2014-08-25 07:57:09 <ncopa> it should be possible to make it post build errors here anyway
2014-08-25 08:00:08 <clandmeter> ok, that would be nice.
2014-08-25 08:02:14 <zenny> ncopa: The coredump and backtrace output is at: http://imgur.com/ebwnJC6
2014-08-25 08:02:24 <zenny> would you mind having a look? thanks!
2014-08-25 08:02:44 <zenny> clandmeter: thanks for the help last week.
2014-08-25 08:03:23 <fabled> zenny, can you install musl-dbg and redo the backtrace?
2014-08-25 08:03:37 <fabled> well
2014-08-25 08:03:40 <fabled> maybe that doesn't matter
2014-08-25 08:04:33 <ncopa> zenny: very well, that was helpful
2014-08-25 08:04:53 <ncopa> zenny: do you think you could file an issue on bugs.alpinelinux.org and upload that image there?
2014-08-25 08:05:05 <zenny> Yep, I shall do.
2014-08-25 08:05:08 <ncopa> something like: openvswitch segfauls
2014-08-25 08:05:11 <ncopa> segfautls*
2014-08-25 08:06:01 <ncopa> fabled: he has v3.0 repos + @edge/testing
2014-08-25 08:06:48 <fabled> is that reproducible or random crash?
2014-08-25 08:06:59 <ncopa> zenny: ^^^
2014-08-25 08:07:06 <ncopa> i think it happens always
2014-08-25 08:07:30 <ncopa> as i understand it segfaults as soon traffic start flowing
2014-08-25 08:07:58 <ncopa> but zenny know how to reproduce
2014-08-25 08:13:22 <ncopa> \o/ hvmloadr finally built
2014-08-25 08:13:28 <ncopa> i dont know if ti works though
2014-08-25 08:14:02 <zenny> ncopa: bug filed- https://bugs.alpinelinux.org/issues/3309
2014-08-25 08:14:37 <zenny> fabled:It is reproducible from the very beginning
2014-08-25 08:14:41 <ncopa> perfect
2014-08-25 08:14:44 <ncopa> thanks!
2014-08-25 08:18:03 <fabled> ah
2014-08-25 08:18:17 <fabled> it's stack-size issue
2014-08-25 08:18:33 <ncopa> ha
2014-08-25 08:18:36 <ncopa> makes sense
2014-08-25 08:19:04 <ncopa> should be relatively easy to fix
2014-08-25 08:19:05 <fabled> seems they have ovs_thread_create(), so making that add a size change should fix it
2014-08-25 08:19:38 <ncopa> do you follow it up and post patch to upstream, or do you want me do it?
2014-08-25 08:19:49 <fabled> i need to run for ~2 hours
2014-08-25 08:20:04 <fabled> if you want it fixed before that, then it's all yours
2014-08-25 08:20:28 <fabled> fix is to have lib/ovs-thread.c: ovs_thread_create set a stack size
2014-08-25 08:20:37 <ncopa> ok
2014-08-25 08:22:20 <zenny> ncopa+fabled: I have added how to reproduce the bug
2014-08-25 08:23:22 <ncopa> perfect
2014-08-25 08:23:25 <zenny> hope it is useful
2014-08-25 08:24:09 <ncopa> when does it segfault?
2014-08-25 08:24:32 <fabled> probably when there's traffic
2014-08-25 08:24:41 <fabled> on first netlink message
2014-08-25 08:24:47 <ncopa> ok
2014-08-25 08:24:59 <fabled> ok. i need to run. back in ~2 hours
2014-08-25 08:38:07 <zenny> ncopa: I updated some more information to the https://bugs.alpinelinux.org/issues/3309 about the segfaults.
2014-08-25 08:38:13 <zenny> with some screenshots
2014-08-25 08:39:16 <zenny> fabled+ncopa: There is no traffic possible as there is no network, so there is hardly any traffic from outside, could be the traffic between bridge and physical interface.
2014-08-25 08:50:19 <zenny> ncopa: added tail of dmesg.
2014-08-25 08:50:58 <ncopa> zenny: we think we know why it happens
2014-08-25 08:51:07 <ncopa> and we might have a patch up soon for testing
2014-08-25 08:54:59 <zenny> ncopa: please update me once it is patched and ready for testing. thanks!
