2016-05-01 14:05:54 <przemoc> I'll be out of town starting today for a few days, so my plans to do some AL stuff before releasing 3.4 apparently failed. :(
2016-05-01 15:56:59 <scadu> [*]
2016-05-01 20:22:32 <coredumb> # ping
2016-05-01 20:22:34 <coredumb> ping: ping is not a valid applet
2016-05-01 20:22:41 <coredumb> is that a known issue ?
2016-05-01 20:23:28 <skarnet> something to do with bbsetuid being removed
2016-05-01 20:23:44 <coredumb> it's actually here
2016-05-01 20:23:49 <skarnet> remove it
2016-05-01 20:24:14 <coredumb> oh you mean it's been removed on edge ?
2016-05-01 20:24:26 <skarnet> ping was bbsuid beforehand, it's vanilla busybox now
2016-05-01 20:24:29 <skarnet> I think so
2016-05-01 20:24:35 <coredumb> ok
2016-05-01 20:24:51 <Adran> busybox ping
2016-05-01 20:24:52 <coredumb> is there a command to automatically fix the symlinks ?
2016-05-01 20:24:56 <coredumb> yep it works fine
2016-05-01 20:25:05 <Adran> ln -s /bin/busybox /bin/ping
2016-05-01 20:25:33 <coredumb> only ping and ping6 impacted ?
2016-05-01 20:25:53 <skarnet> no idea - list the files that bbsuid had
2016-05-01 20:27:16 <chris|> busybox --install -s
2016-05-01 20:27:27 <coredumb> cool
2016-05-01 20:28:11 <Adran> what did bbsuid contain?
2016-05-01 20:29:19 <chris|> bbsuid contains a couple of applets like mount/umount, su, passwd , and some more
2016-05-01 20:29:35 <chris|> only ping was moved out afaik
2016-05-01 20:30:00 <skarnet> yeah, I don't think we can get rid of bbsuid entirely XD
2016-05-01 20:30:22 <skarnet> (or you'll have to tell me how "su" works...)
2016-05-01 20:31:52 <chris|> bbsuid is basically a proxy command that runs busybox with suid privileges
2016-05-01 20:34:12 <skarnet> yes. ping got out because there's a way to ping without being root now.
2016-05-01 20:34:50 <skarnet> but at least "su" still needs to be in bbsuid :P
2016-05-01 20:35:42 <chris|> yeah, and passwd would probably be tricky to get right
2016-05-01 20:35:52 <coredumb> mmmh but I have su to busybox now
2016-05-01 20:36:03 <coredumb> normal ?
2016-05-01 20:36:12 <skarnet> welp if you ran busybox --install -s...
2016-05-01 20:38:05 <chris|> bbsuid --install
2016-05-01 20:38:08 <coredumb> yes I did :D
2016-05-01 20:38:38 <skarnet> oh, bbsuid has a --install option too... that's cute.
2016-05-01 20:39:00 <skarnet> (seriously, long options in a trivial wrapper...)
2016-05-01 20:40:19 <coredumb> # apk info -a bbsuid
2016-05-01 20:40:21 <coredumb> bbsuid-0.6-r0 description:
2016-05-01 20:40:23 <coredumb> Obsolete package that can be removed
2016-05-01 20:40:52 <Adran> nice
2016-05-01 20:41:00 <coredumb> ahhhh needs busybox-suid now
2016-05-01 20:41:12 <skarnet> sigh. Naming is hard.
2016-05-01 20:42:04 <skarnet> sorry for the misdirection.
2016-05-01 20:43:06 <coredumb> bbsuid --install doesn't change any symlink actuayll
2016-05-01 20:44:11 <chris|> hm, it should relink all it cares about
2016-05-01 20:44:32 <coredumb> seems it doesn't care much about anything then :P
2016-05-01 20:45:08 <Adran> i wonder if alpine will ever switch to toybox
2016-05-01 20:45:18 <chris|> keep in mind that /bin/su is busybox while bbsuid installs to /usr/bin
2016-05-01 20:45:19 <chris|> (except for mount)
2016-05-01 20:47:43 <coredumb> ohhhhh
2016-05-01 20:47:57 <coredumb> then it works :)
2016-05-01 20:48:37 <coredumb> gahhhh this new forefox >_<
2016-05-01 20:53:10 <mosez> does somebody know a guide how to install alpine on hetzner servers?
2016-05-01 20:54:28 <chris|> mosez: not a guide, but I would probably just boot to the rescue system and use apk.static to bootstrap the machine
2016-05-01 20:56:16 <mosez> chris|: sounds good... got to look also for software raid configuration :D
2016-05-01 20:58:30 <chris|> haven't tried it, but should be doable. there is also some support for mdadm in the initramfs
2016-05-01 20:59:03 <mosez> sounds good
2016-05-01 20:59:39 <chris|> unless you are thinking of btrfs raid, in which case I'd suspect it will just work
2016-05-01 21:00:38 <mosez> no i would take mdadm with raid 10
2016-05-01 21:01:48 <chris|> you should consult the syslinux manuals if that works on /boot
2016-05-02 00:59:35 <tdtrask>  ncopa: I need some direction on cron/crond #5411 #5462
2016-05-02 01:00:01 <tdtrask> do we roll back to 'cron' or update ACF and add post-upgrade fix
2016-05-02 01:00:04 <tdtrask> ?
2016-05-02 01:01:54 <tdtrask> on another topic, why do I keep seeing redmine internal errors?
2016-05-02 01:16:11 <swswswx> anyone have any opinions on what the biggest limitations of alpine linux is?
2016-05-02 01:25:52 <Adran> xorg
2016-05-02 01:25:55 <Adran> graphical
2016-05-02 07:40:19 <ncopa> packages that needs fixing from build-3-4-x86:  http://tpaste.us/3JZa
2016-05-02 07:42:52 <ncopa> build-3-4-x86_64: http://tpaste.us/A8wR
2016-05-02 08:00:17 <vkris> Hi, if anyone doing the next update for uwsgi can pls have a look at
2016-05-02 08:00:17 <vkris> https://cgit.gentoo.org/repo/gentoo.git/tree/www-servers/uwsgi/uwsgi-2.0.12.ebuild
2016-05-02 08:00:20 <vkris> and see if it would be useful to include/enable modules for "zeromq(+log), php, redis, geoip, amqp, redislog..etc" for AL
2016-05-02 10:28:24 <leo-unglaub> hey :)
2016-05-02 10:37:08 <jirutka> hi fabled, have you looked at https://github.com/alpinelinux/aports/pull/66 ?
2016-05-02 10:38:23 <fabled> jirutka, why is the 'java' package needed as now the packages provides the name too?
2016-05-02 10:39:28 <fabled> the other changes look ok
2016-05-02 10:39:52 <jirutka> because when you install a package that depends on java-jre and there is no actual java-jre package, than APK doesn’t know what dependency to install
2016-05-02 10:39:58 <jirutka> or have I missed something?
2016-05-02 10:40:36 <jirutka> I’ve tried this scenario and it didn’t work, APK detects that two pacakges provides java-jre, but doesn’t now which one to install
2016-05-02 10:41:07 <fabled> use versioned provides
2016-05-02 10:41:14 <fabled> provides="java-jdk=8"
2016-05-02 10:41:27 <jirutka> aha, I didn’t know about this feature!
2016-05-02 10:41:32 <jirutka> thanks, I’ll try it
2016-05-02 10:41:33 <fabled> mmm
2016-05-02 10:41:34 <fabled> though
2016-05-02 10:41:44 <fabled> then the java stuff can't co-exist
2016-05-02 10:41:56 <fabled> that's actually something we might need to fix
2016-05-02 10:42:22 <fabled> you can install multiple packages with same provides only if they don't have version in them
2016-05-02 10:42:29 <fabled> but then no automatic selection is done
2016-05-02 10:42:43 <fabled> adding the version to provides make it treated like real package name
2016-05-02 10:42:48 <fabled> you can have only one of them installed
2016-05-02 10:43:01 <fabled> and the latest version of that name is automatically preferred and installed of possible
2016-05-02 10:44:01 <fabled> but adding the meta-package breaks it too
2016-05-02 10:44:20 <fabled> having a real package java-jdk=xxx would conflict with the provides iirc
2016-05-02 10:45:53 <jirutka> that’s not good, what we can do about it?
2016-05-02 10:47:14 <jirutka> I must to go now, I’ll be back in ~3 hours
2016-05-02 10:47:42 <jirutka> I’ve added note to the PR to not merge it yet
2016-05-02 11:02:57 <leo-unglaub> ncopa: i really loke your efforts to remove old an unmaintained packages from alpine
2016-05-02 11:03:07 <leo-unglaub> better to have less packages but those perfectly maintained!
2016-05-02 11:05:03 <mosez> hum... i dont get it... how can i persist my alpine installation to the hdd? i have doen everything within a chroot env. but boot is empty :(
2016-05-02 11:06:53 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d is what i have so far while https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d#file-gistfile1-txt-L1-L25 is hetzner (server provider) specific setup
2016-05-02 11:08:34 <mosez> extlinux -r -i /boot and dd if=/usr/share/syslinux/mbr.bin of=/dev/sda is not everything what i need, right?
2016-05-02 11:08:57 <mosez> got only ldlinux.c32  ldlinux.sys within the boot folder
2016-05-02 12:57:44 <mosez> ncopa: you can help me for sure 8)
2016-05-02 12:58:43 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d it's still not booting and i have no clue what i'm missing
2016-05-02 12:59:09 <mosez> i have also requested a serial console at the provider to get a view where it stucks... hopefully.
2016-05-02 13:10:16 <mosez> sad... http://cl.ly/0H1e2c2K1y2f
2016-05-02 13:10:32 <mosez> looks like the boot is not set correctly :(
2016-05-02 13:20:25 <jirutka> fabled: I’m back
2016-05-02 13:24:45 <jirutka> so what we can do with java-{jdk,jre,jre-base} metapackages?
2016-05-02 13:25:37 <jirutka> ncopa, are you here?
2016-05-02 13:25:56 <ncopa> jirutka: i'm back
2016-05-02 13:27:13 <jirutka> ncopa: could you please look at https://github.com/alpinelinux/aports/pull/63, https://github.com/alpinelinux/aports/pull/64, and https://github.com/alpinelinux/aports/pull/65 ?
2016-05-02 13:28:01 <jirutka> I’m actually using these packages proposed for moving from testing to master, so they are tested and I can support them
2016-05-02 13:28:50 <jirutka> I didn’t proposed packages that I’m not actively using (yet), like jruby
2016-05-02 13:33:23 <mosez> ncopa: can you tell me what i'm missing at https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d?
2016-05-02 13:34:21 <mosez> hum.... looks like it's an issue to boot from raid10 -.-
2016-05-02 13:37:30 <ncopa> jirutka: re https://github.com/alpinelinux/aports/pull/65
2016-05-02 13:37:36 <ncopa> i'm sceptic to move it to main
2016-05-02 13:37:52 <ncopa> actually
2016-05-02 13:38:03 <ncopa> they depend on openjdk, right?
2016-05-02 13:38:13 <ncopa> they need to go to community
2016-05-02 13:38:25 <ncopa> we cannot have things in main that depends on things in community
2016-05-02 13:41:01 <ncopa> mosez: what error do you get?
2016-05-02 13:42:15 <ncopa> mosez: i suspsect thet the raid10 module does not get included in the initramfs
2016-05-02 13:42:21 <jirutka> You’re right, I’ll move it to the community repo.
2016-05-02 13:42:24 <mosez> http://cl.ly/0H1e2c2K1y2f system is not able to find a boot record
2016-05-02 13:43:30 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d#file-gistfile1-txt-L144 that's really all to write the mbr?
2016-05-02 13:45:26 <ncopa> that should be correct
2016-05-02 13:45:27 <mosez> the virtual console will remain for ~1 more hour... afterwards i can not really debug it anymore :(
2016-05-02 13:45:47 <ncopa> oh
2016-05-02 13:45:51 <mosez> so maybe i should just overwrite the raid for sda1
2016-05-02 13:45:53 <ncopa> you use grub as boot loader?
2016-05-02 13:46:05 <mosez> no, tried it with the default loader
2016-05-02 13:47:03 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d is everything i'm executing. starting at line 46 it's pure alpine related stuff
2016-05-02 13:47:48 <ncopa> and /dev/md0 is raid1?
2016-05-02 13:48:56 <ncopa> i cannot see any place how/where /dev/md0 is created
2016-05-02 13:50:12 <ncopa> you should also verify that you have "raid" listed in `features` in /etc/mkinitfs/mkinitfs.conf
2016-05-02 13:57:22 <mosez> md0 and md1 are both raid10 and they are created with the installimage script from the provider. it's initially installing opensuse which i'm wiping afterwards
2016-05-02 13:57:43 <ncopa> i think thats the problem
2016-05-02 13:57:49 <ncopa> md0 needs to be raid1
2016-05-02 13:57:51 <mosez> because i don't wanted to waste time on creating all the raid and mdadm stuff manually
2016-05-02 13:58:07 <ncopa> syslinux cannot boot raid10 afaik
2016-05-02 13:58:10 <mosez> ok... so i need to create it manually
2016-05-02 13:58:22 <ncopa> yes, at least /boot
2016-05-02 13:59:05 <ncopa> it might be that you need to use --metadata=0.90 or what its called in mdadm
2016-05-02 13:59:15 <ncopa> so that the metadata is at the end of the partition
2016-05-02 13:59:45 <ncopa> that way syslinux can read one of raids (/dev/sda or /dev/sdb)
2016-05-02 14:02:58 <mosez> ok let me try that
2016-05-02 14:05:11 <leo-unglaub> hiawatha 10.2 is available :)
2016-05-02 15:22:08 <mosez> fuck... still missing operating system -.-
2016-05-02 15:22:08 <mosez> looks like it even didn't get the bootloader
2016-05-02 15:22:08 <mosez> ncopa: let's see if it works. that's now my script https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d
2016-05-02 15:22:08 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d is the latest state and i'm running out of time with the console :(
2016-05-02 15:22:08 <poptart> Someones host just died
2016-05-02 15:22:19 <mosez> ncopa: any idea what should be missing?
2016-05-02 15:23:48 <ncopa> mosez: i dunno
2016-05-02 15:23:55 <mosez> :(((
2016-05-02 15:24:14 <mosez> slowly it looks like i won't get an alpine base system...
2016-05-02 15:25:28 <ncopa> :-(
2016-05-02 15:27:58 <mosez> nad writing mbr to sda is correct?
2016-05-02 15:28:13 <mosez> s/nad/and/
2016-05-02 15:28:25 <mosez> it's so long ago that i had to do all that manually -.-
2016-05-02 15:29:37 <mosez> need to look for docs with grub
2016-05-02 15:31:49 <mosez> and if that doesn't work i will install opensuse with autoyast -.-
2016-05-02 15:32:03 <mosez> i have to get this damn storage server running today
2016-05-02 15:34:07 <Adran> you know
2016-05-02 15:34:11 <Adran> if this has an internal usb
2016-05-02 15:34:30 <Adran> you could use a flash storage medium to install the Os leaving the raid / disks availble to utilize fully
2016-05-02 15:43:03 <tdtrask>  ncopa: I need some direction on cron/crond #5411 #5462
2016-05-02 15:43:05 <tdtrask> do we roll back to 'cron' or update ACF and add post-upgrade fix?
2016-05-02 15:43:35 <ncopa> fix post-upgrade and acf
2016-05-02 15:45:59 <tdtrask> ncopa: ok, thanks
2016-05-02 15:46:09 <tdtrask> 
ACTION will handle ACF and leave post-upgrade for you :)

2016-05-02 15:46:22 <ncopa> watch out for not updating it in 3.3-stable
2016-05-02 15:46:29 <tdtrask> of course
2016-05-02 15:48:16 <jirutka> ncopa, when you have a time, could you please look at the problem we’ve discussed with fabled around 10.30 UTC here on IRC? it’s related to java-{jre,jdk} metapackages.
2016-05-02 15:52:59 <ncopa> so we need some kind of 'alternatives' functionality
2016-05-02 15:53:28 <ncopa> something that corresponds to gentoo's eselect?
2016-05-02 15:53:30 <jirutka> what do you mean with alternatives?
2016-05-02 15:53:50 <jirutka> yes, that would be ideal
2016-05-02 15:54:12 <ncopa> ubuntus way to handle it: http://askubuntu.com/questions/233190/what-exactly-does-update-alternatives-do
2016-05-02 15:54:45 <ncopa> talked with void linux ppl today the also mentioned 'alternatives'
2016-05-02 15:55:01 <ncopa> can we postpone that problem til after v3.4?
2016-05-02 15:55:22 <jirutka> however, I think that we almost have what we need, the only problem is that `provides` is exclusive
2016-05-02 15:56:13 <ncopa> if apk needs fix then bounce the problem to fabled
2016-05-02 15:56:34 <jirutka> it would be great to have also a tool for switching between different “implementations”, but that’s just a cherry on the top
2016-05-02 15:56:47 <jirutka> okay
2016-05-02 15:56:54 <ncopa> but i'm not convinced we want solve this with provides in apk
2016-05-02 15:57:30 <jirutka> If this will take some time, then sure, it’d be probably better to fix it after v3.4; if so, then I’ll separate cleanup commits from the PR to merge it separately
2016-05-02 15:58:01 <jirutka> so what’s the purpose of the `provides` functionality?
2016-05-02 15:58:03 <ncopa> better solve it properly without rushing it
2016-05-02 15:58:42 <ncopa> i think it was originally for things like
2016-05-02 15:58:55 <ncopa> when moving to mariadb, we added provides=mysql
2016-05-02 15:59:06 <jirutka> I’ve thought that we can solve it in the same way as in Lua, but I didn’t know that one cannot (?) install multiple Lua versions simultanesly
2016-05-02 15:59:10 <ncopa> so if you apk add mysql you get mariadb
2016-05-02 15:59:45 <ncopa> oh, no. the provides was for giving an alias name
2016-05-02 16:00:04 <ncopa> for example provides=so:libsomething.so.1
2016-05-02 16:00:08 <jirutka> aha
2016-05-02 16:00:23 <jirutka> it would be really helpful to document this :) there’s a TODO in the provides section on wiki
2016-05-02 16:00:50 <ncopa> so you could say, i need libsomething.so.1, but i dont care what the name of the package is
2016-05-02 16:01:04 <jirutka> apk supports this?
2016-05-02 16:01:31 <ncopa> that way we could move the library from one package name to another without needing rebuild everything that had depends="something"
2016-05-02 16:01:34 <ncopa> yes
2016-05-02 16:01:47 <jirutka> that’s great, I’ve seen this on Fedora and it’s really useful; when user doesn’t know name of the package, he can just provide name of the binary or .so lib
2016-05-02 16:01:57 <ncopa> yes
2016-05-02 16:02:13 <ncopa> we have only used it for sonames
2016-05-02 16:02:18 <ncopa> and pkg-config
2016-05-02 16:02:31 <ncopa> i think we added pkg-config files with v3.3
2016-05-02 16:04:08 <jirutka> I’m gonna read the conversation we had about this a month ago, to recall how it’s handled for lua
2016-05-02 16:05:03 <ncopa> lua-foomodule has lua5.[1-3]-foomodule subpackages
2016-05-02 16:05:07 <fabled> jirutka, provides without version is not exclusive
2016-05-02 16:05:08 <jirutka> or I’ll separate the cleanup commits for openjdks first, then read it
2016-05-02 16:05:27 <fabled> the prbolem is that then there's no automatic way to elect which package to install
2016-05-02 16:06:03 <ncopa> each lua5.x-foomodule has install_if="$pkgname=$pkgver lua5.x"
2016-05-02 16:06:42 <ncopa> so if both lua-foomodule *and* lua5.x is installed at the same time, then lua5.x-foomodule is pulled in
2016-05-02 16:08:15 <jirutka> the cleanup patches from 66 https://github.com/alpinelinux/aports/pull/67
2016-05-02 16:10:08 <ncopa> pushed poppler/icu update + a bunch of big rebuilds
2016-05-02 16:10:37 <ncopa> fabled: i'd like to move libreoffice to community
2016-05-02 16:10:49 <fabled> ncopa, go ahead
2016-05-02 16:11:41 <ncopa> there are various deps thats only used by libreoffice
2016-05-02 16:11:46 <ncopa> we should probably move those too
2016-05-02 16:16:59 <ncopa> argh
2016-05-02 16:17:08 <ncopa> how can if find out which is the last chromium version?
2016-05-02 16:17:18 <ncopa> i know its not pkgver=49.0.2623.110
2016-05-02 16:20:04 <jirutka> fabled: how about adding something like replaces_priority – provides_priority?
2016-05-02 16:21:02 <fabled> jirutka, i got the same idea earlier. but i'm not yet sure if there's any potential issues with that.
2016-05-02 16:21:17 <fabled> it would probably work out mostly
2016-05-02 16:21:49 <jirutka> how it handles ArchLinux? it may be a good inspiration
2016-05-02 16:22:36 <fabled> iirc, some distro had system wide defaults file that define the package to install for virtual package name if none was requested explicitly
2016-05-02 16:23:26 <fabled> seems on arch it depends on the package install utility you use
2016-05-02 16:28:01 <jirutka> that’s probably a better solution, but provides_priority would be probably simpler
2016-05-02 16:28:50 <Adran> related to apk / apkbuilds. -'s are not really supported in version numbers for some reason
2016-05-02 16:29:59 <jirutka> Adran: apk follows Gentoo naming rules https://devmanual.gentoo.org/ebuild-writing/file-format/
2016-05-02 16:30:10 <jirutka> or at least it’s stated in the source code
2016-05-02 17:06:26 <leo-unglaub> ncopa: i could kiss you for a22b383e
2016-05-02 17:06:32 <leo-unglaub> because it was driving me nuts
2016-05-02 17:07:02 <sjennings> question, http://lists.alpinelinux.org/alpine-devel/5212.html
2016-05-02 17:07:41 <sjennings> i sent that patch out almost two months ago and haven't gotten a response, was wondering if i'm doing it right.