2014-08-25 09:00:34 <ncopa> zenny: i'm 90% sure openvswitch-2.3.0-r2 fixes the issue
2014-08-25 09:00:43 <ncopa> care to upgrade and test?
2014-08-25 09:09:04 <fabled> ncopa, you forgot pthread_attr_destroy from your patch
2014-08-25 09:09:14 <ncopa> ugh
2014-08-25 09:09:26 <fabled> though
2014-08-25 09:09:28 <fabled> in musl it's no-op
2014-08-25 09:09:47 <fabled> but in some implementations it can free() memory or similar
2014-08-25 09:09:56 <ncopa> understand
2014-08-25 09:10:13 <ncopa> i need read the pthread_attr_init manpage :)
2014-08-25 09:10:47 <ncopa> this thread is related: http://permalink.gmane.org/gmane.linux.network.openvswitch.general/5831
2014-08-25 09:11:17 <ncopa> apparently they think the pthread support uses too much memory
2014-08-25 09:12:06 <fabled> yes, glibc default stack size is 16MB or something crazy like that
2014-08-25 09:12:35 <fabled> or 8MB
2014-08-25 09:12:40 <ncopa> 8MB
2014-08-25 09:12:53 <ncopa> so with 8 cores 10 threads
2014-08-25 09:12:56 <ncopa> 80MB
2014-08-25 09:13:01 <ncopa> down the toilet :)
2014-08-25 09:16:17 <ncopa> i think I'd like to have it 768kb
2014-08-25 09:16:31 <ncopa> according to their tests 512 should work
2014-08-25 09:16:41 <ncopa> but might be "tight"
2014-08-25 09:20:31 <fabled> ncopa, yes, but memory is cheap, you should just buy more memory ;)
2014-08-25 09:21:00 <ncopa> so they say...
2014-08-25 09:23:51 <ncopa> zenny: could you please test openvswitch-2.3.0-r3 ?
2014-08-25 09:24:10 <ncopa> if you can confirm that it makes things run i'll clean up the patch and send it upstream
2014-08-25 09:39:15 <zenny> ncopa: I shall pull r3 and let you know.
2014-08-25 09:52:28 <zenny> ncopa: I pulled r3 and installed, but the situation remained the same. I updated the info at https://bugs.alpinelinux.org/issues/3309
2014-08-25 09:56:05 <ncopa> hm
2014-08-25 09:56:21 <ncopa> are you sure you got -r3 after the reboot?
2014-08-25 09:56:27 <ncopa> does it still segfault?
2014-08-25 09:59:43 <zenny> yep
2014-08-25 09:59:51 <ncopa> hm
2014-08-25 10:00:12 <ncopa> and it causes kernel panic...
2014-08-25 10:01:57 <zenny> ncopa: ovs-vsctl -V shows "ovs-vsctl (Open vSwitch) 2.3.0 \n Compiled Aug 25, 2014 09:22:58 \n DB Schema 7.6.0"
2014-08-25 15:14:50 <jomat> Hi there
2014-08-25 15:16:46 <jomat> I'm just building http://xorg.freedesktop.org/releases/individual/util/util-macros-1.19.0.tar.gz and I'm not sure what to use as # Maintainer: in the APKBUILD
2014-08-25 15:24:40 <jomat> Oh, my bad... util-macros already exists, I looked for it under xorg-util-macros
2014-08-25 15:39:21 <fcolista> ncopa, i've a v4l-utils updated apkbuild, the last one available (1.3.90). Just removed a patch that couldn't apply. Since is quite critical 'cause is a dependence of ffmpeg and other varius packages, would you mind if it pastebin it and you take a look?
2014-08-25 15:39:58 <fcolista> that's the patch: http://sprunge.us/RKAB
2014-08-25 15:40:11 <fcolista> this the APKBUILD: http://sprunge.us/AagN
2014-08-25 15:41:07 <fcolista> if this is ok, we should bump pkgver of gst-plugins-good1,ffmpeg,ffmpeg1,vlc and wine
2014-08-25 15:41:29 <fcolista> (in main)
2014-08-25 15:48:05 <tdtrask> ncopa: ping
2014-08-25 15:48:25 <tdtrask> <tdtrask> ncopa: I just pushed acf-freeradius3 to git.alpinelinux.org:/home/tdtrask/. Can you please move it to /gitroot/ and add the hook for creating the source package?
2014-08-25 15:48:25 <tdtrask> <tdtrask> thanks
2014-08-25 17:30:00 <clandmeter> crow_: ^
2014-08-25 20:16:53 <crow_> clandmeter thank you :)
2014-08-25 20:37:10 <crow_> clandmeter seems -dbg package is broken: ./oscam.debug: line 1: syntax error: unexpected "("
2014-08-25 20:37:17 <crow_> normal one seems working fine
2014-08-25 20:38:35 <clandmeter> dont see any error
2014-08-25 20:38:35 <clandmeter> http://bld1.alpinelinux.org/buildlogs/build-edge-x86/main/oscam/oscam-0_svn20140825-r0.log
2014-08-25 20:38:36 <clandmeter> i guess you use x86?
2014-08-25 20:38:43 <crow_> nop x86_64
2014-08-25 20:39:09 <crow_> i get it from here: http://dl-3.alpinelinux.org/alpine/edge/main/x86_64/
2014-08-25 20:39:32 <crow_> note i just unpack it and used binary, chmod +x oscam.debug and ./oscam.debug -V give me that error
2014-08-25 20:49:09 <clandmeter> crow_: i think its abuild that breaks it
2014-08-25 20:55:31 <clandmeter> crow_: try that
2014-08-26 06:05:49 <ncopa> tdtrask: done
2014-08-26 06:38:03 <zenny1> Hi, has the new release of openvswitch released?
2014-08-26 06:45:27 <ncopa> no, I havent looked more at openvswitch sorry
2014-08-26 06:48:40 <vkrishn> there is Micro Python (a tiny version of Python 3.4) in openwrt trunk, could be useful till v3.x is in testing
2014-08-26 08:15:01 <crow_> clandmeter thanks that one worked: apu4c:/usr/bin# ./oscam.debug -V -> Version: oscam-1.20-unstable_svn-r9836
2014-08-26 08:16:51 <clandmeter> ncopa: is it possible for abuild dbg funct to mess up the bin? i upgraded oscam yesterday and the dbg bin was not working anymore.
2014-08-26 08:17:13 <clandmeter> i created a custum dbg funct now to work around it.
2014-08-26 08:17:51 <ncopa> hm
2014-08-26 08:17:53 <ncopa> i dont know
2014-08-26 08:17:58 <ncopa> it should not mess up the bin
2014-08-26 11:59:41 <ScrumpyJ1ck> ncopa: git pull: Pull is not possible because you have unmerged files
2014-08-26 12:00:05 <ScrumpyJ1ck> does that mean some patches haven't been commited yet?
2014-08-26 12:00:46 <ncopa> it probably means that i have applied some patches that you have sent, but a fixed version of it
2014-08-26 12:00:56 <ncopa> google git pull: Pull is not possible because you have unmerged files
2014-08-26 12:01:48 <ncopa> http://stackoverflow.com/questions/15127078/git-pull-is-not-possible-unmerged-files
2014-08-26 12:02:01 <ScrumpyJack> hmm, perhaps there are some patches that haven't been commited too
2014-08-26 12:02:49 <jarrudaBRA> ncopa, do you know how to fix kamailio failing to stop in AL 3.0.X ? or.. taking some time to do so in fact
2014-08-26 12:02:54 <jarrudaBRA>  * Stopping Kamailio ...
2014-08-26 12:02:55 <jarrudaBRA>  * start-stop-daemon: 1 process refused to stop
2014-08-26 12:03:31 <ScrumpyJack> are these in testing on Master? http://sprunge.us/NWha
2014-08-26 12:06:56 <ncopa> jarrudaBRA: start-stop-daemon sends signal with stop
2014-08-26 12:06:58 <ncopa> sigterm iirc
2014-08-26 12:07:10 <ncopa> kamailio catches it and does a shutdown
2014-08-26 12:07:25 <ncopa> there is some option for --retry or something
2014-08-26 12:07:34 <ncopa> which has a timeout option
2014-08-26 12:08:02 <ncopa> i think its possible to wait N millisec and then try send new, optionally different signal
2014-08-26 12:08:39 <ncopa> i think that is what we need to do, have it wait for kamailio do proper shutdown
2014-08-26 12:09:48 <ncopa> maybe: --retry 5
2014-08-26 12:31:11 <alacerda> broken link: http://mirror1.hs-esslingen.de/pub/Mirrors/alpine//v3.0/releases/x86_64/alpine-3.0.4-x86_64.iso
2014-08-26 13:06:49 <ncopa> needs to be synced
2014-08-26 13:51:12 <tdtrask> ncopa: thanks
2014-08-26 13:52:25 <ncopa> np
2014-08-26 13:57:36 <ScrumpyJack> do folks commonly run X in alpinelinux? I gave it a shot a while back but failed.