2016-05-02 17:11:43 <jirutka> sjennings: just a detail, global variable realpkgver should be prefixed with `_` → _realpkgver
2016-05-02 17:13:03 <jirutka> and patches for aports should be sent to alpine-aports@lists.alpinelinux.org instead of alpine-devel
2016-05-02 17:13:32 <jirutka> these are then available on http://patchwork.alpinelinux.org/project/aports/list/
2016-05-02 17:14:29 <sjennings> jirutka, ah ok thanks!
2016-05-02 17:15:04 <ncopa> sjennings: sorry for not responing
2016-05-02 17:15:08 <jirutka> other way is to send patches via GitHub pull requests https://github.com/alpinelinux/aports
2016-05-02 17:15:28 <jirutka> * other → alternative
2016-05-02 17:16:34 <ncopa> sjennings: oh, one thing more, send patches to alpine-aports@lists.alpinelinux.org
2016-05-02 17:16:50 <ncopa> then they will get tracked by patchwork.alpinelinux.org
2016-05-02 17:18:26 <sjennings> ncopa, jirutka ok thanks for the direction
2016-05-02 17:19:11 <ncopa> sjennings: if you rebase that patch and then do: git format-patch --stdout -1 | tpaste
2016-05-02 17:19:15 <ncopa> and paste the url here
2016-05-02 17:19:21 <ncopa> then i can apply it now
2016-05-02 17:19:34 <sjennings> ncopa, ok i'll do that now
2016-05-02 17:19:44 <ncopa> you might need: apk add tpaste
2016-05-02 17:21:17 <sjennings> ncopa, jirutka just fyi, i was following this https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Send_a_patch which says to sent to alpine-devel.  so that may need updating.
2016-05-02 17:22:01 <ncopa> yes
2016-05-02 17:22:04 <ncopa> i'll fix that now
2016-05-02 17:22:14 <jirutka> too late, I’m already fixing it XD
2016-05-02 17:22:22 <ncopa> thanks!
2016-05-02 17:28:54 <sjennings> ncopa, i'm on fedora. there doesn't seem to be a package for tpaste. i have fpaste (fedora pastebin). that work?
2016-05-02 17:29:21 <ncopa> sure, as long as i can read it as clear text without html
2016-05-02 17:29:35 <ncopa> i can probably copy paste it manually too
2016-05-02 17:29:38 <sjennings> ncopa, http://paste.fedoraproject.org/361858/10155146/raw/
2016-05-02 17:30:21 <sjennings> ncopa, oh wait hold on that
2016-05-02 17:30:36 <jirutka> heh, everyone have own pastebin, we have http://haste.fit.cvut.cz/ at FIT CTU :P
2016-05-02 17:30:36 <sjennings> forgot the realpkgver -> _realpkgver
2016-05-02 17:31:05 <ncopa> jirutka: did you write that?
2016-05-02 17:31:16 <jirutka> sjennings: you could also rename _builddir to builddir, it has been changed in abuild recently
2016-05-02 17:31:32 <sjennings> jirutka, ah ok, i'll update that too
2016-05-02 17:32:09 <jirutka> ncopa: nope, I’ve forked existing project and refactored some stuff; https://github.com/jirutka/haste-server
2016-05-02 17:32:41 <ncopa> looks like node :)
2016-05-02 17:32:49 <jirutka> yeah… and the code is pretty bad :/
2016-05-02 17:33:13 <ncopa> clandmeters paste in lua: https://github.com/clandmeter/turbo-paste
2016-05-02 17:33:20 <ncopa> thats tpaste
2016-05-02 17:33:33 <jirutka> I know, I’ve looked at it few minutes ago, it looks interesting
2016-05-02 17:34:38 <ncopa> i wonder if/how to implement it with lua directly in nginx
2016-05-02 17:34:53 <jirutka> they seems to be the same in the core functionality (simple pasting from CLI), but Hastebin has also nice UI
2016-05-02 17:35:18 <jirutka> that should be possible, thanks to great ngx-lua
2016-05-02 17:35:35 <jirutka> I wrote https://github.com/jirutka/ngx-oauth
2016-05-02 17:36:11 <jirutka> however, I don’t think that writing _applications_ on top of ngx-lua is a good way
2016-05-02 17:36:32 <jirutka> web server/proxy should be just a web server/proxy, not an application server
2016-05-02 17:42:49 <ncopa> sjennings: do you have a new paste or i just use the first one?
2016-05-02 17:43:34 <sjennings> ncopa, i'll get you a new paste with the realpkgver and builddir changed.  i'm rebuilding and retesting just for good measure.
2016-05-02 17:44:13 <ncopa> oh! thats good!
2016-05-02 18:01:08 <ncopa> i will soon have to leave thoush
2016-05-02 18:02:16 <sjennings> ncopa,
2016-05-02 18:02:30 <sjennings> ok, i'll just paste now and let you know if i hit any issues
2016-05-02 18:02:36 <ncopa> ok
2016-05-02 18:02:44 <jirutka> ncopa: before leaving, could you please look at https://github.com/alpinelinux/aports/pull/67? openjdk8 is green, but I don’t know why openjdk7 fails on Travis, it surely works on my machine (maybe just some travis-related problem); fabled has already looked at these and said that they are okay
2016-05-02 18:03:47 <sjennings> ncopa, http://paste.fedoraproject.org/361868/14622122/raw/
2016-05-02 18:03:55 <sjennings> thanks again!
2016-05-02 18:04:46 <ncopa> jirutka: my build server is busy :-/
2016-05-02 18:04:52 <ncopa> opendj7 will take a while
2016-05-02 18:05:14 <jirutka> yeah, openjdk8 took ~30 minutes on Travis :/
2016-05-02 18:05:25 <ncopa> ah thats why it fails
2016-05-02 18:05:30 <ncopa> it takes more than 1 hour
2016-05-02 18:05:35 <ncopa> openjdk8 is fast
2016-05-02 18:05:40 <ncopa> if you compare with openjdk7
2016-05-02 18:05:54 <ncopa> i think its 5 mins or so on my builder
2016-05-02 18:06:18 <jirutka> nope, openjdk8 is fine on travis, it took 30 min and finished
2016-05-02 18:06:24 <jirutka> openjdk7 didn’t even started
2016-05-02 18:06:38 <ncopa> hm
2016-05-02 18:06:39 <ncopa> ok
2016-05-02 18:06:47 <jirutka> https://travis-ci.org/alpinelinux/aports/builds/127284402
2016-05-02 18:06:57 <jirutka> maybe you can apply just openjdk8 patch for now?
2016-05-02 18:08:05 <ncopa> i should push both today
2016-05-02 18:08:09 <ncopa> so they are ready tm morning
2016-05-02 18:08:13 <jirutka> okay
2016-05-02 18:08:29 <jirutka> you’re gonna release v3.4 tomorrow?
2016-05-02 18:08:30 <ncopa> chromium build soon done: [18141/18541]
2016-05-02 18:08:34 <ncopa> no
2016-05-02 18:09:01 <ncopa> i have a bunch of things to fix
2016-05-02 18:09:14 <ncopa> we should go over the bugs in bugs.a.o
2016-05-02 18:09:18 <ncopa> and do the important stuff
2016-05-02 18:09:37 <jirutka> okay
2016-05-02 18:10:04 <ncopa> i have some other thins that i need to do before the release
2016-05-02 18:10:13 <ncopa> but i dont know how i will get the time :-/
2016-05-02 18:11:01 <jirutka> what things? can I help with some of them?
2016-05-02 18:12:08 <ncopa> busybox related
2016-05-02 18:12:25 <ncopa> i groupadd useradd that is portable
2016-05-02 18:12:38 <ncopa> i already sent patch for groupadd but no response
2016-05-02 18:12:46 <skarnet> give Denys some time
2016-05-02 18:12:51 <skarnet> you know he's slow but will get to it
2016-05-02 18:13:04 <ncopa> like alpine project patch handling :)
2016-05-02 18:13:08 <jirutka> so C stuff, that’s unfortunately not something I can help with
2016-05-02 18:13:17 <ncopa> yes, its mostly C stuff
2016-05-02 18:13:47 <ncopa> jirutka: but you could help me look over the bugs.a.o and try move unimportant stuff to 3.4
2016-05-02 18:13:52 <ncopa> or things that can wait
2016-05-02 18:13:52 <skarnet> I want to be next year already, to be able to help with that kind of thing -_-
2016-05-02 18:14:09 <jirutka> actually, Alpine patch handling is much faster than Gentoo… well, when you communicate on IRC (i didn’t tried that on Gentoo)
2016-05-02 18:15:43 <ncopa> jirutka: you can help me figure out why those packages does not build: http://tpaste.us/2z91
2016-05-02 18:15:50 <chris|> depends if anyone is interested in your patches though, just have a look at the patchwork
2016-05-02 18:16:20 <ncopa> jirutka: build logs: http://build.alpinelinux.org/buildlogs/build-3-4-x86_64/main/
2016-05-02 18:16:44 <ncopa> jirutka: you could also look over testing/*
2016-05-02 18:16:54 <ncopa> and see if there are something that we can move to community
2016-05-02 18:17:07 <ncopa> or maybe even main/ if its something we want support for 2 years
2016-05-02 18:17:50 <sjennings> ncopa, package built and installed fine. so good to go :)
2016-05-02 18:18:04 <ncopa> it built for me
2016-05-02 18:18:09 <ncopa> so i ushed it
2016-05-02 18:18:11 <ncopa> pushed
2016-05-02 18:18:17 <ncopa> i should really go now...
2016-05-02 18:18:18 <sjennings> ncopa, awesome, thanks!
2016-05-02 18:18:21 <sjennings> later
2016-05-02 18:19:41 <ncopa> oh build server is building libreoffice + webkit (x2) + linking chromium at the same time
2016-05-02 18:19:47 <ncopa> no wonder if feels slow...
2016-05-02 18:19:51 <jirutka> ldapvi depends on ncruses, when I add ncurses-dev to makedepends, then it build fine
2016-05-02 18:20:11 <ncopa> jirutka: can you do a long pull request
2016-05-02 18:20:18 <jirutka> okay
2016-05-02 18:20:22 <ncopa> so i can pull in the fixes in one shot
2016-05-02 18:30:05 <swswswx> is there any areas with alpine that I can volunteer some time and ansible playbooks for?
2016-05-02 18:30:39 <jirutka> uh, ansible…
2016-05-02 18:44:39 <mosez> adran: if it got a usb drive, yes...
2016-05-02 19:10:26 <mosez> so i really failed to get alpine running on a hetzner server...
2016-05-02 19:25:59 <chris|> why is that?
2016-05-02 19:27:31 <mosez> because i don't get it booting?
2016-05-02 19:29:17 <chris|> again: why? :D
2016-05-02 19:30:57 <chris|> I am willing to bet that noone can help you without you describing your issue in a bit more detail ;)
2016-05-02 19:32:04 <mosez> chris|: i have written about that half of the day :)
2016-05-02 19:32:36 <mosez> now i don't have access to the virtual console anymore. last thing was that it looked like the bootloader have not been detected
2016-05-02 19:34:40 <jirutka> it seems that gtk+3.0 package is quite broken and it breaks packages that depends on it http://haste.fit.cvut.cz/kuyupec.log
2016-05-02 19:36:12 <Adran> x stuff is very broken in general
2016-05-02 19:36:13 <Adran> :(
2016-05-02 19:37:30 <chris|> sorry, I don't follow IRC closely while I'm at work
2016-05-02 19:38:34 <chris|> also you can just request the LARA again if they take it back
2016-05-02 19:40:00 <mosez> chris|: https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d this is the current script... i took lara 2 times today...
2016-05-02 19:48:37 <chris|> did you check that your extlinux.conf is correct?
2016-05-02 19:49:20 <chris|> also: why do you copy the kernel and initramfs from the cd
2016-05-02 19:50:11 <mosez> https://gist.github.com/tboerger/79d0d87643927fc955b8a030d0122b2d that's the extlinux.conf
2016-05-02 19:50:38 <mosez> i've copied the kernel and initramfs because boot is empty otherwise.
2016-05-02 19:50:55 <mosez> or just the stuff copied over because of extlinux
2016-05-02 19:51:58 <mosez> chris|: maybe i just did not find proper guides/docs, this is what i've tried so far.
2016-05-02 19:52:10 <chris|> --allow-untrusted is uncool in any case
2016-05-02 19:52:30 <mosez> this is also from the alpine wiki :)
2016-05-02 19:53:53 <mosez> is update-extlinux all i need to execute? or also extlinux --raid --install?
2016-05-02 19:54:41 <chris|> well, you don't install the kernel and mkinitfs package, so naturally /boot is empty ;)
2016-05-02 19:55:09 <mosez> ok... that was also something missing in the guides i read :D
2016-05-02 19:55:43 <mosez> so let'S give it another try...
2016-05-02 19:56:16 <chris|> have you seen the syslinux boot menu on the LARA?
2016-05-02 19:57:09 <mosez> http://cl.ly/0H1e2c2K1y2f that's what i have seen
2016-05-02 19:58:34 <chris|> okay, that means it does not even find the boot loader
2016-05-02 20:01:08 <jirutka> ncopa, do you want to keep https://pkgs.alpinelinux.org/package/v3.3/main/x86_64/nagiosql in the master? this version is old as hell (around 2010) and the project is quite dead already
2016-05-02 20:01:38 <mosez> chris|: is the dd at the end of the script correct?
2016-05-02 20:01:49 <mosez> chris|: and also the exlinux command?
2016-05-02 20:02:00 <mosez> s/exlinux/extlinux/
2016-05-02 20:02:03 <chris|> maybe you should set your disks to bootable with fdisk
2016-05-02 20:02:15 <chris|> not sure if the hetzner system relies on it
2016-05-02 20:02:33 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d#file-provision-sh-L19-L21
2016-05-02 20:03:22 <chris|> yes, the dd should be ok
2016-05-02 20:03:51 <mosez> ah... i think it's because of the partition table... can that be the case?
2016-05-02 20:04:14 <mosez> i have to take DOS and not GPT partition tables?
2016-05-02 20:04:30 <mosez> sorry but it's ages ago that i did that manually :(
2016-05-02 20:08:46 <chris|> ah, yes, indeed
2016-05-02 20:08:50 <mosez> forget it... gpt should be ok
2016-05-02 20:09:01 <mosez> The GPT partition table is usually used in larger then 2TiB drives (e.g. 3TB HDDs in EX4 and EX6)
2016-05-02 20:09:15 <chris|> if you use gpt you need to create a special mgr partition at the beginning and put the mbr on that
2016-05-02 20:09:41 <chris|> using dd on the disk itself will probably ruin the table with gpt
2016-05-02 20:10:06 <chris|> at least if I remember right
2016-05-02 20:11:37 <jirutka> ncopa: I’m fixing packages from http://tpaste.us/2z91, this is work-in-progress PR https://github.com/alpinelinux/aports/pull/68
2016-05-02 20:11:39 <chris|> https://wiki.archlinux.org/index.php/Syslinux#GUID_partition_table
2016-05-02 20:12:09 <jirutka> audacious and ghostscript fails because of gtk+3.0
2016-05-02 20:14:28 <mosez> chris|: that makes sense... and hopefully that will be the key
2016-05-02 20:17:25 <mosez> chris|: i still have to use /dev/sda --attributes=1:set:2 even if the boot partition is raid1, right?
2016-05-02 20:18:05 <chris|> buest guess: yes, on all your sd{a,b,c,d}
2016-05-02 20:18:21 <mosez> thanks, will update the script and retry it
2016-05-02 20:46:12 <mosez> the damn partition tables are not getting written...
2016-05-02 20:46:19 <mosez> Re-reading the partition table failed.: Device or resource busy
2016-05-02 21:03:32 <chris|> sorry, can't really help you with that
2016-05-02 21:04:13 <chris|> also need to get some sleep ;)
2016-05-02 21:04:39 <mosez> hum... need to search the kernel name :)
2016-05-02 21:06:35 <mosez> chris|: i have to use installkernel?
2016-05-02 21:09:45 <mosez> ah linux- :D
2016-05-02 21:18:54 <mosez> ok, still not available. got to order the console one more time tomorrow...
2016-05-02 21:25:33 <jirutka> I’m idiot, I’ve used exit 0 instead of return 0 in build function… o.O
2016-05-02 22:46:37 <jirutka> ncopa: all the packages from your list fixed https://github.com/alpinelinux/aports/pull/68 :)
2016-05-02 23:26:38 <coredumb> how should one name python3 module ? py3-name ?
2016-05-03 00:00:55 <coredumb> or name-python3 ?
2016-05-03 02:37:30 <Adran>  /det
2016-05-03 07:37:04 <mosez> so lets get to the next try to fix the alpine boot
2016-05-03 07:55:54 <ScrumpyJack> morning
2016-05-03 07:59:59 <mosez> damn... system is still complaining about missing os :(
2016-05-03 08:52:45 <mosez> YES!!!
2016-05-03 08:52:59 <mosez> now it's failing for another reason but i've mastered a step :D
2016-05-03 08:53:14 <mosez> now i need to figure out why it won't work to have the root on lvm
2016-05-03 08:53:18 <ammunta> not it's failing for a *different reason*!!!! YES!!!
2016-05-03 08:53:21 <ammunta> :))
2016-05-03 08:54:08 <mosez> it's a mess to get alpine running on hetzner servers
2016-05-03 08:54:32 <mosez> https://gist.github.com/tboerger/0cd5fde88092b209e56c32254ebc922d
2016-05-03 08:56:40 <ammunta> http://sprunge.us/KHHj these are my notes from last time I did something similar (not hetzner)
2016-05-03 08:57:05 <ammunta> no comment on whether it works
2016-05-03 08:57:10 <ammunta> it probably didn't
2016-05-03 08:57:30 <ammunta> once I got it working I decided I wasn't ever doing that again
2016-05-03 08:58:01 <ammunta> I wasn't doing any raid though
2016-05-03 08:58:21 <ammunta> that would probably have been harder
2016-05-03 08:58:58 <mosez> and for gpt partition tables it changes quite a lot as well ;)
2016-05-03 08:59:13 <mosez> and yeah... software raid and on top of it lvm :)
2016-05-03 09:00:22 <ammunta> mind, I just splatted alpine userspace all over the recovery system to avoid mounting stuff
2016-05-03 09:01:36 <ammunta> haven't tried it with alpine specifically but even manually installing on gpt tables didn't seem that much of a hassle, just need someone else to have done the actually hard parts first :&&
2016-05-03 09:01:48 <mosez> my mkinitfs is missing lvm. so hopefully adding that and executing extlinux -i will fix it
2016-05-03 09:02:26 <ammunta> obviously that'll just make it fail in another new and exciting way right
2016-05-03 09:05:34 <mosez> and obviously i did it wrong again -.-
2016-05-03 09:05:45 <ammunta> did it fail in a new and exciting way
2016-05-03 09:06:02 <mosez> https://wiki.alpinelinux.org/wiki/Setting_up_disks_manually#LVM exactly say i need to do mkinitfs -c $MNT/etc/mkinitfs/mkinitfs.conf -b $MNT :(
2016-05-03 09:12:40 <mosez> nearby... it's booting but ssh failed to start because /etc seems to be readonly :)
2016-05-03 09:20:30 <mosez> damn... i forgot to setup networking -.-
2016-05-03 09:33:38 <mosez> oh god... cannot find device "eth0" oO
2016-05-03 09:44:18 <mosez> looks like my network card doesnt get detected?
2016-05-03 09:44:26 <mosez> ncopa: is that right?
2016-05-03 09:45:38 <skarnet> could be a lot of things. Either your network card isn't detected, or the module isn't properly loaded, or...
2016-05-03 09:45:50 <skarnet> check your kernel logs.
2016-05-03 09:51:03 <mosez> at least 03:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 09) doesnt look special
2016-05-03 09:51:22 <skarnet> lsmod | grep rtl ?
2016-05-03 10:43:30 <mosez> hum... still no luck :(
2016-05-03 10:43:42 <mosez> i will loose my virtual console soon
2016-05-03 10:47:56 <mosez> ncopa: any other idea why eth0 is an unknown device?
2016-05-03 10:48:19 <mosez> within messages and dmesg i can not find anything network related.
2016-05-03 10:48:25 <mosez> rtl is not a module.
2016-05-03 11:43:06 <coredumb> Hey folks
2016-05-03 11:43:23 <coredumb> what's the naming convention for python3 packages ?
2016-05-03 11:43:35 <jirutka1> hi
2016-05-03 11:43:50 <jirutka1> it’s not specified yet, as I know… but it’ll probably be py3-
2016-05-03 11:44:18 <coredumb> I've made an APKBUILD for borgbackup which required to rebuild py-msgpack with python3
2016-05-03 11:44:50 <coredumb> as myspack is officially called msgpack-python I named th pkg msgpack-python3
2016-05-03 11:44:53 <coredumb> well
2016-05-03 11:45:03 <coredumb> was wondering about all that before sending patched
2016-05-03 11:45:04 <jirutka> please stick to py3-*
2016-05-03 11:45:22 <coredumb> fine by me
2016-05-03 11:45:23 <jirutka> python 2 uses prefix py-*
2016-05-03 11:45:45 <coredumb> did sound logical ... but it got late and redid it as -python3 :P
2016-05-03 11:47:48 <jirutka> we have a lot of work with preparing v3.4 release now, so it’s not an ideal time for this
2016-05-03 11:52:17 <jirutka> afk, I have meeting now
2016-05-03 11:52:46 <coredumb> yep that's what I thought
2016-05-03 11:53:01 <coredumb> I'll send it all clean after the release of 3.4 :)
2016-05-03 13:47:56 <jirutka> ncopa?
2016-05-03 14:29:48 <fcolista> required by: php-intl-5.6.21-r0[so:libicudata.so.56]
2016-05-03 14:29:53 <fcolista> php-intl is broken
2016-05-03 14:29:58 <fcolista> on edge
2016-05-03 14:37:59 <fcolista> ok
2016-05-03 14:37:59 <fcolista>  required by: php-intl-5.6.21-r0[so:libicudata.so.56]
2016-05-03 14:38:04 <fcolista> we have .57
2016-05-03 14:38:14 <fcolista> looks we need to rebuild php
2016-05-03 14:55:31 <ncopa> i thought i rebuilt it
2016-05-03 14:56:40 <fcolista> i've just bumped pkgrel
2016-05-03 14:56:50 <ncopa> thanks
2016-05-03 14:57:13 <ncopa> we need rebuild chromium too
2016-05-03 14:57:21 <ncopa> but it also needs update
2016-05-03 14:57:27 <ncopa> but update breaks
2016-05-03 14:57:32 <ncopa> needs a few patches i think
2016-05-03 14:59:48 <jirutka> ncopa: this one is waiting for you https://github.com/alpinelinux/aports/pull/68 ;) travis failed because build of qt5-qtwebkit takes too much time
2016-05-03 15:02:53 <ncopa> i wonder what causes all those incorrect checksums
2016-05-03 15:03:04 <jirutka> me too, it’s really strange
2016-05-03 15:03:51 <ncopa> but
2016-05-03 15:04:01 <ncopa> edge builder may fail
2016-05-03 15:04:06 <ncopa> since it has the old tarball
2016-05-03 15:04:13 <ncopa> with the previous checksum
2016-05-03 15:04:46 <jirutka> when the abuild fails to verify the checksum, then it deletes the tarball…no?
2016-05-03 15:05:24 <jirutka> so when you triet it another time, it should download a new tarball and it should succeed?
2016-05-03 15:05:34 <ncopa> it does not
2016-05-03 15:05:44 <ncopa> does not del tarball
2016-05-03 15:05:53 <jirutka> then abuild lies
2016-05-03 15:06:06 <jirutka> because it prints a message that says that
2016-05-03 15:06:06 <ncopa> yes
2016-05-03 15:06:10 <ncopa> we shoudl fix that
2016-05-03 15:06:24 <jirutka> and i think that it’s a reasonable thing to do when checksum fails
2016-05-03 15:06:46 <ncopa> the thing is that we cache it
2016-05-03 15:06:52 <ncopa> so it picks up the old file from cache
2016-05-03 15:06:58 <jirutka> checksum failure should invalidate the cache ;)
2016-05-03 15:07:07 <ncopa> the v3.4 stores files in other directory
2016-05-03 15:07:23 <ncopa> so we have the exact tarballs we used at build time
2016-05-03 15:07:32 <ncopa> in case upstream remove them or change them
2016-05-03 15:07:59 <ncopa> you can find the cached files on http://distfiles.alpinelinux.org/distfiles/
2016-05-03 15:08:41 <jirutka> well, there are only few abuilds with this problem, so what about manually removing the cached files?
2016-05-03 15:08:51 <ncopa> that what i have done
2016-05-03 15:09:01 <ncopa> but normally i manually untar them both
2016-05-03 15:09:07 <ncopa> and run a diff
2016-05-03 15:09:11 <ncopa> to figure out what changed
2016-05-03 15:09:43 <jirutka> what do you usually find out in these cases?
2016-05-03 15:09:58 <ncopa> sometimes ABPKUILD has bad soruce url
2016-05-03 15:10:09 <ncopa> which may be an autogenerated tarball from cgit or similar
2016-05-03 15:10:19 <ncopa> or maybe even autogenerates tarball from git master
2016-05-03 15:10:27 <ncopa> so next commit the tarball changes....
2016-05-03 15:10:44 <jirutka> when it’s autogenerated from a tag in git, then it *should* be immutable as well
2016-05-03 15:10:54 <jirutka> the same as tarballs
2016-05-03 15:10:59 <ncopa> depends
2016-05-03 15:11:11 <ncopa> gzip stores a timestamp
2016-05-03 15:11:28 <jirutka> well, at least at GitHub this works
2016-05-03 15:11:34 <ncopa> i used bzip2 some time with cgit
2016-05-03 15:11:49 <ncopa> but tarball changed when i upgraded cgit
2016-05-03 15:11:57 <ncopa> github works because i think they cache it
2016-05-03 15:12:13 <ncopa> i have asked them specifically about this issue
2016-05-03 15:12:16 <jirutka> yes, because tagged releases should be immutable
2016-05-03 15:12:17 <ncopa> and they say it will not change
2016-05-03 15:13:00 <ncopa> for our git.a.o i made a git hook that will generate a tarball on git tags
2016-05-03 15:13:08 <ncopa> and store those on dev.a.o/archive/
2016-05-03 15:36:38 <jirutka> ncopa: should I close https://github.com/alpinelinux/aports/pull/68 ?
2016-05-03 15:37:05 <ncopa> jirutka: yes please
2016-05-03 16:06:25 <leo-unglaub> hey friends :)
2016-05-03 16:06:29 <leo-unglaub> how are you?
2016-05-03 17:00:07 <coredumb> o/ leo-unglaub
2016-05-03 17:07:54 <jirutka> ncopa: do you use some tool for monitoring releases of project we have abuild for?
2016-05-03 17:10:50 <scadu> jirutka: uggedal has aports-vercmp to compare pkgs version in aports with upstream or Arch, but this repository is not available now.
2016-05-03 17:11:19 <scadu> or you mean something different.
2016-05-03 17:11:33 <jirutka> I’m thinking about integration with Anitya https://release-monitoring.org/
2016-05-03 17:16:09 <scadu> I doubt it is necessary now. I'd add to planning room :f
2016-05-03 17:16:27 <scadu> As you know, there is a lot more important things to do.
2016-05-03 17:16:28 <jirutka> where’s a planning room? :P
2016-05-03 17:16:34 <jirutka> yeah, I know
2016-05-03 17:17:38 <leo-unglaub> jirutka: planing room -> /dev/null
2016-05-03 17:17:39 <leo-unglaub> *g+
2016-05-03 17:17:49 <jirutka> LOL
2016-05-03 17:17:55 <scadu> with queue in /tmp
2016-05-03 17:17:57 <scadu> maybe.
2016-05-03 17:18:55 <scadu> leo-unglaub: +1
2016-05-03 17:19:19 <jirutka> ah, here it is https://wiki.alpinelinux.org/wiki/Wishlist
2016-05-03 17:19:40 <jirutka> but quite outdated
2016-05-03 17:19:51 <scadu> I didn't know about it.
2016-05-03 17:25:37 <jirutka> there’s something wrong… https://pkgs.alpinelinux.org/package/edge/testing/x86_64/jruby should have more than one dependency https://github.com/alpinelinux/aports/blob/master/testing/jruby/APKBUILD
2016-05-03 17:26:30 <jirutka> ah, I see, why the heck I add depends= to package()
2016-05-03 17:29:29 <scadu> jirutka: commiting at night? :>
2016-05-03 17:29:37 <jirutka> yeah, probably…
2016-05-03 17:30:36 <scadu> that's why I wait until morning :P
2016-05-03 17:31:05 <jirutka> I sleep at morning :)
2016-05-03 18:01:56 <jirutka> what does “fatal: No names found, cannot describe anything.” mean?
2016-05-03 18:02:34 <jirutka> it seems that abuild prints that when updating index
2016-05-03 18:15:26 <poptart> Does anyone else have issues with the packaged version of gradm not actually doing any learning?
2016-05-03 18:24:42 <chris|> mosez: so, any luck?
2016-05-03 18:31:28 <jirutka> if anyone wants to install GitLab CE on Alpine, there’s an abuild for it! https://github.com/jirutka/aports/tree/edge/bundles/gitlab-ce :)
2016-05-03 18:39:16 <chris|> jirutka: very cool, your .gitignore is missing a **/src/* though ;)
2016-05-03 18:39:43 <jirutka> ah, thanks, I forgot to it!
2016-05-03 18:40:06 <skarnet> you accidentally a word, too!
2016-05-03 18:40:19 <jirutka> ?
2016-05-03 18:40:44 <skarnet> "I forgot to it". You accidentally a word.
2016-05-03 18:41:31 <skarnet> ... nevermind. :P
2016-05-03 18:42:02 <jirutka> ah, my English kinda sucks :/
2016-05-03 18:42:25 <skarnet> don't worry, my sentence was (purposefully) broken English.
2016-05-03 18:45:37 <chris|> but since you are building with --allow-untrusted, I am probably not going to trust your repo
2016-05-03 18:45:49 <chris|> s/installing
2016-05-03 18:46:02 <jirutka> you’re wrong, read it more carefully…
2016-05-03 18:46:58 <przemoc> hi, Alpiners.  I'm back in town.  how is 3.4 progressing?  I see it's still not released yet at least. :)  I hope to fiddle w/ AL tomorrow a bit.
2016-05-03 18:46:59 <jirutka> the only package I’m installing with allow-untrasted is alpine-base
2016-05-03 18:47:11 <jirutka> hm, but you’re right, I should do it better
2016-05-03 18:49:50 <jirutka> I actually already wrote this in more secure way, for lxc template, don’t remember why I didn’t used it here
2016-05-03 18:49:56 <chris|> alpine-base pulls in alpine-keys untrusted after which you are free to omit key verification anyways
2016-05-03 18:50:40 <jirutka> well, but the packages are fetched from my mirror with HTTPS
2016-05-03 18:51:12 <jirutka> I think that it’s ok-ish for that purpose, but I agree that I should make it more secure
2016-05-03 18:51:38 <chris|> does your mirror verify package integrity? if yes, then it would be ok
2016-05-03 18:52:32 <jirutka> hm, no, it’s just a sync from rsync.a.o
2016-05-03 18:54:44 <chris|> well, in the end, something always has to give. I checked in the alpine dev keys and pull them via git over ssl, which is also not necessarily the best
2016-05-03 18:55:47 <chris|> fabled: I wanted to ask you, is it intentional that "apk --version" returns a non-zero exit code?
2016-05-03 18:55:58 <fabled> chris|, no, it's fixed ingit
2016-05-03 18:56:19 <fabled> http://git.alpinelinux.org/cgit/apk-tools/commit/?id=f9b4fb469b3df10c3cdfb0740225dd430bf90565
2016-05-03 18:56:44 <chris|> yeah, just found that, thanks :)
2016-05-03 18:57:21 <fabled> should probably fix it in the packages too
2016-05-04 09:42:47 <leo-unglaub> hey :)
2016-05-04 09:55:20 <scadu> hi.
2016-05-04 10:11:28 <leo-unglaub> i am sooo exited :)
2016-05-04 10:11:41 <leo-unglaub> i finally got an internet connection with 75mbit/sec :)
2016-05-04 10:11:43 <leo-unglaub> sooo fun!!!
2016-05-04 10:12:03 <przemoc> and uplink equally good? ;)
2016-05-04 10:15:32 <leo-unglaub> its okay, 7,5mbit/sec
2016-05-04 10:15:41 <leo-unglaub> better than my 90kbit/sec before!
2016-05-04 10:16:02 <przemoc> definitely
2016-05-04 10:16:06 <leo-unglaub> i am currently installing every alpine package i can find ... just because i can !!!!!
2016-05-04 10:16:07 <leo-unglaub> :)
2016-05-04 10:17:21 <przemoc> and so leo-unglaub's AL minimal setup size has been lost...
2016-05-04 10:17:37 <leo-unglaub> przemoc: dont worry, its on my girlfriends laptop *g*
2016-05-04 10:17:45 <leo-unglaub> that one has to be reinstalled anyway *g*
2016-05-04 10:17:56 <scadu> leo-unglaub: she's using Alpine? :>
2016-05-04 10:17:58 <scadu> woohoo.
2016-05-04 10:18:02 <suqdiq> is there a date for 3.4 release ?
2016-05-04 10:18:17 <przemoc> May 2 IIRC :>
2016-05-04 10:18:18 <suqdiq> shit i need to finish my rsyslog package updates in time for it :x
2016-05-04 10:18:21 <leo-unglaub> yes, she is ... sometimes ... she can double boot win 7 and alpine
2016-05-04 10:18:22 <suqdiq> o
2016-05-04 10:18:27 <suqdiq> crap, time flys
2016-05-04 10:18:29 <leo-unglaub> sadly she chooses win7 way to often
2016-05-04 10:18:37 <scadu> is she*
2016-05-04 10:18:39 <scadu> :c
2016-05-04 10:19:29 <przemoc> don't blame her.  solitaire is not that convincing on linux.
2016-05-04 10:19:42 <suqdiq> yeah and freecell/minesweeper too
2016-05-04 10:20:07 <scadu> and indesign.
2016-05-04 10:20:09 <scadu> oh wait.
2016-05-04 10:21:22 <przemoc> https://appdb.winehq.org/objectManager.php?sClass=application&iId=755  apparently it's not that bad
2016-05-04 10:23:08 <scadu> przemoc: just an example :p
2016-05-04 10:25:06 <skarnet> those were some convincing examples of everyday sexism, guys
2016-05-04 10:25:13 <skarnet> even if I know you were joking, that's not cool
2016-05-04 10:35:22 <przemoc> well...  it's your misinterpretation (at least of my words).  if we would talk about someone's boyfriend in the same shoes, I wouldn't hesitate to use the same joke.  it's not sex (or gender, if you prefer) oriented, only OS preference oriented.  but I apologize if that kind of jesting is unwelcome here.
2016-05-04 10:35:37 <suqdiq> yeah replace she with he, same shit
2016-05-04 10:35:43 <suqdiq> non-computer people :)
2016-05-04 10:39:18 <skarnet> sorry if I'm a bit too sensitive to those issues.
2016-05-04 10:40:00 <skarnet> (you may joke about me, I use Windows all the time)
2016-05-04 10:40:08 <przemoc> (so do I)
2016-05-04 10:42:33 <scadu> oh noes.
2016-05-04 10:42:38 <scadu> burn them all.
2016-05-04 10:43:23 <scadu> another anti joke.
2016-05-04 10:43:59 <suqdiq> WINDOWS ? WHERE WE'RE GOING, WE DONT NEED NO WINDOWS!
2016-05-04 10:44:40 <scadu> we don't need computers.
2016-05-04 10:45:46 <scadu> move to Alaska and fight with bears.
2016-05-04 10:57:41 <suqdiq> i just said windows :p
2016-05-04 10:57:47 <suqdiq> hard to avoid computers nowadays :)
2016-05-04 11:01:48 <skarnet> not sure you need computers for bear-fighting though
2016-05-04 11:02:40 <scadu> exactly.
2016-05-04 11:03:58 <skarnet> although, with IoT, you never know. "This bear-fighting knife lets you share your hunting exploits live on your Facebook account." @internetofshit
2016-05-04 11:04:53 <suqdiq> i wouldnt fight a bear :) i'd either flee or get eaten
2016-05-04 11:05:16 <przemoc> #InternetOfTurd  FTFY
2016-05-04 11:06:06 <skarnet> przemoc: https://twitter.com/internetofshit :)
2016-05-04 11:06:51 <przemoc> o-kay...
2016-05-04 11:08:09 <scadu> suqdiq: look at Dicaprio in The Revenant
2016-05-04 11:08:46 <scadu> #alpine-devel, yeah
2016-05-04 11:09:35 <skarnet> bear-fighting is always on-topic
2016-05-04 11:09:37 <skarnet> always
2016-05-04 11:11:37 <scadu> dunno when #alpine-devel converted to underground fight club chat :p
2016-05-04 11:22:13 <suqdiq> FIGHT CLUB ? im up for that :d
2016-05-04 11:22:20 <suqdiq> fighting humans better than other animals :)
2016-05-04 11:32:34 <scadu> ncopa: is there any news regarding python transition?
2016-05-04 11:46:33 <ScrumpyJack> afternoon
2016-05-04 12:04:41 <xming> does anyone has monotone AKPBUILD hidden somewhere?
2016-05-04 12:23:33 <suqdiq> monotone
2016-05-04 12:23:33 <suqdiq> ?
2016-05-04 12:31:54 <ncopa> scadu: i dont think we can remove/replace python2
2016-05-04 12:32:03 <ncopa> too much will break
2016-05-04 12:32:16 <ncopa> so i think we need to incrementally add python3
2016-05-04 12:32:27 <jirutka> not remove, just do similar thing as with Lua
2016-05-04 12:32:38 <scadu> ncopa: that's what I wanted to do eventually.
2016-05-04 12:33:34 <ncopa> something similar lua would be great
2016-05-04 12:33:54 <ncopa> i wonder if should rename "python" to python2
2016-05-04 12:34:04 <ncopa> and make it provides=python
2016-05-04 12:40:17 <scadu> hm...
2016-05-04 12:43:29 <jirutka> then we have to solve the same problem as with jdk/jre, so we can kill two birds with one stone :)
2016-05-04 12:45:17 <scadu> ncopa: what have we exactly done with lua? just want to make sure. I didn't find anything in aports quickly.
2016-05-04 12:45:24 <xming> this monotone http://www.monotone.ca/
2016-05-04 12:46:07 <jirutka> scadu: this isn’t exactly the same case, because you don’t install lua5.1 and lua5.2 at the same time, but you sometimes need to install python2 and python3 at the same time
2016-05-04 12:46:40 <xming> there is lua and luajit, which you could install at the same time
2016-05-04 12:47:22 <xming> python has even more, there is also pypy
2016-05-04 12:47:31 <jirutka> right
2016-05-04 12:49:35 <ncopa> scadu: the main point is that we build lua5.1-* lua5.2-* and lua5.3-* from same apkbuild
2016-05-04 12:50:38 <ncopa> simple example: http://git.alpinelinux.org/cgit/aports/tree/main/lua-microlight/APKBUILD
2016-05-04 12:51:51 <xming> not maintaining backward compatibility seems to be the norm these days
2016-05-04 12:52:25 <jirutka> it depends…
2016-05-04 12:53:09 <scadu> ncopa: I see now/.
2016-05-04 12:53:14 <xming> on my gentoo I have ruby[1,2].[0-9] and python-[2,3].[6,7,2,3,4,5]
2016-05-04 12:53:41 <xming> so messy
2016-05-04 12:54:31 <jirutka> why messy? slots works great on Gentoo
2016-05-04 12:57:22 <xming> it works, but it's polluting
2016-05-04 13:38:59 <fcolista> ncopa, where's the page with the status of build servers'
2016-05-04 13:39:03 <fcolista> ?
2016-05-04 13:39:17 <jirutka> fcolista: do you mean this? http://build.alpinelinux.org/
2016-05-04 13:39:34 <fcolista> yeah jirutka
2016-05-04 13:39:35 <fcolista> thx
2016-05-04 13:39:52 <jirutka> np
2016-05-04 14:04:32 <ncopa> i think i will changes the builders to do 'halt-on-error'
2016-05-04 14:15:10 <ScrumpyJack> fcolista: you can also track the build processes with mqtt
2016-05-04 14:15:16 <ScrumpyJack> mosquitto_sub -h msg.alpinelinux.org -t 'build/#' -v
2016-05-04 14:15:59 <fcolista> ScrumpyJack, yes. Thanks for recall it.
2016-05-04 14:18:23 <przemoc> I don't remember whether I asked it already:  why MQTT in particular has been chosen for AL infra?
2016-05-04 14:51:54 <suqdiq> cause MQTT FUCKING ROX!
2016-05-04 14:52:12 <suqdiq> didnt know AL used it :) +1 rep points on my book
2016-05-04 15:24:47 <ScrumpyJack> any alpine on rpi3 users about?
2016-05-04 15:26:26 <ScrumpyJack> does 3.3.3 work on rpi3?
2016-05-04 15:28:54 <ScrumpyJack> i get kernel image not found after a standard untar of rpi-3.3.3
2016-05-04 15:30:39 <ScrumpyJack> (green light blinks 7 times)
2016-05-04 15:32:08 <ScrumpyJack> looks like there are some differences in files. now only one cmdline.txt file needed?
2016-05-04 15:33:01 <ScrumpyJack> and config.txt has changed?
2016-05-04 15:34:02 <ScrumpyJack> i don't see BOOT_IMAGE=/boot/vmlinuz-rpi2 in the new cmdline.txt, is that correct?
2016-05-04 16:14:39 <ScrumpyJack> my bad
2016-05-04 16:15:35 <leo-unglaub> clandmeter: ping
2016-05-04 16:26:38 <ScrumpyJack> fabled: can you show me your config for the rpi with a pitft screen attached? I think i remember you had one working
2016-05-04 17:11:54 <clandmeter> leo-unglaub: pong
2016-05-04 17:12:06 <leo-unglaub> clandmeter: you maintain pkgs.alpinelinux.org, right?
2016-05-04 17:12:17 <clandmeter> yes
2016-05-04 17:12:25 <clandmeter> did you break it?
2016-05-04 17:12:42 <leo-unglaub> hehe, i did not try :)
2016-05-04 17:13:07 <leo-unglaub> i wanted to ask you for a rest api to fetch packge informations in a json format
2016-05-04 17:15:30 <clandmeter> leo-unglaub: ask vkris
2016-05-04 17:15:45 <clandmeter> he has done some dev work on it
2016-05-04 17:15:59 <clandmeter> i think he put it on his github
2016-05-04 17:16:09 <clandmeter> based on phalconphp
2016-05-04 17:18:03 <leo-unglaub> clandmeter: he ordered someone to work on it?
2016-05-04 17:18:12 <leo-unglaub> nah, i want the master himself to proide the json :)
2016-05-04 17:18:37 <clandmeter> https://github.com/insteps/aport-api/commits/master
2016-05-04 17:19:05 <clandmeter> i made an attempt with lua and lua-turbo
2016-05-04 17:19:39 <clandmeter> but it seems a full rest api is more work then i expected
2016-05-04 17:19:51 <leo-unglaub> clandmeter: nah, fuck full rest
2016-05-04 17:20:00 <leo-unglaub> just reading :)
2016-05-04 17:20:17 <leo-unglaub> a master.json that lists all package names and then a <packagename>.json
2016-05-04 17:20:22 <leo-unglaub> thats all i would need :)
2016-05-04 17:20:35 <clandmeter> :)
2016-05-04 17:20:56 <clandmeter> i would prefer to use some standard like jsonapi.org
2016-05-04 17:21:14 <leo-unglaub> the great thing about rest is, that no one who calls it "rest" actually read the rfc and everyone does what they want *g*
2016-05-04 17:21:26 <leo-unglaub> 99% of all "rest apis" out there completely ignore the standard *g*
2016-05-04 17:22:08 <clandmeter> if you want some webinterface to aports which does nothing but what you ask, dont call it rest :p
2016-05-04 17:22:58 <clandmeter> the api is already live with some test data. i bet you can guess the uri.