2014-08-26 13:57:48 <f15RyWpgGTO6> i do
2014-08-26 13:57:50 <ScrumpyJack> Maybe I'll give it another go. Anything I should look out for?
2014-08-26 13:58:03 <f15RyWpgGTO6> not really
2014-08-26 13:58:28 <ScrumpyJack> f15RyWpgGTO6: what graphics chipset do you have?
2014-08-26 13:58:36 <f15RyWpgGTO6> intel
2014-08-26 13:58:45 <ScrumpyJack> on a laptop?
2014-08-26 13:59:24 <f15RyWpgGTO6> +
2014-08-26 14:02:00 <ScrumpyJack> i have the Intel 945GM chipset on a Dell, so I used to xf86-video-intel package
2014-08-26 14:02:57 <f15RyWpgGTO6> worksforme (TM)
2014-08-26 14:03:50 <ScrumpyJack> do you have to load the intel_drv.so module from that package? (I can't load it)
2014-08-26 14:04:09 <ScrumpyJack> (can you check if yours is running?)
2014-08-26 14:04:37 <f15RyWpgGTO6> sorry not now
2014-08-26 14:07:38 <jomat> I also run X
2014-08-26 14:08:05 <jomat> On a Thinkpad X230 with xf86-video-intel
2014-08-26 14:08:58 <jomat> /usr/lib/xorg/modules/drivers/intel_drv.so isn't a kernel module
2014-08-26 14:09:29 <jomat> ScrumpyJack: What's your problem with X?
2014-08-26 14:29:49 <ScrumpyJack> startx just gives me a blank screen after probing. no X. can't change tty after that
2014-08-26 14:33:02 <jomat> Try Ctrl-Alt-Print-e
2014-08-26 14:33:26 <jomat> Or wherever your sysreq-key is :-)
2014-08-26 14:33:32 <jomat> that should kill the xserver
2014-08-26 14:34:00 <jomat> You also need running udev for the input drivers
2014-08-26 14:34:30 <jomat> s/drivers/devices/
2014-08-26 14:35:34 <jomat> and something in your .xinitrc that is executed (i. e. a windowmanager (i. e. awesome) or just a terminal (i. e. urxvt))
2014-08-26 14:36:01 <jomat> Otherwise xfce4 is tried to be started which will fail if you haven't installed it
2014-08-26 14:50:33 <ScrumpyJack> running startx as a normal user seems to get me further
2014-08-26 14:57:16 <ScrumpyJack> now X works if i run it on it's own (although no mouse in sight) and startx exits and kills X
2014-08-26 14:58:20 <zenny1> Confused with data mode installation of 3.0.3 (standard). When the installer asked for the disk, it was 'none', but the installer died without asking for 'request mode' which should be 'data' according to http://wiki.alpinelinux.org/wiki/Installation#Basics, or did I miss something?
2014-08-26 15:01:11 <ScrumpyJack> can you run setup-disk
2014-08-26 17:23:19 <shafire> does alpine-xen v3 now work?
2014-08-26 17:43:52 <zenny1> hi: tried to configure alternative network for lxc as described here: http://wiki.alpinelinux.org/wiki/Talk:LXC, but the vlans do not start up. any clue?
2014-08-26 17:48:19 <zenny1> even lxc-checkconfig outputs "Try recompiling with IKCONFIG_PROC, installing the kernel headers"
2014-08-26 17:54:10 <crow_> should not last command work? last: can't open '/dev/null/wtmp': Not a directory
2014-08-26 18:00:46 <zenny1> The instructions for lxc does not seem to work as the latest stable release (3.0.4) seems to lack the cgroup+namespaces module enabled in the kernel because when trying to start a lxc guest, it gives: "lxc-start: command get_cgroup failed to receive response"
2014-08-26 18:00:57 <zenny1> Any inputs to make lxc work?