2016-05-04 17:22:58 <leo-unglaub> okay
2016-05-04 17:23:12 <leo-unglaub> let me rephrase my question
2016-05-04 17:23:36 <leo-unglaub> would you be willing to provide a overview.json and a <package>.json for every package?
2016-05-04 17:25:46 <clandmeter> leo-unglaub: i think if vkris finishes his api, it would provide what you need.
2016-05-04 17:26:04 <leo-unglaub> okay, is it already testable?
2016-05-04 17:26:05 <clandmeter> take a look at api.a.o
2016-05-04 17:26:15 <clandmeter>  /packages
2016-05-04 17:27:25 <leo-unglaub> muuuuaaaahhhhhh
2016-05-04 17:27:26 <leo-unglaub> hahaha
2016-05-04 17:27:58 <leo-unglaub> guys, we really have to do a course in web security or we should stop calling alpine linux secure *g*
2016-05-04 17:28:12 <clandmeter> ?
2016-05-04 17:28:20 <leo-unglaub> the server broadcasts the php version number out in every header ...
2016-05-04 17:28:34 <leo-unglaub> and btw, the server needs an php update based on the header *h*
2016-05-04 17:28:54 <leo-unglaub> expose_php = off
2016-05-04 17:28:54 <ncopa> which server?
2016-05-04 17:29:01 <leo-unglaub> that needs to be in every php config
2016-05-04 17:29:17 <clandmeter> then send us a pr :p
2016-05-04 17:29:26 <ncopa> leo-unglaub: which site?
2016-05-04 17:29:31 <clandmeter> diff patch whatever
2016-05-04 17:29:40 <ncopa> alpinelinux.org should not have any php at all
2016-05-04 17:30:02 <clandmeter> should not or does not?
2016-05-04 17:30:04 <ncopa> the webserver cannot run anything, just serve static pages
2016-05-04 17:30:10 <ncopa> does not
2016-05-04 17:30:11 <leo-unglaub> curl -I https://api.alpinelinux.org/
2016-05-04 17:30:21 <clandmeter> api is not an offical alpine site
2016-05-04 17:30:39 <clandmeter> but i think i already said that.
2016-05-04 17:30:58 <ncopa> ghostscript tarball chage: http://tpaste.us/GZy6
2016-05-04 17:31:13 <ncopa> which is why checksum fails
2016-05-04 17:37:53 <leo-unglaub> clandmeter: i think i can work with the api you gave me
2016-05-04 17:38:27 <leo-unglaub> its all for a stupid bet anyway *g*
2016-05-04 17:38:28 <clandmeter> ok. but like i said, its not ready.
2016-05-04 17:38:45 <clandmeter> and the data is not uptodate
2016-05-04 17:38:52 <leo-unglaub> does not matter :)
2016-05-04 17:39:02 <clandmeter> but for details ask vkris
2016-05-04 17:39:25 <leo-unglaub> a friend of mine was so happy about the new apple xcode and how fast he can drag and drop together an app and i bet him that i can do it faster in vim ...
2016-05-04 17:39:56 <leo-unglaub> now we are coding an app that lists all alpine packages and allows you to flag then as out of date and see who is faster
2016-05-04 17:40:01 <leo-unglaub> me in vim or him in xcode
2016-05-04 17:44:46 <leo-unglaub> ncopa: but dont worry, they propobly never get released *g*
2016-05-04 18:27:06 <fabled> ScrumpyJack, the dtbs?
2016-05-04 18:28:07 <fabled> ScrumpyJack, https://github.com/rameplayerorg/rameplayer-alpine/blob/master/dts/rame-cid4-lcd-overlay.dts
2016-05-04 18:39:12 <ScrumpyJack> fabled: thanks dude
2016-05-04 18:40:46 <fabled> leo-unglaub, we get new record with openssl
2016-05-04 18:40:57 <fabled> 8mins after announce mail I got it flagged as outdated
2016-05-04 18:41:37 <skarnet> fabled: you should be used to that kind of thing with openssl by now :)
2016-05-04 18:41:37 <fabled> i upgraded edge. waited a while. and got another flagging... apparently due to having -stables still not updated
2016-05-04 18:41:51 <fabled> yeah
2016-05-04 18:41:59 <fabled> i think we want libressl after 3.4 is out
2016-05-04 18:42:13 <fabled> there's new major openssl release anyway that is not abi compatible, and requires full os recompile
2016-05-04 18:42:24 <fabled> so probably worthwhile to go libressl while at it
2016-05-04 18:42:37 <ScrumpyJack> did you get boot msgs displaying on the pitft?
2016-05-04 18:44:04 <fabled> the dts needs to be loaded with dtoverlay command in config.txt
2016-05-04 18:44:32 <fabled> after that when hwdetect probes the dts exposed modules i get:
2016-05-04 18:44:44 <fabled> [    1.861726] fbtft: module is from the staging directory, the quality is unknown, you have been warned.
2016-05-04 18:44:44 <fabled> [    1.862978] fb_st7735r: module is from the staging directory, the quality is unknown, you have been warned.
2016-05-04 18:44:44 <fabled> [    1.864106] fbtft_of_value: buswidth = 8
2016-05-04 18:44:44 <fabled> [    1.864122] fbtft_of_value: debug = 0
2016-05-04 18:44:45 <fabled> [    1.864129] fbtft_of_value: rotate = 270
2016-05-04 18:44:47 <fabled> [    1.864137] fbtft_of_value: fps = 25
2016-05-04 18:44:49 <fabled> [    1.864144] fbtft_of_value: txbuflen = 32768
2016-05-04 18:45:01 <fabled> [    2.755967] graphics fb1: fb_st7735r frame buffer, 160x128, 40 KiB video memory, 32 KiB DMA buffer memory, fps=25, spi0.0 at 3
2016-05-04 18:45:01 <fabled> 2 MHz
2016-05-04 18:45:54 <fabled> that's the relevant stuff... i have about 5 peripherals added via dtbs
2016-05-04 18:46:04 <ScrumpyJack> i'm using the overlay that comes in /overlays (pitft-capacitive)
2016-05-04 18:46:08 <fabled> rtc chip, gpio stuff, keybard, ...
2016-05-04 18:46:24 <leo-unglaub> hehehe
2016-05-04 18:46:36 <fabled> push buttons. rotary encoder.
2016-05-04 18:46:44 <fabled> hd audio card
2016-05-04 18:47:08 <ScrumpyJack> i'm using the old pitft for the RPi1 model B, so i don't think your overlay will work for me
2016-05-04 18:47:27 <fabled> right. yes, the dtb is for specific adafruit clone sourced from china
2016-05-04 18:50:21 <ScrumpyJack> do you boot messages show up on the TFT?
2016-05-04 19:15:57 <fabled> no
2016-05-04 19:16:04 <fabled> i turned them off intentionally
2016-05-04 19:16:21 <fabled> you can define which fb device should get message from kernel cmdline though
2016-05-04 19:20:07 <ScrumpyJack> yes, doesn't work for me.
2016-05-04 19:20:34 <ScrumpyJack> do you use X ?
2016-05-04 19:22:29 <fabled> no
2016-05-04 19:22:29 <ScrumpyJack> wifi dongle dies when i unplug ethernet cable!
2016-05-04 19:22:36 <ScrumpyJack> that must be udev right?
2016-05-04 19:23:09 <ScrumpyJack> fabled: what do you use your setup for? sounds interesting
2016-05-04 19:23:34 <fabled> web controlled media player (clips, live streams)
2016-05-04 19:24:16 <ScrumpyJack> omxplayer?
2016-05-04 19:27:42 <fabled> yes, patched to support alsa devices
2016-05-04 19:28:31 <ScrumpyJack> how do you output omxplayer to /dev/fb1
2016-05-04 19:29:24 <fabled> omxplayer outputs using omx only; the main display is mirror to fb1 + overlay text added with https://github.com/rameplayerorg/rameplayer-utils/tree/master/ramefbcp
2016-05-04 19:29:50 <fabled> it basically takes snapshot of master output every 50-100ms or so
2016-05-04 19:30:40 <fabled> you cannot redirect omx accelerated video directly to fb1 in any way.
2016-05-04 19:30:49 <fabled> that's the only practical way to do it
2016-05-04 19:32:29 <ScrumpyJack> perhaps there is another way
2016-05-04 19:32:49 <ScrumpyJack> mplayer, or X+vlc or something
2016-05-04 19:35:29 <ScrumpyJack> ah, i forgot what replaced mplayer in AL
2016-05-04 19:35:54 <scadu> mpv?
2016-05-04 19:38:13 <fabled> ScrumpyJack, yes, mpv/vlc works, but then it uses the main cpu to decode/render. and depending on the thing it may or may not work.
2016-05-04 19:44:14 <ScrumpyJack> it looks like mpv need sdl2 to be built with directfb support
2016-05-05 07:28:48 <fcolista> who's around here that is root@dev.alpinelinux.org ?
2016-05-05 07:29:56 <fabled> fcolista, yes?
2016-05-05 07:30:16 <fcolista> fabled, can you please delete Net_Socket from distfiles?
2016-05-05 07:30:29 <fcolista> look at #alpine-commits
2016-05-05 07:30:47 <fabled> fcolista, done
2016-05-05 07:30:54 <fcolista> thx fabled
2016-05-05 07:44:51 <ncopa> any idea what to do with php7?
2016-05-05 07:45:05 <ncopa> we rename php -> php5 and upgrade php to 7.x?
2016-05-05 07:45:50 <rnalrd> ncopa, what's the EoL for php5?
2016-05-05 07:46:19 <ncopa> dunno
2016-05-05 07:46:55 <rnalrd> http://php.net/eol.php
2016-05-05 07:48:25 <ncopa> http://php.net/supported-versions.php
2016-05-05 07:49:01 <ncopa> 5.6 is EOL 31 Dec 2018
2016-05-05 07:49:07 <ncopa> we can keep it for v3.4
2016-05-05 07:49:13 <ncopa> i think we need to keep it
2016-05-05 07:49:21 <rnalrd> yes
2016-05-05 07:50:08 <ncopa> but how do we do it?
2016-05-05 07:50:16 <ncopa> currently we have only added php7
2016-05-05 07:50:24 <ncopa> and have php7-*
2016-05-05 07:50:26 <ncopa> i dont like that
2016-05-05 07:50:32 <ncopa> well
2016-05-05 07:50:46 <ncopa> it means that if you dont care what version and just do : apk add php
2016-05-05 07:50:48 <ncopa> you'll get the older
2016-05-05 07:50:53 <ncopa> that is what i dont like
2016-05-05 07:54:11 <fcolista> ncopa, this would affect also the way we call python packages then.
2016-05-05 07:54:34 <ncopa> fabled: what happens if both `php5` and `php7` has `provides=php`, and you `apk add php`?
2016-05-05 07:54:40 <fcolista> py-* should be python3, and then remove py-* to py2
2016-05-05 07:54:55 <ncopa> its more difficult with py
2016-05-05 07:55:06 <ncopa> but yes
2016-05-05 07:55:14 <ncopa> py-* should be renamed py2-*
2016-05-05 07:55:24 <fcolista> yes. I'm just saying that we should have the same approach
2016-05-05 07:55:46 <fcolista> that's going to be much more difficult, yes.
2016-05-05 07:56:18 <ncopa> i think upstream python suggests adding pyhon3
2016-05-05 07:56:32 <ncopa> and not have python = python3
2016-05-05 07:59:41 <fcolista> that's makes sense. When python4 will be shipped we should have no issue with this packages anymore
2016-05-05 08:01:42 <fabled> ncopa, if it's provides=php, they can be installed side by side
2016-05-05 08:01:55 <fabled> if it's provides="php=5" then only one can exist at a time
2016-05-05 08:02:20 <ncopa> which will it pick?
2016-05-05 08:02:44 <ncopa> the one with highest pkgver?
2016-05-05 08:10:37 <suqdiq> http://seclists.org/oss-sec/2016/q2/238 -> oh wow.
2016-05-05 08:11:31 <fabled> ncopa, "apk add php" will error out if there's no other dependency
2016-05-05 08:11:45 <fabled> it says "multiple packages provide virtual package php, please select which to install"
2016-05-05 08:14:41 <ncopa> ok
2016-05-05 08:14:53 <ncopa> sounds good actually
2016-05-05 08:54:51 <leo-unglaub> hey friends, i am in a bit of a hurry with horrible internet so i cannot reseatch it myself
2016-05-05 08:55:03 <leo-unglaub> how can i connecto to a wifi without a passphrase in wpa_supplicant?
2016-05-05 09:02:04 <scadu> wpa_cli, then scan and add_network?
2016-05-05 09:02:42 <ncopa> iirc there was a config option to disable password
2016-05-05 09:04:00 <scadu> set_network 0 key_mgmt NONE
2016-05-05 09:04:11 <scadu> in wpa_cli
2016-05-05 09:08:33 <ncopa> i am moving php modules from /usr/lib/php/modules to /usr/lib/php5/modules
2016-05-05 09:08:59 <ncopa> i wonder if i can move /usr/share/pear to /usr/share/php5/pear
2016-05-05 09:14:40 <suqdiq> sorry for the joke but like, could we move all php related stuff to /dev/null ? :D
2016-05-05 09:15:18 <suqdiq> leo-unglaub: no clue, i use iwconfig for that purpose, dont know if wpa supplicant can connect to non-passworded networks since it will try to negotiate EAP and such all the time
2016-05-05 09:17:40 <leo-unglaub> suqdiq: any idea on how do do that?
2016-05-05 09:18:45 <suqdiq> sure, like iwconfig wlan0 essid 'network-name'
2016-05-05 09:18:56 <suqdiq> replace wlan0 accordingly :p
2016-05-05 09:19:07 <suqdiq> also then depending on nic, you might need to ifconfig wlan0 up
2016-05-05 09:21:13 <przemoc> leo-unglaub: maybe put only ssid="YOUR_SSID" and key_mgmt=NONE in wpa_supplicant's network description?
2016-05-05 09:55:32 <xming> I've got botan/monotone ABUILD if there is any interest I will send PR
2016-05-05 11:01:17 <blueness> i can’t get alpine-desktop working in vbox, has any work been done there before i start re-inventing the wheel?
2016-05-05 11:17:17 <blueness> xf86-video-vmware is not working in virtualbox latest
2016-05-05 11:31:14 <ncopa> blueness: i think its a known problem that it does not work
2016-05-05 11:31:26 <ncopa> would be awesome if you could help figure out why and fix it
2016-05-05 11:33:24 <ncopa> there is a virtualbox-guest-modules-grsec package
2016-05-05 11:33:42 <ncopa> i wonder if we should rebuild that against the linux-virtgrsec kernel
2016-05-05 11:39:38 <suqdiq> yeah X in alpine is a bit broken in many scenarios :D been meaning to get to the bottom of that on the virtualbox field at least
2016-05-05 11:56:15 <ncopa> i'd love to see bug reports on bugs.a.o with explanation of what goes wrong and how to fix it
2016-05-05 12:23:54 <leo-unglaub> hey :)
2016-05-05 12:23:58 <leo-unglaub> thanks for the help before
2016-05-05 12:24:11 <leo-unglaub> i was irc on my phone and researching stuff there sucks soo hard :)
2016-05-05 12:24:13 <leo-unglaub> thanks anyway
2016-05-05 12:52:15 <xming> grsec: denied resource overstep by requesting 135168 for RLIMIT_MEMLOCK against limit 65536 for /usr/bin/mtn[mtn:3005] uid/euid:1001/1001 gid/egid:1001/1001, parent /bin/busybox[sh:3000] uid/euid:1001/1001 gid/egid:1001/1001
2016-05-05 12:52:26 <xming> how do I raise RLIMIT_MEMLOCK?
2016-05-05 12:53:19 <ncopa> ulimit?
2016-05-05 12:53:40 <ncopa> rc_ulimit="...." in /etc/conf.d/<service>
2016-05-05 12:54:10 <ncopa> ulimit -l 128
2016-05-05 12:54:16 <xming> it's from the command line, not a service
2016-05-05 12:56:09 <xming> thanks it's -l
2016-05-05 12:56:53 <xming> doesn't alpine have /etc/security/limits* ?
2016-05-05 12:58:04 <xming> okay I can't raise it as normal user
2016-05-05 12:58:05 <ncopa> i think thats a bash thing?
2016-05-05 13:02:32 <xming> it's from pam, at least according to my gentoo box
2016-05-05 13:05:07 <ncopa> ok
2016-05-05 13:05:17 <ncopa> we dont use pam by default
2016-05-05 13:08:35 <xming> so I need to create a init script to raise the hard limit for users?
2016-05-05 13:12:42 <ncopa> i think you can set it in /etc/rc.conf globally
2016-05-05 13:13:16 <jirutka> how to set it in rc.conf?
2016-05-05 13:14:10 <xming> jirutka: # Pass ulimit parameters
2016-05-05 13:14:15 <xming> ncopa: thanks
2016-05-05 13:14:27 <xming> jirutka: #rc_ulimit="-u 30"
2016-05-05 13:14:39 <jirutka> thanks
2016-05-05 13:59:33 <xming> yes! I've some some more alpine users
2016-05-05 13:59:58 <xming> promoted Alpine a bit on dn42 network
2016-05-05 14:03:14 <jirutka> great! :)
2016-05-05 14:05:42 <xming> anyone here on dn42?
2016-05-05 15:14:25 <poptart> yes xming
2016-05-05 15:15:31 <xming> poptart: hiya
2016-05-05 15:16:34 <poptart> I've been using alpine on it too :P
2016-05-05 15:18:36 <ncopa> rnalrd: would it be ok to move owncloud to community?
2016-05-05 15:18:48 <ncopa> it means that you will need upgrade every 6 months
2016-05-05 15:34:07 <xming> poptart: nice
2016-05-05 15:43:25 <ncopa> i will push the php -> php5 rename now
2016-05-05 15:43:33 <ncopa> its a scary move
2016-05-05 15:43:51 <swswswx> everyone hols onto your pants
2016-05-05 15:43:55 <swswswx> hold
2016-05-05 15:44:04 <ncopa> +1 :)
2016-05-05 17:34:17 <blueness> ncopa: sorry i had to go teach, i succeeded with my stuff to get musl + hardened kernel + hardened toolchian + xorg work on vbox, so there must be something i’m doing right, i’m not sure what yet
2016-05-05 18:00:56 <skarnet> "I must be doing something right, but I have no idea why" :D
2016-05-05 18:01:03 <skarnet> what*
2016-05-05 20:59:38 <jirutka> please someone close http://bugs.alpinelinux.org/issues/5390, it’s already fixed
2016-05-05 20:59:51 <jirutka> #5390
2016-05-05 21:00:23 <jirutka> and also close #5386
2016-05-06 04:36:18 <kvda> Is there any info on how to become a package maintainer for a package that's not already listed?
2016-05-06 04:37:04 <kvda> Gah just found it https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2016-05-06 07:32:55 <suqdiq> kvda: submit it to -testing, should be ok
2016-05-06 07:42:01 <ScrumpyJack> morning
2016-05-06 07:59:48 <scadu> morning.
2016-05-06 10:54:17 <ScrumpyJack> armhf xf86-video-modesetting seems to be empty. is that just me?
2016-05-06 10:55:00 <ScrumpyJack> (inedge)
2016-05-06 12:07:23 <ncopa> ScrumpyJack: it might be that it is shipped with xorg-server
2016-05-06 12:07:56 <ncopa> so it may be its only tehre for backwards compat
2016-05-06 12:29:28 <stwa> hi
2016-05-06 12:29:58 <stwa> can i submit a patch to enable automatic channel selection with hostapd?
2016-05-06 12:42:40 <ncopa> stwa: is it an intrusive patch? is it only config?
2016-05-06 12:42:57 <ncopa> if its not only config change, has upstream accepted the patch?
2016-05-06 12:48:24 <stwa> ncopa: it's only one config line, and no new dependencies
2016-05-06 12:48:45 <ncopa> go for it
2016-05-06 12:49:31 <ncopa> paste it here within 10 mins and i'll apply it now
2016-05-06 12:52:47 <stwa> diff -u a/APKBUILD b/APKBUILD
2016-05-06 12:52:47 <stwa> --- a/APKBUILD	2016-05-05 12:42:18.764995768 +0200
2016-05-06 12:52:47 <stwa> +++ b/APKBUILD	2016-05-06 14:51:46.918261839 +0200
2016-05-06 12:52:47 <stwa> @@ -51,6 +51,7 @@
2016-05-06 12:52:47 <stwa>  		-e '/^#CONFIG_IEEE80211AC=y/s/^#//' \
2016-05-06 12:52:50 <stwa>  		-e '/^#CONFIG_FULL_DYNAMIC_VLAN=y/s/^#//' \
2016-05-06 12:52:52 <stwa>  		-e '/^#CONFIG_LIBNL32=y/s/^#//' \
2016-05-06 12:52:54 <stwa> +		-e '/^#CONFIG_ACS=y/s/^#//' \
2016-05-06 12:52:57 <stwa>  		defconfig >> .config
2016-05-06 12:52:59 <stwa>  	echo "CC ?= ${CC:-gcc}" >> .config
2016-05-06 12:53:02 <stwa>  	echo "CFLAGS += -I/usr/include/libnl3" >> .config
2016-05-06 12:54:02 <stwa> ncopa: just one additional line
2016-05-06 12:55:55 <ncopa> git format-patch --stdout HEAD^ | tpaste
2016-05-06 12:56:05 <ncopa> first: apk add tpaste
2016-05-06 12:56:28 <ncopa> then i get proper commit with commit message and you as author
2016-05-06 12:58:30 <stwa> ok, sorry ...
2016-05-06 12:58:31 <stwa> http://tpaste.us/24bl
2016-05-06 12:59:03 <ScrumpyJack> ncopa: could you slip in patches 1092 to 1096?