2014-08-26 18:16:00 <zenny1> Just wondering whether it is possible to put alpine into production? Tried with openvswitch for four days, didn't work. And now trying to experiment with lxc, it does not work either! Discouraged my will to deploy alpine in prod! ;(
2014-08-26 21:32:34 <clandmeter> crow_: its a known issue. musl doesnt support wtmp
2014-08-27 07:01:39 <ncopa> i think I'll make abuild exit with error if suid-root executable is found
2014-08-27 07:02:10 <ncopa> and have an option to allow it
2014-08-27 07:02:19 <ncopa> eg options="suid-root"
2014-08-27 07:02:21 <ncopa> or similar
2014-08-27 07:10:36 <ScrumpyJack> updated and upgraded (which included an update to xf86-video-intel, and now back to black screen when running startx
2014-08-27 07:20:02 <darkfader> ScrumpyJack: i once fell for that when the output went to the wrong port
2014-08-27 08:00:52 <ScrumpyJack> darkfader: still tinkering. dbus wasn't set to start, user starting X wasn't in video group ... I'm sure we'll get there
2014-08-27 08:13:32 <ScrumpyJack> this is my X log file http://sprunge.us/FTFB
2014-08-27 08:50:19 <ScrumpyJack> Xorg -configure gives me
2014-08-27 08:50:29 <ScrumpyJack> http://sprunge.us/aKBR
2014-08-27 08:50:40 <ScrumpyJack> anyone seen that before?
2014-08-27 08:52:22 <jomat_> ScrumpyJack: Doesn't look bad, does it?
2014-08-27 08:53:07 <ScrumpyJack> this looks bad: No devices to configure.  Configuration failed
2014-08-27 08:53:42 <ScrumpyJack> Failed to load /usr/lib/xorg/modules/drivers/intel_drv.so: Error relocating /usr/lib/xorg/modules/drivers/intel_drv.so: fbPutImage: symbol not found
2014-08-27 08:56:40 <ncopa> ScrumpyJack: it might be lazy loading issue
2014-08-27 08:57:53 <ncopa> try add the module that provide fbPutImage to /etc/X11/xorg.conf.d/20-modules.conf
2014-08-27 08:58:18 <jomat> Ah, the other file... http://sprunge.us/FTFB doesn't look bad
2014-08-27 08:59:04 <ScrumpyJack> ncopa: do i really want ramebuffer?
2014-08-27 08:59:16 <ncopa> i think not
2014-08-27 08:59:22 <ncopa> jomat:
2014-08-27 08:59:24 <ncopa> [    73.913] removing GPU device /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1 (null)
2014-08-27 08:59:24 <ncopa> [    73.913] removing GPU device /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-VGA-1 (null)
2014-08-27 08:59:24 <ncopa> [    73.914] removing GPU device /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-DVI-D-1 (null)
2014-08-27 08:59:24 <ncopa> [    73.914] removing GPU device /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-SVIDEO-1 (null)
2014-08-27 08:59:44 <ncopa> [    73.913] (EE) /dev/dri/card0: failed to set DRM interface version 1.4: Permission denied
2014-08-27 08:59:47 <ScrumpyJack> yes
2014-08-27 09:00:05 <ncopa> it could be grsec kernel's sysfs protection
2014-08-27 09:00:57 <ncopa> try: find /sys -type d -exec chmod o+rx {} \;
2014-08-27 09:01:17 <ScrumpyJack> turtle:~# ls -l /dev/dri/card0
2014-08-27 09:01:17 <ScrumpyJack> crw-rw----    1 root     video     226,   0 Aug 27 10:15 /dev/dri/card0
2014-08-27 09:03:51 <ScrumpyJack> ncopa: that didn't change anything
2014-08-27 09:13:51 <zenny> Hi. This seems very specific to alpine. Tried to create aliases as mentioned at http://wiki.alpinelinux.org/wiki/Talk:LXC with 3.0.4, but fails to initialize. Installed iprouter2 and pax-utils as suggested by ncopa at http://dev.alpinelinux.org/irclogs/%23alpine-devel-2014-04.log to address safire's problem, but no go. always "Cannot find device "eth0.8" and so on for every alias.
2014-08-27 09:16:45 <ncopa> zenny: do you have the vlan package installed?