2016-05-06 12:59:39 <ncopa> %1092 %1096
2016-05-06 13:01:15 <ncopa> stwa: thanks. i pubmped pkgrel
2016-05-06 13:01:21 <ScrumpyJack> sorry, i meant 1092 to 1096 (including 1093, 1094, and 1095 :)
2016-05-06 13:01:26 <stwa> ncopa: great, thank you!
2016-05-06 13:01:48 <ScrumpyJack> omg, sorry
2016-05-06 13:01:51 <ncopa> ScrumpyJack: which patches?
2016-05-06 13:01:59 <ScrumpyJack> 1902 to 1906
2016-05-06 13:02:11 <ncopa> %1902 %1906
2016-05-06 13:02:31 <ncopa> the avr stuff
2016-05-06 13:02:31 <ScrumpyJack> %1903 %1904 %1905
2016-05-06 13:02:35 <ncopa> hum
2016-05-06 13:02:54 <ScrumpyJack> it's the AVR programming toolchain
2016-05-06 13:03:01 <ncopa> i would like to take a closer look at it
2016-05-06 13:03:08 <ncopa> i dont have anything against it
2016-05-06 13:03:10 <ScrumpyJack> ok np
2016-05-06 13:03:26 <ncopa> fabled: do you have time to take a quick look at those? ^^^
2016-05-06 13:03:34 <ncopa> and apply if they look ok
2016-05-06 13:03:43 <stwa> ncopa: yes, especially the gcc-avr and binutils-avr should be reviewed
2016-05-06 13:04:51 <ncopa> generally
2016-05-06 13:05:11 <ncopa> at this point i ignore testing/* unless its moving to main or community
2016-05-06 13:05:21 <ncopa> i want wrap up the v3.4 release
2016-05-06 13:05:42 <ncopa> the -avr might be nice to sneak in, but i am afraid it will consume time and delay the release
2016-05-06 13:06:18 <ncopa> i want the dovecot plugin fixes though
2016-05-06 13:12:36 <stwa> looking forward to v3.4! :)
2016-05-06 13:13:38 <ncopa> i'm looking forward to get it done too :)
2016-05-06 13:57:01 <ncopa> leaving now for GBR
2016-05-06 13:57:11 <ncopa> if there will be a docker meetup in london next week (wednesday?) then I'll be there
2016-05-06 13:57:13 <ncopa> see u
2016-05-06 14:18:00 <xming> o/
2016-05-06 17:19:56 <scadu> Is there any reason why my latest patch is not visible in patchwork?
2016-05-06 17:38:43 <Shogun> Hi all!
2016-05-06 17:39:09 <Shogun> Can someone help me on a little custom APKBUILD issue?
2016-05-06 17:39:55 <Shogun> Specifically, I downloaded the APKBUILD for nodejs, bump version to 6.0 and try to build it. I got this error:
2016-05-06 17:39:58 <Shogun> >>> nodejs-dev*: Stripping binaries scanelf: mmap on './usr/lib' of 4096 bytes failed :(
2016-05-06 20:57:27 <jirutka> the next release of GitHub Linguist will recognize APKBUILD and OpenRC runscripts :)
2016-05-06 20:57:27 <jirutka> https://github.com/github/linguist/pull/2934
2016-05-06 20:57:27 <jirutka> https://github.com/github/linguist/pull/2935
2016-05-06 20:57:51 <jirutka> this means that they will be correctly categorized and syntax highlighted
2016-05-06 21:28:08 <vkris> I am not sure, but seems busybox-static->crond looks for /bin/sh, is it ok ?
2016-05-06 21:29:43 <vkris> is it a good idea to start http server using busybox-static->inetd, tried couple of times , did not work
2016-05-07 06:46:10 <Shogun> Hi all.
2016-05-07 06:46:21 <Shogun> Is there anyone familiar with scanelf?
2016-05-07 06:51:45 <Shogun> Or even someone familiar with package building.
2016-05-07 06:53:02 <ammunta> what's wrong
2016-05-07 06:55:07 <Shogun> I was trying to build Node 6.0, so I downloaded the APKBUILD of the current official one (5.11.0), bumped the version and try to build it. Compilation and installation go fine, but creating packages doesn't. It return errors like this on all the three (nodejs, nodejs-dev, nodejs-doc) (sub)packages:
2016-05-07 06:55:08 <Shogun> >>> nodejs-dev*: Stripping binaries scanelf: mmap on './usr/lib' of 4096 bytes failed :(
2016-05-07 06:56:44 <ammunta> ... sigh the indentation got messed in that apkbuild, didn't it
2016-05-07 06:57:27 <Shogun> O.o What do you mean?
2016-05-07 07:07:44 <ammunta> not related, I was just trying to look at it
2016-05-07 07:10:42 <Shogun> Oh, I see. The thing is that it seems like it detects $pkgdir/usr/lib as file. And I don't know why
2016-05-07 07:12:11 <Shogun> FYI: I tried fixing the indentation. No luck. :)
2016-05-07 07:16:08 <ammunta> well of course, the indentation doesn't do anything
2016-05-07 07:17:34 <Shogun> I'm so desperate I'd believe to anything XD
2016-05-07 07:46:11 <Shogun> The weird thing is I also get the reverse error. tar cannot create an archive because it thinks .PKGINFO is a directory rather than a file. Bah.
2016-05-07 08:09:33 <Shogun> Nevermind, it's happening even with the regular package (5.11). There must be something wrong with my system. Let me try with a clean OS install.
2016-05-07 11:18:28 <Mo0O> hello
2016-05-07 11:19:24 <Mo0O> do you know how can I fix this cc1plus error: unrecognized command line option '-Wno-unused-local-typedef' [-Werror]
2016-05-07 11:19:26 <Mo0O> ?
2016-05-07 16:41:02 <jirutka> is there anyone how has access to Alpine servers? TLS certificates on alpinelinux.org and bugs.alpinelinux.org has expired! #5545
2016-05-08 10:33:15 <Tsutsukakushi> still no new cert?
2016-05-08 10:33:34 <jirutka> still not
2016-05-08 10:33:44 <Tsutsukakushi> :/
2016-05-08 10:34:39 <jirutka> I’m trying to find email to Nathan Angelacos, he’s apperantly ignoring bug reports on bugs.a.o >_<
2016-05-08 10:34:47 <Tsutsukakushi> hm
2016-05-08 10:34:58 <Tsutsukakushi> Natanael Copa <ncopa@alpinelinux.org>
2016-05-08 10:35:00 <Tsutsukakushi> ?
2016-05-08 10:35:01 <Tsutsukakushi> this?
2016-05-08 10:35:15 <Tsutsukakushi> no
2016-05-08 10:35:17 <Tsutsukakushi> :/
2016-05-08 10:35:19 <jirutka> nope
2016-05-08 10:36:36 <jirutka> ncopa is currently on some conference
2016-05-08 10:36:43 <jirutka> Nathan Angelacos is responsible for this
2016-05-08 10:37:05 <Tsutsukakushi> ah
2016-05-08 10:37:05 <Tsutsukakushi> ok
2016-05-08 10:38:23 <jirutka> Nathan Angelacos is nangel
2016-05-08 10:38:41 <Tsutsukakushi> hmm
2016-05-08 10:38:54 <jirutka> still can’t find email, damn
2016-05-08 10:39:21 <Tsutsukakushi> nangel@users.sourceforge.net
2016-05-08 10:39:21 <Tsutsukakushi> there
2016-05-08 10:39:23 <Tsutsukakushi> from git
2016-05-08 10:39:33 <Tsutsukakushi> nangel/haserl
2016-05-08 10:39:42 <jirutka> I don’t think this is a real email…
2016-05-08 10:39:43 <Tsutsukakushi> at git.alpinelinux.org
2016-05-08 10:39:53 <jirutka> srsly, who would use email at sourceforge.net?
2016-05-08 10:39:59 <ammunta> last time didn't skarnet get nangel's attention by throwing a tantrum in the mail logs :))
2016-05-08 10:40:00 <Tsutsukakushi> it probably forwards somewhere
2016-05-08 10:40:24 <jirutka> hmm… skarnet, are you here? we need your help :P
2016-05-08 10:40:52 <jirutka> ha, nangel@alpinelinux.org
2016-05-08 10:40:58 <Tsutsukakushi> you should try sending a mail to that address
2016-05-08 10:40:58 <jirutka> uh, obviously…
2016-05-08 10:41:05 <Tsutsukakushi> it probably forwards to his mail
2016-05-08 10:41:11 <Tsutsukakushi> lol
2016-05-08 10:41:11 <Tsutsukakushi> ok
2016-05-08 10:42:23 <jirutka> okay, I’ll try to do like 5 self-censorships loops to not offend him as irresponsible admin… :)
2016-05-08 10:42:51 <scadu> It's Sunday. I doubt anyone will answer :f
2016-05-08 10:43:02 <jirutka> it’s expired since friday
2016-05-08 10:43:25 <jirutka> and StartSSL sends dozen of emails before cert actually expire
2016-05-08 10:44:11 <Tsutsukakushi> :/
2016-05-08 10:45:00 <scadu> maybe nangel lost the bear fight :<
2016-05-08 10:45:10 <jirutka> bear fight? o.O
2016-05-08 10:52:33 <ammunta> bear fight
2016-05-08 10:52:54 <jirutka> okay, e-mail sent
2016-05-08 10:53:12 <jirutka> what’s bear fight? you don’t mean actual fight with a bear, do you?
2016-05-08 10:54:53 <scadu> I was joking. Have no idea where nangel is.
2016-05-08 10:55:08 <jirutka> :)
2016-05-08 10:55:47 <skarnet> bear fighting does involve an actual fight with a bear, obviously
2016-05-08 10:55:51 <skarnet> what do you think it's about
2016-05-08 10:56:14 <Tsutsukakushi> skarnet: it could be a metaphorical bear fight
2016-05-08 10:56:26 <jirutka> ^ that
2016-05-08 10:56:30 <skarnet> real bear fight or gtfo
2016-05-08 10:56:49 <Tsutsukakushi> you could call some gay sex a "bear fight"
2016-05-08 10:56:57 <jirutka> LOL
2016-05-08 10:57:20 <jirutka> since we don’t have many wild bears in Czech Republic, I implicitly don’t expect to be a real bear fight :)
2016-05-08 10:57:34 <skarnet> with "bear swordfight" you'd have a case
2016-05-08 10:57:55 <skarnet> but in a void, a bear fight isn't sexual at all
2016-05-08 10:58:11 <skarnet> (or you have very specific kinks)
2016-05-08 10:58:44 <scadu> my mind has a kernel panic.
2016-05-08 10:58:50 <jirutka> XD
2016-05-08 10:58:59 <jirutka> try to restart it :P
2016-05-08 10:59:06 <jirutka> like, hard restart…
2016-05-08 10:59:38 <skarnet> anyway, yes, nangel was responsible for mail server management, I needed something done with the mail server, I threw a tantrum here and somehow that caught nangel's attention, but it could simply have been that someone called him
2016-05-08 11:00:04 <skarnet> like, nangel, please come to irc and silence this guy
2016-05-08 11:00:38 <jirutka> well, maybe we could try to use the some tactic? :)
2016-05-08 11:00:45 <skarnet> I can't guarantee it would work for certificate renewal, but you can try :)
2016-05-08 11:31:05 <jirutka> WHAT THE HELL?!
2016-05-08 11:31:06 <jirutka> <mailto:nangel@alpinelinux.org>: host http://mail.alpinelinux.org/[74.117.189.114] said: 452
2016-05-08 11:31:06 <jirutka>    4.3.1 Insufficient system storage (in reply to MAIL FROM command)
2016-05-08 11:32:19 <jirutka> is he f*cking kidding me?! not just that he failed to renew TLS certs before expiration, so he must ignored dozen of mails from StartSSL, but also run out of space on mail server?!
2016-05-08 11:33:32 <scadu> wee
2016-05-08 11:38:03 <Tsutsukakushi> what do the alpine mailing lists have in their List-id ?
2016-05-08 11:38:15 <jirutka> dunno?
2016-05-08 11:38:57 <Tsutsukakushi> i wish listinfo pages started including that info  :/
2016-05-08 11:39:12 <Tsutsukakushi> so i wouldn't have to wait to receive mail from lists before i could make a filter rule for them
2016-05-08 11:39:17 <Tsutsukakushi> it's kind of annoyinh
2016-05-08 11:39:21 <Tsutsukakushi> annoying
2016-05-08 11:59:22 <skarnet> insufficient system storage? I haven't seen that error message since, like, 1996
2016-05-08 11:59:33 <jirutka> exactly…
2016-05-08 12:00:16 <skarnet> is it possible to contribute financially to Alpine? I'd like to contribute a hard disk
2016-05-08 12:00:24 <jirutka> :)
2016-05-08 12:00:50 <jirutka> as ncopa told me recently, he has enough HW resources
2016-05-08 12:01:51 <skarnet> it's not showing
2016-05-08 12:02:00 <jirutka> this isn’t problem in a HW
2016-05-08 12:02:03 <jirutka> but sysadmin
2016-05-08 12:02:03 <skarnet> quota?
2016-05-08 12:02:13 <skarnet> who the heck still uses quotas?
2016-05-08 12:02:34 <Tsutsukakushi> skarnet: shared hosts
2016-05-08 12:02:37 <Tsutsukakushi> probably do
2016-05-08 12:02:44 <jirutka> there were some info about PayPal, but I can’t find it now
2016-05-08 12:02:50 <skarnet> mail.alpinelinux.org
2016-05-08 12:02:56 <skarnet> does it look like a shared host?
2016-05-08 12:02:59 <jirutka> Alpine doesn’t run on a shared host
2016-05-08 12:03:05 <skarnet> does it look like it should run on a shared host?
2016-05-08 12:03:15 <skarnet> CHEESUS
2016-05-08 12:03:15 <Tsutsukakushi> yeah
2016-05-08 12:03:16 <Tsutsukakushi> but
2016-05-08 12:03:25 <Tsutsukakushi> i was just answering the question
2016-05-08 12:03:31 <Tsutsukakushi> shared hosts use quotas
2016-05-08 12:03:35 <skarnet> it was a rhetorical question
2016-05-08 12:03:46 <Tsutsukakushi> geez
2016-05-08 12:03:49 <skarnet> like "does it look like a shared host?"
2016-05-08 12:03:50 <Tsutsukakushi> you could say that then
2016-05-08 12:03:54 <skarnet> it's a rhetorical question too
2016-05-08 12:03:58 <jirutka> :)
2016-05-08 12:04:01 <Tsutsukakushi> it doesn't really work over text
2016-05-08 12:08:23 <skarnet> https://goo.gl/rznAWY  classical example of aggressive questioning: there's just no good answer
2016-05-08 12:12:46 <jirutka> anyway, does anyone know any ncopa’s email not hosted on alpine.org?
2016-05-08 12:35:52 <jirutka> skarnet: do you what is the attitude towards setuid in Alpine?
2016-05-08 12:38:01 <coredumb> ncopa@docker.com ?
2016-05-08 12:38:24 <coredumb> just throwing ideas ^^
2016-05-08 12:38:36 <Mo0O> lol
2016-05-08 12:38:41 <jirutka> close, natanael.copa@docker.com, acccording to some git commits :)
2016-05-08 12:38:47 <jirutka> thanks
2016-05-08 12:38:50 <coredumb> ah see!
2016-05-08 12:41:18 <pluesch> there is a wiki entry about donations
2016-05-08 12:41:56 <pluesch> https://wiki.alpinelinux.org/wiki/Alpine_Linux:Developers
2016-05-08 12:42:52 <leo-unglaub> hey friends :)
2016-05-08 12:51:13 <jirutka> aha, there’s `options="suid"`!
2016-05-08 13:02:07 <chris|> imo, there are far too many packages that require suid, so any new package should at least attempt to no rely on it
2016-05-08 13:02:29 <jirutka> qemu-bridge-helper needs suid
2016-05-08 13:02:55 <jirutka> to be able to start qemu VM as non-privileged user
2016-05-08 13:03:06 <jirutka> is it sufficient to just add options="suid" to APKBUILD?
2016-05-08 13:46:23 <jirutka> hmm, even debian refused setuid on qemu-bridge-helper, so I’ll try to find better solution
2016-05-08 13:53:33 <jirutka> skarnet: I wonder if https://github.com/qemu/qemu/blob/master/qemu-bridge-helper.c can be simply modified to not require suid, but instead use sudo to gain root privileges, so instead of setting suid on this binary, we would add sudo rule for qemu user; what do you think?
2016-05-08 13:58:05 <skarnet> I can't give you a good answer without examining the thing in detail, and honestly, it's Sunday, it's sunny, and I don't want to. :P
2016-05-08 13:59:03 <skarnet> The best solution also depends on general Alpine policy, and I'm not the person to ask about that.
2016-05-08 14:01:23 <skarnet> Ew. Just a glance at the qemu-bridge-helper.c code and I understand why debian doesn't want it to be setuid. But a sudo rule wouldn't be much better - either way you're running that thing as root.
2016-05-08 14:03:30 <Mo0O> aync alpine mirror check using golang: http://sprunge.us/HZOg?go
2016-05-08 14:08:47 <chris|> if the admin wants to give users permission to run vms, he can set suid flag explicitly (or grant sudo), I really see no reason why the package should default that
2016-05-08 14:13:39 <jirutka> because this is the whole point of qemu-bridge-helper
2016-05-08 14:14:27 <jirutka> we should at least add some warn message, because currently it doesn’t work and without deeper examine user can’t know why
2016-05-08 14:16:54 <jirutka> hm, you’re right that sudo doesn’t change anything, I’ve forgotten that we can just set 04750 and group qemu; if admin wants to allow some user to run VM, then he will explicitly add him to the qemu group; exactly the same as for kvm to use paravirtualization (e.g. /dev/kvm device)
2016-05-08 14:18:18 <jirutka> btw actually I don’t want to allow users to run VMs, I just don’t want to run VMs as root, but as some unprivileged user
2016-05-08 14:31:25 <skarnet> then 4750 is the way to go
2016-05-08 14:42:34 <chris|> sounds reasonable
2016-05-08 14:42:54 <jirutka> okay, I’m gonna send a patch :)
2016-05-08 15:02:18 <ammunta> 06:57 < skarnet> but in a void, a bear fight isn't sexual at all
2016-05-08 15:02:19 <ammunta> 06:58 < skarnet> (or you have very specific kinks)
2016-05-08 15:02:21 <ammunta> 06:57 < skarnet> but in a void, a bear fight isn't sexual at all
2016-05-08 15:02:21 <ammunta> 06:58 < skarnet> (or you have very specific kinks)
2016-05-08 15:18:54 <skarnet> twice for emphasis
2016-05-08 15:46:03 <Tsutsukakushi> in the void it wouldn't be much of a fight
2016-05-08 15:46:13 <Tsutsukakushi> both you and the bear would suffocate right away
2016-05-08 15:55:18 <leo-unglaub> i did not read the entire backlog, but problem do people have all of the sudden with suid?
2016-05-08 15:55:44 <leo-unglaub> if you suid it or if you sudo it or if you su root and then execute it have the same effect in the security context
2016-05-08 15:55:57 <leo-unglaub> shit code is always dangerous, no matter how you execute it
2016-05-08 16:02:03 <chris|> leo-unglaub: I think the main problem is that package maintainers (not necessarily alpine related) tend to use suid or sudo as a tool to increase convenience at the cost of security, and I think that is the wrong way to go. It is not the job of the maintainer to grant permissiosns to anything, that's the prerogative of the person actually responsible for system configuration
2016-05-08 16:03:17 <leo-unglaub> its the job of the software developer to develop code that is secure and clean. if a program needs suid in order to work nice then use it.
2016-05-08 16:03:33 <skarnet> unfortunately that's not how it works :/
2016-05-08 16:04:09 <leo-unglaub> in my opinion in the last couple of months some hipster idiots started to get rid of suid withoutunderstanding the problem ... jsut because its cool now everyone wants to get rid if it
2016-05-08 16:04:09 <skarnet> in an ideal world, everything could run as root and it would all be unicorns, rainbows and butterflies
2016-05-08 16:04:16 <leo-unglaub> see ping and traceroute in alpine
2016-05-08 16:04:36 <skarnet> getting rid of suid is actually a good idea, no matter whether the cool kids are doing it or not
2016-05-08 16:04:44 <leo-unglaub> using sudo just to ping a host and see if a network is reachable is very stupid in my mind
2016-05-08 16:05:33 <skarnet> uh, yeah, that's not how "removing the s bit on ping" is supposed to work.
2016-05-08 16:15:05 <jirutka> > using sudo just to ping a host and see if a network is reachable is very stupid in my mind
2016-05-08 16:15:05 <jirutka> aaah, that’s the reason why on Gentoo I can’t ping as normal user anymore!
2016-05-08 16:15:47 <jirutka> I didn’t know that ping needs root privileges… wait a moment… why?
2016-05-08 16:21:07 <skarnet> ICMP sockets.
2016-05-08 16:21:16 <skarnet> normal users can only get TCP or UDP sockets.
2016-05-08 16:21:59 <skarnet> to talk raw IP, or ICMP, you need root. Which makes sense if you don't want your normal users to DoS your network in the blink of an eye.
2016-05-08 16:22:40 <jirutka> you can’t DoS network using TCP or UDP packets…?
2016-05-08 16:30:43 <leo-unglaub> in general it makes sence to limit icmp to root
2016-05-08 16:30:52 <leo-unglaub> but ping is very important for other users
2016-05-08 16:30:58 <leo-unglaub> so the suid is the perfect solution
2016-05-08 16:36:59 <chris|> leo-unglaub: suid on ping is obsolete since 2010 or something when ICMP DGRAM protocol was introduced. The only reason any ping implementation still needs root is because they are too lazy to upgrade
2016-05-08 16:39:36 <chris|> and thay should also make the point against developers being responsible for providing "secure and clean" code since even ping authors seem to strugle
2016-05-08 16:44:11 <leo-unglaub> not really
2016-05-08 19:33:09 <coredumb> isn't all that been solved with capabilities 2 years or so ago ?
2016-05-08 19:33:43 <skarnet> theoretically, yes
2016-05-08 19:44:16 <coredumb> so what's the debate about ?
2016-05-08 19:44:25 <coredumb> we still need suid bins on alpine ?
2016-05-08 19:44:35 <jirutka> yes
2016-05-08 19:44:41 <chris|> the principle of the thing
2016-05-08 19:46:01 <coredumb> how hard is that to change ?
2016-05-08 19:46:21 <chris|> I would make the same argument for capabilities, they should be granted by admins, not by package maintainers
2016-05-08 19:46:30 <chris|> unless absolutely necessary
2016-05-08 19:46:45 <coredumb> ok so
2016-05-08 19:46:56 <coredumb> is it better to have suid or capability ?
2016-05-08 19:47:13 <chris|> capabilities
2016-05-08 19:47:16 <skarnet> capabilities are finer-grained
2016-05-08 19:47:19 <skarnet> but less programs use them
2016-05-08 19:48:24 <skarnet> (and arguably they're right, because capabilities are not portable :D)
2016-05-08 19:48:40 <coredumb> so we just need to add cap_net_raw+ep to /bin/busybox and ping is fixed
2016-05-08 19:49:04 <coredumb> what's the blockage against it ?
2016-05-08 19:49:07 <skarnet> not exactly
2016-05-08 19:49:07 <chris|> that and it is more complicated to maintain them from a packagers perspective
2016-05-08 19:49:24 <chris|> ping is already fixed if I am not mistaken
2016-05-08 19:49:27 <skarnet> if you do that, it means /bin/sh also has the capability to make raw sockets
2016-05-08 19:49:36 <coredumb> skarnet: exactly
2016-05-08 19:49:38 <skarnet> and you just killed your LAN
2016-05-08 19:51:02 <ammunta> are caps inherited on exec
2016-05-08 19:51:37 <skarnet> I very much hope they are - if they're not, it's unusable
2016-05-08 19:52:41 <ammunta> it would be stupid, but not surprising tbh
2016-05-08 19:54:45 <coredumb> damn we discussed that bbsuid thing last time
2016-05-08 20:01:38 <chris|> if I remember correctly, capabilities are not inherited (or rather: passed down) by default unless both the calling and called process are set allow it throuth the inheriable set
2016-05-08 20:02:31 <coredumb> ok so there's traceroute
2016-05-08 20:02:42 <coredumb> I think it needs the same fix as ping right ?
2016-05-08 20:04:44 <skarnet> chris|: how could a called process set anything before being execve()'d ?
2016-05-08 20:07:20 <chris|> skarnet: obviously it can't, caps on that process need to be set with setcap, but beyond that, I'd probably have to go read up on caps to provide a complete answer
2016-05-08 20:08:51 <skarnet> eh, don't bother, I'll go read up on caps myself the day I'm interested enough in them ;)
2016-05-08 20:09:38 <chris|> yeah, same here ;)
2016-05-08 20:10:25 <jirutka> skarnet: just wonder, what do you have against Linux caps?
2016-05-08 20:10:40 <skarnet> the answer is contained in your question :)
2016-05-08 20:10:54 <jirutka> Linux?
2016-05-08 20:11:03 <skarnet> yes.
2016-05-08 20:11:06 <jirutka> that’s the only reason?
2016-05-08 20:11:10 <skarnet> Capabilities are not POSIX, not portable.
2016-05-08 20:11:35 <jirutka> POSIX is very old, so we should stick to it like forever?
2016-05-08 20:12:16 <skarnet> no, things change, slowly, and there are a few extensions used on several systems and that can reasonably be used in a portable program
2016-05-08 20:12:29 <skarnet> but capabilities are still a Linux-only mutant feature
2016-05-08 20:13:14 <jirutka> is there any better (systematic) solution?
2016-05-08 20:14:10 <skarnet> standard Unix permissions already go a long way *if they're used correctly*, which unfortunately is very rarely the case
2016-05-08 20:14:36 <jirutka> how would you solve e.g. problem with ping with just Unix permissions?
2016-05-08 20:15:49 <skarnet> keeping ping root, and maybe rewrite it if security really is a problem. ping isn't rocket science, you should be able to run that code as root without getting pwned.
2016-05-08 20:16:24 <skarnet> A security hole in ping means the writer was a group of one to fifteen monkeys, approximately
2016-05-08 20:16:27 <jirutka> so every user should be root? o.O
2016-05-08 20:16:56 <jirutka> isn’t that stupid that non-privileged users cannot use such a basic tool like ping?
2016-05-08 20:17:06 <skarnet> no, suid it if you want your users to have access to it
2016-05-08 20:17:10 <jirutka> aha
2016-05-08 20:17:58 <skarnet> that's my point: it's a basic tool, if there's an issue in suiding it, then it needs to be audited and maybe rewritten
2016-05-08 20:20:48 <skarnet> depending on the icmp socket implementation, it should also be possible to drop privileges after binding the socket, and dramatically reduce the amount of trusted code
2016-05-08 20:21:34 <chris|> so s6-ping? :)
2016-05-08 20:22:30 <skarnet> I'd like to avoid that if possible :P
2016-05-08 20:58:52 <chris|> who wouldn't ;)
2016-05-08 20:59:19 <jirutka> Lennart? systemd-ping…
2016-05-08 20:59:46 <jirutka> I bet that it’s on his fucking list of things he wants to mess up
2016-05-08 21:01:21 <chris|> yeah, but he is probably basing that on kdbus, so...
2016-05-08 21:01:50 <skarnet> You don't understand. It's important that other applications can get access to the information exported by ping, and dbus is the ideal framework for that.
2016-05-08 21:02:54 <skarnet> Also, for security, it's essential that ping has its own cgroup, and since centralized cgroups are mandatory, only systemd can decently control when ping is executed.
2016-05-08 21:03:07 <jirutka> it’ll be also hard-depend on Pulse Audio, because you know, it’s everyone needs to _play_ ping.
2016-05-08 21:03:40 <skarnet> Auditive network feedback *is* a feature that has been asked for by users.
2016-05-08 21:04:04 <skarnet> I really don't understand the amount of hate. The Linux community is really a bunch of childish ragers.
2016-05-08 21:04:16 <jirutka> XD
2016-05-08 21:05:04 <skarnet> I should work on that impersonation for next Halloween...
2016-05-08 21:05:29 <chris|> obviously, children take after their father
2016-05-08 21:18:44 <jirutka> abuild question: I’m trying to write a package for R; it installs libR.so to /usr/lib/R/lib and when abuild traces dependencies, it can’t find it… how to deal with it? export LD_LIBRARY_PATH in the APKBUILD?
2016-05-09 02:45:06 <iShyion> hi
2016-05-09 02:45:10 <iShyion> move your asses
2016-05-09 02:45:16 <iShyion> dump of idiots
2016-05-09 02:45:32 <iShyion> asdasdasdasd
2016-05-09 08:16:50 <leo-unglaub> hey :)
2016-05-09 08:25:52 <ScrumpyJack> hey
2016-05-09 08:25:58 <ScrumpyJack> clandmeter: you about?
2016-05-09 08:53:27 <ncopa> hi
2016-05-09 08:53:39 <ncopa> looks like bugs.a.o cert expired
2016-05-09 08:54:14 <jirutka> hi
2016-05-09 08:54:30 <jirutka> bugs.a.o has been _finally_ fixed, but https://alpinelinux.org/ is still expired
2016-05-09 08:56:45 <jirutka> what, Scrumpy Jack has closed the issue despite that it’s not actually fixed #5546
2016-05-09 08:59:12 <Tsutsukakushi> ncopa: all your certs did
2016-05-09 08:59:52 <ncopa>  /o\
2016-05-09 09:00:01 <Tsutsukakushi> ncopa: you should do let's encrypt
2016-05-09 09:00:14 <jirutka> ^ agree
2016-05-09 09:00:36 <ScrumpyJack> jirutka: that was a mistake
2016-05-09 09:01:00 <jirutka> I see, you’ve spammed me with dozen of emails with status change, yet the bug is still not fixed
2016-05-09 09:01:03 <ScrumpyJack> I closed  #5546 as duplicate of  #5545 and redmine closed both
2016-05-09 09:01:26 <jirutka> I must go
2016-05-09 09:01:34 <ScrumpyJack> I'm now more familiar with redmine! :)
2016-05-09 09:24:20 <ncopa> it seems that someone thinks that you can hack redline by posting an issue as html?
2016-05-09 09:24:22 <ncopa> https://bugs.alpinelinux.org/issues/5549
2016-05-09 10:01:28 <clandmeter> redmine certs updated
2016-05-09 10:17:36 <ScrumpyJack> .
2016-05-09 10:49:27 <fcolista> ncopa, after renaming php to php5
2016-05-09 10:49:37 <fcolista> how to fix broken dependencies?
2016-05-09 10:49:37 <fcolista> ERROR: unsatisfiable constraints:
2016-05-09 10:49:38 <fcolista>   php-5.6.19-r0:
2016-05-09 10:49:38 <fcolista>     conflicts: php5-5.6.21-r1
2016-05-09 10:49:38 <fcolista>     satisfies: php-ctype-5.6.19-r0[php] php-curl-5.6.19-r0[php] php-dom-5.6.19-r0[php] php-sockets-5.6.19-r0[php] php-pdo-5.6.19-r0[php] php-pgsql-5.6.19-r0[php] php-sqlite3-5.6.19-r0[php]
2016-05-09 10:49:39 <fcolista>                php-pdo_pgsql-5.6.19-r0[php] php-bcmath-5.6.19-r0[php] php-gd-5.6.19-r0[php] zabbix-webif-3.0.1-r0[php] php-iconv-5.6.19-r0[php] php-gettext-5.6.19-r0[php] php-xmlreader-5.6.19-r0[php]
2016-05-09 10:49:42 <fcolista>   php5-5.6.21-r1:
2016-05-09 10:49:44 <fcolista>     conflicts: php-5.6.19-r0[php]
2016-05-09 10:49:46 <fcolista>     satisfies: world[php5] zabbix-webif-3.0.1-r0[php] php-ctype-5.6.19-r0[php] php-curl-5.6.19-r0[php] php-dom-5.6.19-r0[php] php-sockets-5.6.19-r0[php] php-pdo-5.6.19-r0[php] php-pgsql-5.6.19-r0[php]
2016-05-09 10:49:49 <fcolista>                php-sqlite3-5.6.19-r0[php] php-pdo_pgsql-5.6.19-r0[php] php-bcmath-5.6.19-r0[php] php-gd-5.6.19-r0[php] php-iconv-5.6.19-r0[php] php-gettext-5.6.19-r0[php] php-xmlreader-5.6.19-r0[php]
2016-05-09 10:52:20 <fcolista> for i in $(apk info | grep php); do apk add  ${i/php/php5}; done
2016-05-09 10:52:21 <fcolista> ?
2016-05-09 10:52:51 <fcolista> ncopa, after renaming php to php5
2016-05-09 10:52:52 <fcolista> for i in $(apk info | grep php); do apk add  ${i/php/php5}; done
2016-05-09 10:53:00 <fcolista> would this ^ solve broken dependencies?
2016-05-09 13:03:30 <jirutka> fcolista: this really isn’t a safe solution!
2016-05-09 13:03:50 <fcolista> no is not
2016-05-09 13:04:01 <fcolista> and doesn't work, also
2016-05-09 13:04:09 <jirutka> ;)
2016-05-09 13:09:21 <clandmeter> ScrumpyJack: you still needed me?
2016-05-09 13:10:03 <jirutka> scrumpyjack: https://alpinelinux.org/ … what do you see?
2016-05-09 13:10:08 <jirutka> I see The certificate is only valid for the following names: bugs.alpinelinux.org, msg.alpinelinux.org, patchwork.alpinelinux.org, www.alpinelinux.org, forum.alpinelinux.org .
2016-05-09 13:10:40 <clandmeter> ah thats my mistake
2016-05-09 13:10:46 <clandmeter> well
2016-05-09 13:10:54 <clandmeter> i thought alpinelinux.org should be included.
2016-05-09 13:12:58 <ScrumpyJack> clandmeter: It's good thanks
2016-05-09 13:15:24 <ScrumpyJack> clandmeter: you put bugs.alpinelinux.org in a Subject CN and as an alt DNS, thinking you had put only alpinelinux.org in the S. CN
2016-05-09 13:19:37 <clandmeter> ScrumpyJack: startssl support up to 5 domains on a single cert.
2016-05-09 13:19:47 <clandmeter> i though the main domain would be auto included.
2016-05-09 13:21:07 <barthalion> why not let's encrypt?
2016-05-09 13:24:04 <ScrumpyJack> will startssl issue a cert for longer than 1 year?
2016-05-09 13:24:42 <clandmeter> barthalion: becausde nobody took tht time to set it up.
2016-05-09 13:24:52 <clandmeter> ScrumpyJack: no, not for free
2016-05-09 13:25:31 <ScrumpyJack> ah yes, 5 domains 1 year, $0.00
2016-05-09 13:26:41 <ScrumpyJack> well, i guess that gives us 1 year to setup LE :)
2016-05-09 13:27:36 <clandmeter> caddyserver is nice, but its missing some features.
2016-05-09 13:32:35 <jirutka> I think that someone spammed https://bugs.alpinelinux.org/issues/5408 – incef.php.fla, incef.php.sh
2016-05-09 13:35:06 <clandmeter> jirutka: i removed them and deleted the user.
2016-05-09 13:42:21 <BastienM> Hello guys,  is there a more elaborate doc on how to build a custom iso (apart the one from the wiki) ?
2016-05-09 13:42:47 <BastienM> currently looking to build a livecd with KVM embedded (ESX like)
2016-05-09 13:43:55 <BastienM> https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image seems outdated (no lbu apk available, nor busybox -p)
2016-05-09 13:52:06 <clandmeter> BastienM: busybox -p 127.0.0.1:80 is a mistake
2016-05-09 13:52:20 <clandmeter> it should be the applet httpd -p ....
2016-05-09 13:53:16 <clandmeter> im not sure what you mean with a livecd with kvm embedded?
2016-05-09 13:55:38 <BastienM> ho ok, gonna note that somewhere. Thanks
2016-05-09 13:58:42 <BastienM> will be back later, thanks again for the hint clandmeter :)
2016-05-09 14:26:52 <Tsutsukakushi> would anyone else be interested in having a linux-libre-grsec kernel in the alpine linux repos?
2016-05-09 14:26:58 <Tsutsukakushi> i could try making that
2016-05-09 14:28:18 <Rhy0lite> Who should I contact to start the process to add more architectures to Alpine Linux?
2016-05-09 14:28:30 <Tsutsukakushi> Rhy0lite: ncopa
2016-05-09 14:28:35 <Tsutsukakushi> i think
2016-05-09 14:28:49 <Tsutsukakushi> Rhy0lite: which archs are you thinking of?
2016-05-09 14:28:57 <Rhy0lite> PPC64 and s390x
2016-05-09 14:29:03 <Rhy0lite> I work for IBM
2016-05-09 14:29:21 <Tsutsukakushi> oh
2016-05-09 14:29:43 <Rhy0lite> musl libc now runs on PPC64 and port has been started for s390x
2016-05-09 14:29:55 <Tsutsukakushi> i don't know if the alpine team has enough
2016-05-09 14:30:05 <Tsutsukakushi> enough extra resources to support much more
2016-05-09 14:30:08 <ncopa> Rhy0lite, hi
2016-05-09 14:30:21 <Rhy0lite> Hi, ncopa!
2016-05-09 14:30:27 <ncopa> Rhy0lite: funny, we actually talked about PPC64 today
2016-05-09 14:30:34 <Rhy0lite> Great!
2016-05-09 14:30:36 <Tsutsukakushi> alpine on talos would be pretty nice package
2016-05-09 14:30:52 <Rhy0lite> Thanks, Tsutsukakushi!
2016-05-09 14:31:08 <Tsutsukakushi> np
2016-05-09 14:31:13 <Rhy0lite> I understand the limited resources
2016-05-09 14:32:03 <Rhy0lite> I want to start the discussion to understand the process and bottlenecks
2016-05-09 14:33:18 <Tsutsukakushi> are you here representing IBM or did you just mention that for some other reason?
2016-05-09 14:33:37 <Rhy0lite> I am representing IBM
2016-05-09 14:33:42 <Tsutsukakushi> kwl
2016-05-09 14:33:52 <ncopa> cool indeed
2016-05-09 14:34:30 <ncopa> so what we need is a builder
2016-05-09 14:34:35 <ncopa> the hw
2016-05-09 14:34:45 <ncopa> i suppose thats not an issue for IBM :)
2016-05-09 14:35:06 <ncopa> then someone with the time to bootstrap alpine
2016-05-09 14:35:19 <Rhy0lite> HW is not always that easy
2016-05-09 14:35:29 <ncopa> ok..
2016-05-09 14:35:42 <Rhy0lite> I assume that you want it hosted for you somewhere
2016-05-09 14:36:03 <Rhy0lite> There are a number of systems for Open Source developers at various locations around the world
2016-05-09 14:36:30 <Rhy0lite> OSUOSL, FIT-Brno, SuperVessel in China, Unicamp in Brazil
2016-05-09 14:36:32 <Tsutsukakushi> (soon with talos you could get a dev machine for ncopa for 3k)
2016-05-09 14:38:26 <Rhy0lite> BTW, Alpine Linux support already is listed as a focus area for the "Open Mainframe Project"
2016-05-09 14:38:27 <Rhy0lite> https://wiki.linuxfoundation.org/openmainframeproject/tsc/projectplanning
2016-05-09 14:38:50 <ncopa> then we�d need cross compiler to cross compile a native toolchain, gcc,musl,busybox
2016-05-09 14:39:09 <ncopa> and from there we need to bootstrap abuild/apk-tools
2016-05-09 14:39:31 <Tsutsukakushi> Rhy0lite: Redirection loop detected (https://wiki.linuxfoundation.org/openmainframeproject/tsc/projectplanning)
2016-05-09 14:39:57 <ncopa> i think the trickiest parts are to build the native toolchain
2016-05-09 14:40:03 <Rhy0lite> The link works for me
2016-05-09 14:40:36 <Tsutsukakushi> keeps redirecting between identity.linuxfoundation.org and wiki.linuxfoundation.org
2016-05-09 14:40:40 <Tsutsukakushi> CAS Authentication wanted!
2016-05-09 14:42:18 <Rhy0lite> ncopa: If someone bootstraps it, what is the process to add the arch to the distro?
2016-05-09 14:43:35 <Rhy0lite> Is Alpine Linux a legal entity?  A foundation?
2016-05-09 14:43:44 <ncopa> we need an official builder, which will listen to mqtt event to pick up git pushes
2016-05-09 14:44:02 <ncopa> no, alpine linux is not (yet) a legal entity
2016-05-09 14:44:02 <Rhy0lite> Okay
2016-05-09 14:44:23 <Rhy0lite> Solely technical issue
2016-05-09 14:46:18 <ncopa> Rhy0lite, would you be able to work on setting up ppc64?
2016-05-09 14:46:46 <Rhy0lite> in terms of hardware or the full bootstrap?
2016-05-09 14:46:51 <ncopa> both
2016-05-09 14:47:16 <Rhy0lite> I probably would not do it all myself
2016-05-09 14:48:25 <Rhy0lite> Let me provide this feedback internally and figure out how my colleagues want to proceed
2016-05-09 14:49:39 <ncopa> i can probably help setting up the build server once the native toolchain is bootstrapped
2016-05-09 14:49:45 <Rhy0lite> I am leading a number of projects utilizing financial bounties to expand the ecosystem on POWER and Z
2016-05-09 14:49:51 <ncopa> bootstrapping apk-tools/abuild
2016-05-09 14:50:20 <Rhy0lite> We may choose to create some bounties to solicit help
2016-05-09 14:50:37 <Rhy0lite> I first wanted to know the requirements
2016-05-09 15:11:31 <Tsutsukakushi> ncopa: ever considered getting a bouncer?
2016-05-09 15:12:11 <skarnet> 
ACTION volunteers to be a bouncer, as in the big guy who gets to kick people out

2016-05-09 15:12:42 <Tsutsukakushi> t-that's called an operator
2016-05-09 17:16:19 <jirutka> could please someone help me with writing abuild for R? I don’t know how to deal with:
2016-05-09 17:16:19 <jirutka> >>> R*: Tracing dependencies...
2016-05-09 17:16:19 <jirutka> >>> ERROR: R*: libR.so: path not found
2016-05-09 17:56:28 <fcolista> jirutka, does libR.so exists? And maybe is in a different path where the linker expects to find him?
2016-05-09 17:56:51 <fcolista> Or is called differently, like libR.so.xyz ?
2016-05-09 17:57:43 <jirutka> yes, it exists, but in /usr/lib/R/libs… Alba already gave me a hint to use ldpath
2016-05-09 17:58:59 <fcolista> ok
2016-05-09 17:59:10 <ScrumpyJack> fcolista: thanks for bumping moc. It was on my list but i guess I'm not quick enough
2016-05-09 18:00:37 <jirutka> however, I have another question… R also provides libRmath.so, called standalone shared lib; as I googled, it’s used as a dependency for C programs, but it’s not quite standalone, it’s linked against libR.so… I’m thinking how to name subpackage for this… should I add it to -dev, or something else?
2016-05-09 18:01:22 <jirutka> or -dev should be just header files, pkgconfig and similar stuff?
2016-05-09 18:05:50 <fcolista> jirutka, " --enable-R-shlib        build the shared/dynamic library 'libR' [no]"
2016-05-09 18:06:28 <jirutka> nope, see https://git.archlinux.org/svntogit/packages.git/tree/r/trunk/PKGBUILD#n56
2016-05-09 18:07:05 <fcolista> then?
2016-05-09 18:07:14 <skarnet> -dev is header files, static libraries, pkg-config and .so links
2016-05-09 18:07:15 <fcolista> --enable-R-shlib is enabled in PKGBUILD
2016-05-09 18:07:23 <fcolista> but not in APKBUILD
2016-05-09 18:07:26 <skarnet> real shared libraries are in the original package
2016-05-09 18:07:53 <jirutka> I’m writing new APKBUILD, the current one is… well… very bad
2016-05-09 18:07:57 <fcolista> skarnet: +1.