2014-08-27 09:16:49 <ncopa> apk add vlan
2014-08-27 09:17:31 <zenny> Then tried to implement docker, everything goes fine until 'docker pull docker pull uggedal/alpine-3.0'. "/var/run/docker.sock" not available problem. Appended the /etc/conf.d/docker with "-H tcp://0.0.0.0:4243 -H unix:///var/run/docker.sock", and docker is running, yet while trying to pull, it says: "2014/08/27 11:09:07 Cannot connect to the Docker daemon. Is 'docker -d' running on this host?"
2014-08-27 09:17:43 <zenny> ncopa: let me check.
2014-08-27 09:19:03 <zenny> ncopa: thanks, apk add vlan solved the issue.
2014-08-27 09:19:39 <zenny> How is apk evoked to figure out the dependencies?
2014-08-27 09:21:40 <ncopa> ?
2014-08-27 09:23:37 <zenny> but the eth0.8 alias does not get IP from DHCP server directly and eth0.129 is not up when rebooted (the setup is exactly like http://wiki.alpinelinux.org/wiki/Talk:LXC, any thoughts?
2014-08-27 09:27:01 <zenny> and there is no network to the lxc guests with the setup like in http://wiki.alpinelinux.org/wiki/Talk:LXC
2014-08-27 09:32:24 <zenny> ncopa: even with a stable and nothing extra added except lxc, it sometimes gives kernel panic upon reboot as seen at http://imgur.com/NudZnrB
2014-08-27 09:32:56 <ncopa> do you have the start of the kernel dump?
2014-08-27 09:33:23 <ncopa> what kernel version and arch is it?
2014-08-27 09:34:38 <zenny> ncopa: kernel version: 3.14.17-0-grsec and the arch is x86_64
2014-08-27 09:35:18 <ncopa> ok
2014-08-27 09:35:35 <ncopa> do you have the first part of the dump?
2014-08-27 09:36:06 <zenny> ncopa: I don't have the start of the kernel dump because it is a new machine I deployed after I failed with openvswitch and docker.
2014-08-27 09:37:57 <ncopa> but it is triggered with openvswitch?
2014-08-27 09:38:20 <zenny> nope, this is completly vanilla machine
2014-08-27 09:38:27 <zenny> just lxc is installed onto it
2014-08-27 09:38:51 <zenny> as stated at http://wiki.alpinelinux.org/wiki/Talk:LXC
2014-08-27 12:38:08 <alacerda> I got this on TW: "@alpinelinux Any hope of an ARM build sometime for example for the Pi? Would be very interesting!"
2014-08-27 12:38:13 <alacerda> any answer?
2014-08-27 12:48:03 <fabled> we have experimental ARM build for the Pi
2014-08-27 12:48:38 <fabled> latest Pi boot image from some months back http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz
2014-08-27 12:48:43 <fabled> and the armhf repository is up-to-date
2014-08-27 12:57:07 <ScrumpyJack> i have some PIs doing nothing. i really should try/test it
2014-08-27 13:05:19 <alacerda> fabled, thanks
2014-08-27 15:18:22 <clandmeter> anyone knows how to reboot from emergency console?
2014-08-27 17:46:04 <zenny> clandmeter: you told me once that kvm with libvirt works fine. Is it possible for you to share your configs? As openvswitch didn't work nor lxc nor docker, I would like to test with your tested way. Appreciate that!
2014-08-27 18:31:53 <kaniini> clandmet1r: echo 'b' > /proc/sysrq-trigger
2014-08-27 21:28:24 <clandmet1r> kaniini: thx ill write that down.
2014-08-28 06:13:07 <fabled> larena_, use $CLIBC
2014-08-28 06:13:23 <fabled> grepping libc name in $CHOST is not fully reliable
2014-08-28 06:13:32 <larena_> k tnx
2014-08-28 07:55:10 <clandmeter> ncopa: you around?
2014-08-28 08:06:33 <clandmeter> ncopa: i have a problem with alpine_dev
2014-08-28 08:07:30 <clandmeter> even if i specify alpine_dev=sda1:vfat it will still mount as usb
2014-08-28 08:09:02 <clandmeter> and randomly switch to sda1
2014-08-28 08:13:33 <clandmeter> ncopa: so i could just use usbdisk as device, but then it sometimes tries to mount the second partition as alpine_dev (which is ext4)
2014-08-28 12:41:51 <jarrudaBrazil> clandmeter, ping
2014-08-28 12:42:24 <jarruda> clandmeter, ping
2014-08-28 12:52:44 <clandmeter> hi
2014-08-28 12:56:46 <clandmeter> jarruda: whats up?