2016-05-09 18:08:20 <jirutka> I’ve already enabled --enable-R-shlib, but that’s not what I’m asking about
2016-05-09 18:11:06 <fcolista> you can call the subpackage for libRmath with r-mathlibs, as debian does
2016-05-09 18:11:18 <fcolista> *r.mathlib
2016-05-09 18:11:40 <fcolista> you build the package, then in the subackage you move the lib
2016-05-09 18:12:11 <jirutka> what should r-mathlibs contain? just libRmath.so?
2016-05-09 18:12:57 <jirutka> btw built-in R packages, basically a standard library, should be in a subpackage *-libs, right?
2016-05-09 18:13:00 <fcolista> yes. Also the include file .h might have another subpackage, as skarnet correctly stated, as r-mathlibs-dev
2016-05-09 18:13:32 <jirutka> I’m quite confused from plural, why mathlibs when it contains just a single lib? :)
2016-05-09 18:13:42 <fcolista> it's mathlib
2016-05-09 18:13:49 <fcolista> fcolista> *r.mathlib
2016-05-09 18:13:53 <fcolista> *r-mathlib
2016-05-09 18:14:05 <jirutka> aha, okay :)
2016-05-09 18:14:14 <fcolista> https://packages.debian.org/sid/r-mathlib
2016-05-09 18:14:15 <skarnet> jirutka: you're confusing it with skalibs ;)
2016-05-09 18:14:20 <fcolista> :)
2016-05-09 18:14:23 <jirutka> XD
2016-05-09 18:15:19 <jirutka> so, r, r-libs, r-doc, r-dev, r-mathlib, r-mathlib-dev?
2016-05-09 18:15:45 <fcolista> why r-libs?
2016-05-09 18:16:02 <jirutka> standard library of R
2016-05-09 18:16:08 <fcolista> it's part of R
2016-05-09 18:16:14 <fcolista> or
2016-05-09 18:16:22 <fcolista> if you want to follow the debian approac
2016-05-09 18:16:25 <fcolista> *approach
2016-05-09 18:16:27 <jirutka> hm, for example Ruby and Lua has separated stdlib
2016-05-09 18:16:34 <fcolista> you might do r-core
2016-05-09 18:17:07 <fcolista> which distributions are packaging "r-libs" ?
2016-05-09 18:17:38 <jirutka> dunno, I’ve looked just in Arch and Gentoo
2016-05-09 18:17:51 <jirutka> but Alpine has different rules for subpackages
2016-05-09 18:18:14 <fcolista> Is Arch doing r-libs?
2016-05-09 18:18:17 <jirutka> nope
2016-05-09 18:18:30 <fcolista> ah ok
2016-05-09 18:19:14 <jirutka> I followed example of Ruby and Lua in Alpine… and also Erlang, according to names https://pkgs.alpinelinux.org/packages?name=*-stdlib&branch=&repo=&arch=x86_64&maintainer=
2016-05-09 18:19:24 <jirutka> however, I’m not sure if it’s really useful in this case
2016-05-09 18:20:04 <fcolista> which library are you using for blas?
2016-05-09 18:20:24 <jirutka> maybe someone may need r-mathlib, but not whole stdlib; however, there’s a problem that r-mathlib depends on r and r should also install stdlibs
2016-05-09 18:20:43 <jirutka> lapack
2016-05-09 18:21:26 <jirutka> but, I’m not enterily sure if it really provides blas… but according to https://git.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/lapack it should
2016-05-09 18:21:41 <jirutka> I’m currently trying to figure out this
2016-05-09 18:23:33 <fcolista> Alpine has openblas
2016-05-09 18:23:39 <fcolista> dunno if it helps, though
2016-05-09 18:25:27 <jirutka> what’s difference between lapac, openblas, …?
2016-05-09 18:25:50 <fcolista> Dunno
2016-05-09 18:26:19 <jirutka> and according to https://pkgs.alpinelinux.org/contents?branch=edge&name=lapack&arch=x86_64&repo=testing, I lapack seems to provide libblas
2016-05-09 18:26:33 <jirutka> also ldd on compiled .so files from R seems to be correct
2016-05-09 19:12:02 <jirutka> fcolista: I’m done, could I ask you for a review? https://github.com/jirutka/alpine-aports/blob/r/testing/R/APKBUILD (https://github.com/alpinelinux/aports/pull/74/files)
2016-05-09 21:03:41 <mitchty> okies, so once ghc 8.0 is released I'm ready to update my ghc port to that so that arm support works, was hoping it would hit prior to 3.4 but not happening anytime nearish soon
2016-05-09 21:04:23 <mitchty> so i assume the order of operations is testing->community?
2016-05-09 21:13:08 <barthalion> jirutka: stdlib != -libs subpackage…
2016-05-09 21:14:01 <jirutka> it is in the case of Erlang https://pkgs.alpinelinux.org/package/edge/community/x86_64/erlang-stdlib
2016-05-09 21:14:16 <jirutka> but not in Lua, hmm
2016-05-10 01:49:44 <Sachiru> Does Alpine Linux have support for running the check_mk agent?
2016-05-10 05:46:11 <barthalion> Sachiru: if it is packaged or you plan to, why not?
2016-05-10 09:47:11 <ScrumpyJack> skarnet: you about?
2016-05-10 09:47:35 <skarnet> hey :) what's up?
2016-05-10 09:48:03 <ScrumpyJack> do you have any instructions on replacing open-rc with s6 in AL?
2016-05-10 09:48:15 <skarnet> woah there.
2016-05-10 09:48:22 <skarnet> I don't, not yet.
2016-05-10 09:48:58 <skarnet> There needs to be much work done before it's easy to do (and automatable etc.)
2016-05-10 09:49:36 <skarnet> replacing init is relatively easy; replacing the service manager (OpenRC) is another enchilada
2016-05-10 09:51:15 <skarnet> I'll do that work next year, I'll promise. I'm currently having a $dayjob in order to be able to set aside enough money and time to work on this project full-time until completion.
2016-05-10 09:52:38 <skarnet> Until then, it's obviously possible to use s6, and s6-rc, and other parts of the s6 ecosystem, on AL, but it won't integrate smoothly with the way AL manages its services, and you'll have to do the conversion by hand.
2016-05-10 10:25:36 <ScrumpyJack> kk
2016-05-10 10:27:59 <ScrumpyJack> anyone tried X on an rpi 3 with edge?
2016-05-10 10:42:39 <ScrumpyJack> arg, AL rpi3 wiki page says Disable glx. well *don't* :)
2016-05-10 11:20:31 <ScrumpyJack> does lbu support standard globbing?
2016-05-10 11:21:35 <ScrumpyJack> i want to save a dirname but not the contents
2016-05-10 11:26:59 <leo-unglaub> hey friends :)
2016-05-10 11:27:49 <leo-unglaub> how are you?
2016-05-10 11:33:21 <jomat> morning leo… just reading the "news" about grsec *sigh*
2016-05-10 11:43:08 <leo-unglaub> hehe, its not as big as people make it
2016-05-10 11:43:16 <leo-unglaub> you have something like that every couple weeks
2016-05-10 11:46:59 <jomat> But I finally resolved the dependency hell that opened its gates when trying to install ncmpcpp :-)
2016-05-10 11:48:12 <leo-unglaub> thats good to know :)
2016-05-10 11:49:14 <jirutka> serious question: how hard it would be to make Alpine compatible also with OpenBSD kernel? for example Gentoo runs even on FreeBSD kernel, but I don’t know how much effort it cost
2016-05-10 11:59:21 <leo-unglaub> jirutka: well, it depends on how you want to do it
2016-05-10 11:59:32 <leo-unglaub> there are several ways of doing it
2016-05-10 11:59:57 <leo-unglaub> if you want to keep the OpenBSD kernel and put Alpine tools on top of it like apk and so on its very easy
2016-05-10 12:02:13 <jirutka> other options?
2016-05-10 12:02:50 <leo-unglaub> you try to get some binary compat working and want to use the same packges on both versions -> impossible task
2016-05-10 12:04:18 <leo-unglaub> jirutka: are you interrested in this subject?
2016-05-10 12:04:54 <jirutka> well, I am, but to be honest, it’s not so improtant for me to actual do any work on it
2016-05-10 12:19:52 <ScrumpyJack> alpine-linux, alpine-bsd, alpine-minix, alpine-hurd, alpine-takes-over-the-world! :)
2016-05-10 12:20:02 <jirutka> XD
2016-05-10 12:20:18 <jirutka> alpine-hurd would be just on the paper? XD
2016-05-10 14:36:48 <ScrumpyJack> anyone know anything about font aports?
2016-05-10 16:13:29 <jirutka> ncopa, could you please look at https://github.com/alpinelinux/aports/pull/72 and https://github.com/alpinelinux/aports/pull/73 ?
2016-05-10 16:22:45 <ncopa> jirutka, the only thing i am in doubt with is the /usr/lib/qemu -> /usr/libexec/qemu change
2016-05-10 16:28:13 <ncopa> I think the reason was that FHS 2.3 does not mention /usr/libexec
2016-05-10 16:29:21 <skarnet> why do you care about FHS at all? :/
2016-05-10 17:24:10 <nmeum> what's the difference between testing/libyaml and main/yaml if I am not mistaken it is the exact same sofware only with different versions
2016-05-10 17:32:09 <nmeum> If nobody has any objections I will remove testing/libyaml (the only package explicitly depending on it seems to be testing/h2o)
2016-05-10 17:35:19 <Tsutsukakushi> skarnet: <- that
2016-05-10 17:35:40 <Tsutsukakushi> FHS is not nearly comprehensive enough
2016-05-10 17:53:57 <skarnet> Tsutsukakushi: the problem I have with FHS is not that it isn't comprehensive, it's that it addresses obsolete issues, misses current issues, and makes arbitrary, technically bad policy decisions
2016-05-10 17:54:33 <skarnet> I don't like comprehensive stuff. Less is better than more.
2016-05-10 17:54:41 <Tsutsukakushi> well
2016-05-10 17:54:56 <Tsutsukakushi> the misses current issues is what i meant by not comprehensive enough
2016-05-10 17:55:04 <skarnet> I can agree with that :)
2016-05-10 21:09:14 <jirutka> ncopa: Alpine already uses /usr/libexec, so I’m just trying to be consistent ;)
2016-05-10 21:09:43 <jirutka> ~ # ls -la /usr/libexec/
2016-05-10 21:09:44 <jirutka> total 56
2016-05-10 21:09:44 <jirutka> drwxr-xr-x    1 root     root            28 May  8 17:13 .
2016-05-10 21:09:44 <jirutka> drwxr-xr-x    1 root     root           116 May  3 19:41 ..
2016-05-10 21:09:44 <jirutka> drwxr-xr-x    1 root     root            48 May  3 19:41 gcc
2016-05-10 21:09:44 <jirutka> -rwxr-xr-x    1 root     root         55336 Oct 29  2015 rmt
2016-05-10 21:09:44 <jirutka> drwxr-xr-x    1 root     root          4168 May  3 19:41 git-core
2016-05-10 21:20:27 <jirutka> regarding to FHS… is there any rational reason to keep /usr in Alpine? it’d be much simpler and cleaner to use just /bin, /lib, …
2016-05-10 21:23:38 <skarnet>  /usr/include, for libc headers ? You don't want /include :)
2016-05-10 21:23:50 <jirutka> why not?
2016-05-10 21:24:24 <skarnet> I don't like clutter in /, but that's just my sense of aesthetics
2016-05-10 21:24:44 <mitchty> /usr/bin/env for one no?
2016-05-10 21:25:02 <skarnet> no, /bin/env is valid as well
2016-05-10 21:25:20 <jirutka> I understand that, but when dropping /usr, we should be consistent; it doesn’t make much sense to keep it for something, that would be just arbitrary decisions
2016-05-10 21:25:21 <mitchty> just thinking of all the scripts with /usr/bin/env
2016-05-10 21:25:30 <jirutka> yeah, that’s true :(
2016-05-10 21:25:37 <jirutka> we can’t remove /usr/bin/env
2016-05-10 21:25:39 <skarnet> mitchty: saying "use #!/usr/bin/env myinterpreter for reliable shebangs" is wrong
2016-05-10 21:25:54 <skarnet> there is NO standard that says that /usr/bin/env has to exist
2016-05-10 21:26:09 <skarnet> and using /usr/bin/env in shebangs is only displacing the problem
2016-05-10 21:26:10 <mitchty> i'm not saying it has to exist
2016-05-10 21:26:17 <jirutka> maybe, but the fact is that it’s “standard”
2016-05-10 21:26:27 <skarnet> #!/usr/bin/env scripts are *broken*
2016-05-10 21:26:28 <jirutka> you will break too much software by dropping /usr/bin/env
2016-05-10 21:26:49 <mitchty> then all scripts with /usr in them are broken
2016-05-10 21:26:56 <jirutka> it’s still less broken than hardcoding absolute paths to shebang…
2016-05-10 21:27:06 <mitchty> you'll be patching up a lot of scripts either way
2016-05-10 21:27:08 <skarnet> no, it's exactly as broken
2016-05-10 21:27:12 <skarnet> not more, not less
2016-05-10 21:27:18 <jirutka> ?
2016-05-10 21:27:32 <jirutka> so what you prefer? to not use shebang at all or what?
2016-05-10 21:27:34 <skarnet> in both cases, you don't have the guarantee that the absolute path to the binary is correct
2016-05-10 21:27:56 <skarnet> my own preference is slashpackage, which does guarantee absolute paths
2016-05-10 21:28:09 <jirutka> ?
2016-05-10 21:28:15 <skarnet> the FHS does not offer a good solution to that, so you just have a choice of brokenness
2016-05-10 21:28:18 <jirutka> slashpackage?
2016-05-10 21:28:34 <skarnet> a convention for fixed absolute paths.
2016-05-10 21:28:50 <skarnet> which, unfortunately, very few people follow.
2016-05-10 21:29:23 <skarnet> which makes it basically useless, since a convention is only as good as the number of people who follow it. But it still offers that very guarantee that the FHS does not.
2016-05-10 21:29:29 <jirutka> I’ve never seen that
2016-05-10 21:29:32 <jirutka> you mean https://cr.yp.to/slashpackage.html?
2016-05-10 21:29:37 <skarnet> yes.
2016-05-10 21:29:49 <jirutka> what distro use this? like, no one?
2016-05-10 21:30:14 <skarnet> unfortunately, no. There's GoboLinux which uses something similar (if not exactly slashpackage).
2016-05-10 21:30:24 <skarnet> But that doesn't make scripts portable.
2016-05-10 21:30:42 <skarnet> anyway, the problem comes from hardcoding paths in shebangs.
2016-05-10 21:30:44 <mitchty> as a note, nixos, which is the most deviant unix i know of, only has /usr/bin/env as a link to env
2016-05-10 21:31:03 <jirutka> I agree that FHS is not good, but we can rewrite the whole world, I’d like to discuss actually feasable solutions, not unreachable ideal
2016-05-10 21:31:13 <skarnet> The real solution to the problem would be to make the kernel perform PATH search in shebangs.
2016-05-10 21:31:20 <skarnet> #!interpreter should work.
2016-05-10 21:31:33 <jirutka> yeah, but it doesn’t
2016-05-10 21:31:47 <jirutka> and how it’s different from /usr/bin/env? not so much if you think about it…
2016-05-10 21:31:52 <skarnet> (it works on some versions of Linux, but it doesn't anymore)
2016-05-10 21:31:58 <jirutka> since every distro has /usr/bin/env
2016-05-10 21:32:05 <skarnet> nope.
2016-05-10 21:32:24 <mitchty> and if you need to run on a bsd then what?
2016-05-10 21:32:44 <skarnet> then the BSD kernels should *also* perform PATH searchs in shebangs. :)
2016-05-10 21:32:47 <mitchty> /usr/bin/env is pretty much a defacto standard
2016-05-10 21:32:51 <jirutka> I have /usr/bin/env even on OS X, which is based on FreeBSD
2016-05-10 21:33:16 <skarnet> try some router installations which use busybox with the "don't use /usr" option.
2016-05-10 21:33:53 <jirutka> #!ruby, or #!/usr/bin/env ruby… it’s basically just a syntax difference
2016-05-10 21:34:05 <skarnet> And the fact that "every distro does it that way!!!one!" is NOT the same thing as "it's the right way to do it"
2016-05-10 21:34:09 <jirutka> #!ruby would probably do exactly the same thing as env
2016-05-10 21:34:17 <jirutka> I agree
2016-05-10 21:34:32 <skarnet> #!ruby does not depend on where the env binary lives
2016-05-10 21:34:45 <skarnet> #!/usr/bin/env ruby only works if env is in /usr/bin
2016-05-10 21:34:52 <skarnet> which it does not have to be.
2016-05-10 21:35:11 <jirutka> but we don’t live in an ideal world… it’s not reasanable to patch every software in the world to work on Alpine just because of this decission
2016-05-10 21:35:15 <mitchty> only problem is #!thing is supposed to mean #!./thing in posix iirc, let me look
2016-05-10 21:35:49 <skarnet> jirutka: I agree, I didn't say breaking it voluntarily was a good idea.
2016-05-10 21:35:54 <jirutka> we should be glad that at least /usr/bin/env exists and it’s de-facto standard
2016-05-10 21:36:09 <skarnet> For practical purposes, keeping /usr/bin/env is an acceptable solution (it's not worse than any other place)
2016-05-10 21:36:17 <jirutka> exactly
2016-05-10 21:36:44 <skarnet> I was just insisting on the fact that it's not a requirement and scripts should *not* rely on it, because I find the distinction important.
2016-05-10 21:37:10 <jirutka> btw is it Fedora? that dropped /bin, /lib, … in favor of /usr/bin, /usr/lib…
2016-05-10 21:37:24 <skarnet> it's Lennart
2016-05-10 21:37:31 <jirutka> uh, okay
2016-05-10 21:37:31 <skarnet> so yeah, Fedora.
2016-05-10 21:37:35 <jirutka> I didn’t say anything!
2016-05-10 21:37:41 <jirutka> please remove this from the record
2016-05-10 21:38:39 <jirutka> once again, what solution would you suggest? #!thing doesn’t work on Linux, period. absolute paths are much worse. as I see it, /usr/bin/env is still the only working solution and less evil
2016-05-10 21:39:19 <skarnet> the fun thing is that they want to abolish the distinction between / and /usr (which still has a semantic value today) and they don't say anything about merging /bin and /sbin (the distinction doesn't make sense at all on modern systems)
2016-05-10 21:39:21 <jirutka> and envpackage doesn’t work on Linux as well
2016-05-10 21:40:36 <skarnet> The correct solution is not to hardcode shebangs in scripts. To distribute scripts, have an installation script that puts the administrator-chosen location in the shebang.
2016-05-10 21:41:02 <jirutka> so just add yet another moving part…
2016-05-10 21:41:07 <jirutka> and increase complexity
2016-05-10 21:41:10 <jirutka> not a good solution
2016-05-10 21:41:17 <skarnet> it's the price of flexibility. You can't have it both ways.
2016-05-10 21:42:06 <jirutka> the better solution is IMO to stick to one and the only one reasonable standard
2016-05-10 21:43:00 <skarnet> s/reasonable/vastly deployed/
2016-05-10 21:43:28 <jirutka> well… I can’t agree with that
2016-05-10 21:43:51 <jirutka> yes, it’s true in this case, but I don’t see /usr/bin/env as a totally bad or evil solution
2016-05-10 21:44:02 <jirutka> it’s not ideal, but it’s also not bad
2016-05-10 21:44:02 <skarnet> and your "better" solution will piss off people like me who build minimal environments by hand, don't use /usr, and have to patch broken scripts.
2016-05-10 21:44:41 <jirutka> I would never ever argue to stick to the only one vastly deployed way… like I would never say that everyone should use systemd crap
2016-05-10 21:44:58 <skarnet> oh, I wasn't even thinking about that.
2016-05-10 21:45:11 <jirutka> come on, you can just symlink it
2016-05-10 21:45:17 <jirutka> it’s just one, only one binary
2016-05-10 21:45:27 <jirutka> and we’re talking just about some arbitrary path
2016-05-10 21:45:28 <skarnet> and make a /usr and a /usr/bin directory just for it?
2016-05-10 21:45:45 <skarnet> If I don't have a /usr, I don't have a /usr, period
2016-05-10 21:46:05 <jirutka> yes… not ideal, but to be honest, this is almost just an aesthetic problem
2016-05-10 21:46:28 <jirutka> still better than force people to template all the scripts
2016-05-10 21:46:36 <jirutka> or to patch all the script myself
2016-05-10 21:46:46 <jirutka> or any other actually available option
2016-05-10 21:46:55 <mitchty> http://www.in-ulm.de/~mascheck/various/shebang/#env from 2001, apparently only sco openserver and redhat had /bin/env the latter with a link in /usr/bin
2016-05-10 21:47:57 <jirutka> but of course, #!thing would be more systematic solution, but sadly it’s not supported
2016-05-10 21:48:08 <skarnet> some day, I hope.
2016-05-10 21:50:24 <jirutka> btw today I’ve met one friend who works as a system architect for critical systems for goverment… I think that I’m gonna have nightmares for few days…
2016-05-10 21:50:34 <ammunta> hey remember
2016-05-10 21:51:00 <ammunta> #!/bin/sh
2016-05-10 21:51:00 <ammunta> #! -*-perl-*-
2016-05-10 21:51:00 <ammunta> eval 'exec perl -x -wS $0 ${1+"$@"}'
2016-05-10 21:51:00 <ammunta> if 0;
2016-05-10 21:51:19 <jirutka> what’s that?
2016-05-10 21:51:22 <ammunta> :&&
2016-05-10 21:51:39 <skarnet> ammunta: gtfo.
2016-05-10 21:51:45 <ammunta> the voodoo perlrun documents for figuring out where perl is
2016-05-10 21:51:50 <jirutka> ${1+"$@"} ??
2016-05-10 21:52:01 <skarnet> standard shell idiom.
2016-05-10 21:52:03 <jirutka> shouldn’t it be ${1:+"$@"} ?
2016-05-10 21:52:24 <ammunta> no
2016-05-10 21:52:46 <ammunta> $1 can be empty
2016-05-10 21:54:15 <jirutka> I’ve justr tried it and it seems to work exactly the same as just "$@"… what I’m missing here?