2014-08-28 12:56:50 <jarruda> hi clandmeter
2014-08-28 12:57:22 <jarruda> Quick question... I found a log from #freeswitch IRC from about 2 months ago
2014-08-28 12:57:43 <jarruda> where you was asking about  [ERR] switch_xml.c:1715 Error
2014-08-28 12:58:04 <jarruda> did you manage to find the solution for that?
2014-08-28 12:58:07 <clandmeter> asterisk?
2014-08-28 12:58:20 <jarruda> no.. freeswitch
2014-08-28 12:58:39 <jarruda> i'm trying freeswitch on AL 3.0.4 and getting this error [ERR] switch_xml.c:1715 Error[[error near line 21]: unexpected <]
2014-08-28 12:58:58 <jarruda> not sure where it's getting the unexpected <
2014-08-28 12:59:38 <clandmeter> not sure i fixed that one. could be i just stopped trying.
2014-08-28 12:59:48 <jarruda> :)
2014-08-28 12:59:49 <jarruda> ok..
2014-08-28 13:00:08 <clandmeter> i remember i asked in freeswitch channel, but didnt get much input.
2014-08-29 09:16:13 <ScrumpyJack> morning all
2014-08-29 09:16:32 <ScrumpyJack> still no joy with X under alpine on my Dell D620
2014-08-29 09:28:00 <ScrumpyJack> Xorg -configure give me No devices to configure. Configuration failed
2014-08-29 11:24:15 <rnalrd> clandmeter, ^^^
2014-08-29 11:41:57 <clandmeter> rnalrd: thx!
2014-08-29 12:50:53 <pnutzh4x0r> ScrumpyJack: try the setup-xorg-base script
2014-08-29 12:51:35 <pnutzh4x0r> that installs udev and allowed me to have a mouse and keyboard
2014-08-29 13:22:47 <Guest9405> hey all. anyone around for a troubleshooting tip?
2014-08-29 13:25:42 <shafire> hi, does XEN now work on v3?
2014-08-29 13:30:57 <fabled> shafire, i think ncopa got it fixed for the latest release.
2014-08-29 13:32:07 <shafire> 4.4?
2014-08-29 13:34:51 <fabled> alpine 3.0.4 / xen 4.3.2-r4
2014-08-29 13:35:09 <fabled> at least there was commit saying "fix hvmloader"
2014-08-29 19:27:16 <jomat> Hm... Is there a common practice for packaging bash-/zsh-/fish-completion files? Should they be packet to the main package or to subpackages? And if they go to subpackages, one subpackage for each shell?
2014-08-29 19:27:23 <jomat> *packed
2014-08-29 19:56:21 <shafire> jomat: http://alpinelinux.org/apk/main/x86_64/nginx-vim
2014-08-29 19:56:24 <shafire> I think so, yes
2014-08-29 19:57:59 <jomat> Ok, thanks, makes sense :-)
2014-08-29 20:21:40 <neeh> Who's designing the new Alpine website and logo? I asked for feedback from graphic designers: https://boards.4chan.org/gd/thread/171228
2014-08-30 17:51:04 <jomat_> I want to submit a patch for dnsmasq, and dnsmasq is in main, should i copy it to testing?
2014-08-30 18:27:51 <barthalion> jomat: no, it's ok to submit it as "main" patch
2014-08-30 18:29:30 <jomat> thx
2014-08-30 18:39:44 <jomat> Although I'm not sure if it's worth a patch: https://github.com/jomat/aports/commit/aba2f264681b3879d58bd013723ecf5c5d50c4fc#diff-d41d8cd98f00b204e9800998ecf8427e
2014-08-31 17:57:48 <barthalion> are we looking for one arm buildbox more?
2014-08-31 17:58:06 <barthalion> friend of mine has spare wandboard quad for 70€
2014-08-31 17:58:21 <barthalion> ncopa_, rnalrd, clandmeter ↑
2014-08-31 19:24:56 <smak> Have you guys considered s6 as alpine's init system? It's, apparently, more robust and simple than OpenRC, since it's implemented in fewer lines of code and has high-level functionality like process supervision (instant restarting of failed processes).
2014-08-31 19:25:05 <smak> http://skarnet.org/software/s6/
2014-08-31 19:25:40 <smak> Is s6 known to the alpine developers? Why did you choose OpenRC instead?