2016-05-10 21:54:58 <skarnet> http://unix.stackexchange.com/questions/68484/what-does-1-mean-in-a-shell-script-and-how-does-it-differ-from
2016-05-10 21:55:35 <ammunta> https://www.gnu.org/software/autoconf/manual/autoconf-2.67/html_node/Shell-Substitutions.html#Shell-Substitutions old implementations, apparently
2016-05-10 21:57:11 <jirutka> mind = blown
2016-05-10 21:58:18 <ammunta> also autoconf docs about portable shell are great for totally useless trivia about unices that haven't existed for thirty years
2016-05-10 21:58:40 <skarnet> ^
2016-05-10 21:58:41 <poptart> some of us still use the 1+ because I want my scripts to run on anything lol
2016-05-10 21:59:23 <ammunta> they won't
2016-05-10 22:00:53 <mitchty> after reading the almquist shell source quick, yeah shebangs are broken
2016-05-10 22:01:09 <ammunta> the sort of systems where the 1+ is relevant are so far removed from modern systems that you can basically assume it won't work
2016-05-10 22:01:38 <jirutka> I prefer to write shell scripts that are actually *readable* and robust, not to run on systems that even doesn’t exist anymore…
2016-05-10 22:02:03 <jirutka> but since using Alpine I now prefer to stick to POSIX shell and not Bash as before
2016-05-10 22:02:34 <skarnet> nice :)
2016-05-10 22:03:01 <jirutka> btw what’s that nonsense: "x$foo" = "x" … why some scripts uses this oddity?
2016-05-10 22:04:11 <ammunta> some shell's test didn't handle empty strings well probably
2016-05-10 22:04:31 <jirutka> what shells? yet another bug of ancient Born Shell?
2016-05-10 22:04:49 <mitchty> original bourne yes
2016-05-10 22:05:11 <ammunta> or possibly mistreated it if $foo started with -
2016-05-10 22:05:47 <ammunta> or any number of things could have been wrong
2016-05-10 22:05:54 <ammunta> don't dive too deep into the portability rabbit hole
2016-05-10 22:09:02 <mitchty> i used solaris and hpux, old school bourne /bin/sh was fun
2016-05-10 22:09:35 <skarnet> ew. /usr/xpg4/bin/sh please.
2016-05-10 22:09:57 <mitchty> no way /usr/ucb/bin/sh iirc, its been a while
2016-05-10 22:10:26 <skarnet> hm, I forget what /usr/ucb is. Wasn't that GNU stuff?
2016-05-10 22:10:37 <mitchty> i thought it was the system5 stuff
2016-05-10 22:10:37 <ammunta> isn't ucb berkeley
2016-05-10 22:10:54 <mitchty> might have been the bsd stuff from old sunos, i forget tbh
2016-05-10 22:10:57 <skarnet> ucb sounds like University of California Berkeley indeed
2016-05-10 22:11:14 <mitchty> solaris was an odd beast
2016-05-10 22:11:31 <skarnet> odd and slow
2016-05-10 22:11:49 <ammunta> okay but doors
2016-05-10 22:12:01 <skarnet> we don't talk about doors
2016-05-10 22:13:21 <ammunta> how about STREAMS :DD
2016-05-10 22:13:54 <jirutka> doors?
2016-05-10 22:14:18 <ammunta> some insane variety of bespoke ipc
2016-05-10 22:14:19 <skarnet> doors and STREAMS were Solaris-specific stuff that you don't want to know about.
2016-05-10 22:14:32 <jirutka> aha :)
2016-05-10 22:16:03 <ammunta> https://www.gnu.org/software/autoconf/manual/autoconf-2.67/html_node/Slashes.html#Slashes
2016-05-10 22:17:12 <mitchty> oh god STREAMS
2016-05-11 00:07:16 <saintdev> the modules and hwclock init scripts have circular dependencies. modules has 'before hwclock' and hwclock has 'before *'
2016-05-11 03:51:07 <ScrumpyJack> morning
2016-05-11 04:13:53 <coredumb> morning
2016-05-11 07:20:44 <xming> slowaris, oh that's so long ago
2016-05-11 09:45:34 <leo-unglaub> hey :)
2016-05-11 09:52:08 <clandmeter> howdy
2016-05-11 10:00:25 <leo-unglaub> last night my windows 7 gaming PC did an update to windows 10 without asking me ... of course the update failed and my windows pc was broken when i wanted to play world of warcraft with friends. so i reinstalled everything. in the process windows 10 killed all bootloaders on all my 8 harddrives to write the windows loader there ... killing my alpine linux and my openbsd in the process ... but then windows did not know that 
2016-05-11 10:00:25 <leo-unglaub> but it gets even better ... i did not want to argue and wanted to play games because i am on a lan party ... so i bought it over the online store ... but after logging into the online store microsoft changes my local account password to the one from the online store ... so i could not login anymore because my online password is only in my password manager ... so i was logged out again
2016-05-11 10:00:40 <leo-unglaub> all i can say is: thanks open source for not beeing sooo fucked up as windows and apple are
2016-05-11 10:03:33 <clandmeter> dont play games, its useless anyway.
2016-05-11 10:05:02 <leo-unglaub> never *g*
2016-05-11 10:05:24 <leo-unglaub> i am a lvl 100 paladin tank ... i am never giving that up *g*
2016-05-11 10:05:29 <clandmeter> but from a business point of view, i should probably advise you to do it :)
2016-05-11 10:06:16 <leo-unglaub> hehe, from a business point of view you are right ... on the other hand .. i am brok since i am 15 years old because 99% of my work is open source and unpayed
2016-05-11 10:06:37 <coredumb> just make kids
2016-05-11 10:06:46 <coredumb> no more time to play WoW
2016-05-11 10:07:26 <coredumb> s/to play WoW/for anything/
2016-05-11 10:08:00 <leo-unglaub> *g*
2016-05-11 10:08:38 <leo-unglaub> i know ... for around a year or so i took care of a child of one friend of mine ...
2016-05-11 10:08:44 <leo-unglaub> absolutly awesome time ...
2016-05-11 10:09:21 <coredumb> hence no need for a windows gaming rig
2016-05-11 10:09:21 <coredumb> \o/
2016-05-11 10:09:54 <clandmeter> i love kids, especially on an airplane.
2016-05-11 10:11:30 <coredumb> I like the "..." behind awesome time
2016-05-11 10:12:46 <leo-unglaub> coredumb: thats because i dont see him anymore because they moved away .. and i miss that little guy
2016-05-11 10:13:23 <coredumb> :)
2016-05-11 10:13:39 <coredumb> leo-unglaub: that's the nice thing about kids
2016-05-11 10:13:46 <coredumb> it's easy to make ^^
2016-05-11 10:14:17 <leo-unglaub> yeah, its not as hard as making a rust package for alpine *g*
2016-05-11 10:14:29 <skarnet> tell that to the mother
2016-05-11 10:15:53 <coredumb> skarnet: that's the good part of being the father :)
2016-05-11 10:18:29 <coredumb> althought I'm sure my wife would agree on leo-unglaub statement that it's easier to make a kid than a rust package :P
2016-05-11 10:19:25 <skarnet> Finding a woman who doesn't want kids was a blessing. There's "us two" time, and there's "she reads, I code" time. Both of which would be sorely missing if we had kids.
2016-05-11 10:21:38 <coredumb> agree
2016-05-11 10:21:42 <coredumb> I'm missing this
2016-05-11 10:21:46 <coredumb> but well
2016-05-11 10:22:12 <coredumb> now there's also "me and the kids" time if I don't have to much work
2016-05-11 10:22:27 <coredumb> and it's even better than "code" time ;)
2016-05-11 10:23:18 <coredumb> so actually what happens is that code time happens when everyone else sleeps :D
2016-05-11 10:24:31 <skarnet> that can work as long as you don't need much sleep
2016-05-11 10:25:03 <coredumb> luckily I can live OK with 4h
2016-05-11 10:25:17 <skarnet> that's lucky indeed!
2016-05-11 10:25:38 <coredumb> "OK"
2016-05-11 10:26:45 <coredumb> I'm better with 5 but I tend to go as low as 2
2016-05-11 10:26:54 <coredumb> I shouldn't ... I'm getting old :D
2016-05-11 10:27:13 <skarnet> 2 is crazy, don't do that.
2016-05-11 10:27:45 <coredumb> yeah not often
2016-05-11 10:28:01 <coredumb> you feel old all day long :D
2016-05-11 10:28:33 <skarnet> I'm quite functional with 6, but I do need my 6 hours - else I need a nap.
2016-05-11 10:29:32 <coredumb> hehe I can't nap :D
2016-05-11 10:30:21 <skarnet> can't as in you won't fall asleep, or as in you're at work and don't have the possibility to nap?
2016-05-11 10:31:07 <coredumb> I won't fall
2016-05-11 10:31:14 <ScrumpyJack> great thing about kids is that one day they leave
2016-05-11 10:31:24 <coredumb> and If I do I'm dead when waking up
2016-05-11 10:31:46 <ScrumpyJack> (I'm kidding of course)
2016-05-11 10:32:01 <coredumb> ScrumpyJack: When they're 5 you can start making them cycling in the basement to generate electricity !
2016-05-11 10:32:06 <leo-unglaub> 2 hours of sleep is okay
2016-05-11 10:32:15 <ncopa> fabled, re py-willow #include "wchar.h" and -D_GNU_SOURCE
2016-05-11 10:32:17 <leo-unglaub> when i drive at night at the red cross i have no sleep at all
2016-05-11 10:32:31 <ncopa> is -D_GNU_SOURCE requirement for #include <wchar.h>?
2016-05-11 10:33:25 <skarnet> leo-unglaub: it's ok when you're young and strong!
2016-05-11 10:33:29 <ScrumpyJack> mine is 18 and off to uni in september. she'll be missed, of course, but it's going to be a quiet household after that
2016-05-11 10:34:32 <leo-unglaub> ncopa: did you get some new irc permissions?
2016-05-11 10:34:40 <leo-unglaub> because my client is now writing your name bold
2016-05-11 10:35:04 <coredumb> ScrumpyJack: is she a good teen ?
2016-05-11 10:35:08 <leo-unglaub> ncopa: http://i.imgur.com/vrT740n.png
2016-05-11 10:35:49 <skarnet> leo-unglaub: you and I are in bold too. The only person you don't like is ScrumpyJack.
2016-05-11 10:35:54 <ncopa> just new irc client
2016-05-11 10:36:02 <ncopa> i�m on mac atm
2016-05-11 10:36:16 <leo-unglaub> ah, interresting :)
2016-05-11 10:36:50 <skarnet> ncopa: mac? Docker is slowly corrupting you, beware. :P
2016-05-11 10:37:04 <coredumb> hehe
2016-05-11 10:37:09 <jirutka> skarnet: OS X is great system
2016-05-11 10:37:20 <skarnet> shoo.
2016-05-11 10:37:21 <coredumb> really ?
2016-05-11 10:37:27 <coredumb> like REALLY ?
2016-05-11 10:37:28 <fabled> ncopa, mempcpy is #include <string.h> + _GNU_SOURCE
2016-05-11 10:37:44 <coredumb> I really don't get people finding nice OS X
2016-05-11 10:37:52 <jirutka> yes, it’s a unix with actually usable GUI
2016-05-11 10:38:44 <xming> I like OSX except for its GUI
2016-05-11 10:39:03 <jirutka> Linux is the best for servers, but I really don’t like it on desktop
2016-05-11 10:39:10 <coredumb> the GUI of OS X makes me wanna puke everytime
2016-05-11 10:39:35 <xming> it's probably the worst GUI which I have ever used
2016-05-11 10:39:37 <skarnet> jirutka: but the year of Linux on the desktop is happening real soon now!
2016-05-11 10:39:39 <jirutka> I’m still on Mavericks, because I hate how they destroyed it in newer versions
2016-05-11 10:40:01 <xming> snow leopard is the last good OSX
2016-05-11 10:40:05 <jirutka> you mean the year of RedHat dictatorship?
2016-05-11 10:40:06 <ScrumpyJack> coredumb: she'd call herself a young adult now. and yes, she's been fantastic. we've been very lucky. I hope it stays that way
2016-05-11 10:40:16 <xming> skarnet: yes beacuse desktop market is dying :D
2016-05-11 10:40:33 <coredumb> ScrumpyJack: cool :)
2016-05-11 10:41:03 <ScrumpyJack> skarnet: " The only person you don't like is ScrumpyJack" I'm sure you mean someting else :)
2016-05-11 10:41:03 <skarnet> look up "year of the linux desktop" if you're not 100% I was sarcastic...
2016-05-11 10:41:09 <skarnet> 100% certain*
2016-05-11 10:41:28 <xming> I am sure the year of Linux on the desktop will be soon
2016-05-11 10:41:49 <skarnet> ScrumpyJack: damn, my attempt at sowing trouble between you and leo-unglaub has failed
2016-05-11 10:42:43 <ncopa> fabled, interestingly enough, if i do #include <string.h> before <wchar.h> it also works
2016-05-11 10:42:46 <xming> when the general public all uses laptops/tablets/phablets/phones
2016-05-11 10:43:34 <fabled> ncopa, huh. that sounds weird.
2016-05-11 10:44:11 <jirutka> desktop and laptop is the same for me
2016-05-11 10:44:22 <coredumb> actually as I'm using Linux as a desktop for the last ~15 years I couldn't bother less if it's THE year or not :D
2016-05-11 10:44:24 <jirutka> I don’t use desktop since… I don’t even know… use just laptop
2016-05-11 10:44:30 <xming> I have an issue with nginx, the APKBUILD has pkgusers="nginx" and install -dm700 -o $pkgusers -g $_grp_ngx "$pkgdir"/var/lib/$pkgname/tmp
2016-05-11 10:44:51 <xming> but somehow /var/lib/nginx/tmp is owned by root
2016-05-11 10:45:10 <xming> coredumb: same here
2016-05-11 10:45:25 <jirutka> and phablets/phones… gosh I don’t understand why so many ppl prefer these over laptops… I suffer every time I use phone, the user experience is more than horrible
2016-05-11 10:45:38 <xming> I hate laptops
2016-05-11 10:46:02 <ncopa> fabled, yes, something is weird.  don't think they use memcpy at all either
2016-05-11 10:46:05 <jirutka> even trivial tasks that I can do on desktop/laptop without even thinking in less than few miliseconds are extremly complicated on these mobile craps
2016-05-11 10:46:34 <jirutka> I use MacBook Pro as my primary and the only one computer
2016-05-11 10:46:53 <jirutka> well, only one desktop-like computer
2016-05-11 10:47:03 <skarnet> To be productive, I need a screen, a keyboard, an Internet connection, and a decent terminal + ssh client. All the rest is completely irrelevant to me - but it rules out phones.
2016-05-11 10:47:07 <fabled> ncopa, probably the fortify headers
2016-05-11 10:47:56 <ncopa> yes
2016-05-11 10:47:58 <fabled> or
2016-05-11 10:48:11 <ncopa> wchar pulls in string.h
2016-05-11 10:48:14 <fabled> the app defining _GNU_SOURCE somewhere between string.h and wchar.h
2016-05-11 10:49:01 <jirutka> ncopa: Alpine already uses /usr/libexec, so I’ve just wanted to be consistent ;)
2016-05-11 10:49:15 <jirutka> # ls -la /usr/libexec/
2016-05-11 10:49:16 <jirutka> total 56
2016-05-11 10:49:16 <jirutka> drwxr-xr-x    1 root     root            28 May  8 17:13 .
2016-05-11 10:49:16 <jirutka> drwxr-xr-x    1 root     root           116 May  3 19:41 ..
2016-05-11 10:49:16 <jirutka> drwxr-xr-x    1 root     root            48 May  3 19:41 gcc
2016-05-11 10:49:16 <jirutka> -rwxr-xr-x    1 root     root         55336 Oct 29  2015 rmt
2016-05-11 10:49:16 <jirutka> drwxr-xr-x    1 root     root          4168 May  3 19:41 git-core
2016-05-11 10:49:31 <fabled> ncopa, i bet that is it. _GNU_SOURCE is defined in some header that is included in between
2016-05-11 10:49:44 <fabled> the time wchar.h is pulled in, it's not defined.
2016-05-11 10:49:49 <fabled> then gnu_source defined
2016-05-11 10:49:50 <jirutka> ncopa: this is related to the qemu patch
2016-05-11 10:49:58 <fabled> string.h included, but dismissed since wchar.h pulled it in earlier
2016-05-11 10:50:43 <ncopa> fabled, *nod* sounds like it
2016-05-11 10:52:01 <ncopa> wchar.h is included first
2016-05-11 10:52:19 <ncopa> if i move it to under #include "Python.h" it also seems to work
2016-05-11 10:54:03 <fabled> /usr/include/python2.7/pyconfig.h:#ifndef _GNU_SOURCE
2016-05-11 10:54:04 <fabled> /usr/include/python2.7/pyconfig.h:# define _GNU_SOURCE 1
2016-05-11 10:54:04 <fabled> /usr/include/python2.7/pyconfig.h:#define _GNU_SOURCE 1
2016-05-11 10:54:34 <fabled> seems python's pyconfig.h is broke
2016-05-11 10:54:45 <fabled> alternatively the applications needs to include it before any system headers
2016-05-11 10:54:54 <fabled> _GNU_SOURCE needs to be defined before system includes
2016-05-11 10:55:17 <ncopa> yup
2016-05-11 11:08:29 <barthalion> jirutka: some packages use libexec wrongly and you infer from that that we should use it everywhere? nope
2016-05-11 11:09:01 <barthalion> and wrt OS X, I kinda miss the times when we mostly spoke here about technical stuff…
2016-05-11 11:09:30 <ncopa> maybe open new channel?
2016-05-11 11:09:34 <ncopa> #alpine-blabla
2016-05-11 11:09:36 <ncopa> :)
2016-05-11 11:09:47 <jirutka> barthalion: how should I know it?
2016-05-11 11:10:01 <jirutka> ncopa: so the Alpine rule is to not use /usr/libexec?
2016-05-11 11:10:13 <barthalion> well, ncopa already said he doesn't like it
2016-05-11 11:10:17 <ncopa> jirutka, we have used both /usr/libexec and /usr/lib/$pkgname/
2016-05-11 11:10:35 <ncopa> some packages uses /usr/libexec by default
2016-05-11 11:10:37 <barthalion> and grep on aports is your best friend, really
2016-05-11 11:10:50 <ncopa> i didn't say i don't like it. i said im in doubt
2016-05-11 11:11:00 <ncopa> it would be nice to be consequent, yes
2016-05-11 11:11:06 <barthalion> then I'll make your doubt go away, /usr/lib/$pkgname is way to go
2016-05-11 11:11:10 <jirutka> barthalion: I use that… e.g. git uses /usr/libexec… so?
2016-05-11 11:11:17 <ncopa> but i though /usr/libexec was a  no-no
2016-05-11 11:11:30 <barthalion> so that makes two already
2016-05-11 11:11:39 <ncopa> but it appears that it is ok in FHS 3.0
2016-05-11 11:11:46 <jirutka> okay, I have no problem with changing it back to /usr/lib/qemu
2016-05-11 11:11:51 <barthalion> except nobody follows FHS, I guess
2016-05-11 11:11:55 <barthalion> except jirutka 8)
2016-05-11 11:11:59 <ncopa> :)
2016-05-11 11:12:22 <jirutka> I like consistence, so if you say that /usr/libexec is no-no, then we should fix packages that installs into /usr/libexec
2016-05-11 11:12:51 <ncopa> i don't really have strong opinion in either way. posslib slightly biased agains /usr/lib/$pkgname
2016-05-11 11:13:25 <jirutka> agains? so you prefer /usr/libexec?
2016-05-11 11:13:28 <ncopa> i think the /usr/libexec was to be avoided because it doesn't support multi arch
2016-05-11 11:13:40 <ncopa> jirutka, no, the other way
2016-05-11 11:13:43 <barthalion> 98 of /usr/lib/$pkgname vs 34 of /usr/libexec
2016-05-11 11:13:43 <jirutka> okay
2016-05-11 11:13:51 <barthalion> grep as a service
2016-05-11 11:13:58 <barthalion> $3 for these queries I just ran
2016-05-11 11:14:01 <ncopa> yes
2016-05-11 11:14:35 <ncopa> we have preferred /usr/lib/$pkgname, but i have not cared enough to patch the packages that hardcode /usr/libexec
2016-05-11 11:14:45 <jirutka> barthalion: what if these 98 packages are wrong and 34 are correct? “some packages use libexec wrongly and you infer from that that we should use it everywhere? nope”
2016-05-11 11:15:00 <barthalion> what if you don't read what I wrote?
2016-05-11 11:15:03 <barthalion> ncopa is in doubt, I'm not
2016-05-11 11:15:06 <barthalion> case solved
2016-05-11 11:15:35 <ncopa> barthalion, why you prefer /usr/lib/$pkgname over /usr/libexec?
2016-05-11 11:15:35 <barthalion> I don't infer anything from these numbers, I merely use it as an argument to stop this meaningless discussion
2016-05-11 11:16:04 <jirutka> barthalion: you’re self-claimed source of the only truth or what?
2016-05-11 11:16:30 <barthalion> ncopa: no gain for such fragmentation
2016-05-11 11:16:34 <skarnet> can't we just install every package into /opt/$pkgname-version and be done with the FHS altogether? :P
2016-05-11 11:16:46 <jirutka> skarnet: nope :/
2016-05-11 11:16:55 <skarnet> maybe not today
2016-05-11 11:17:01 <barthalion> jirutka: no, I'm one of developers, and I believe it counts
2016-05-11 11:17:08 <barthalion> skarnet: gobolinux to the rescue
2016-05-11 11:17:19 <skarnet> barthalion: that's far from the worst thing they're doing
2016-05-11 11:18:00 <ncopa> jirutka, do you have any good reason why /usr/libexec would be preferred over /usr/lib/$pkgname?
2016-05-11 11:18:46 <jirutka> ncopa: no, I don’t
2016-05-11 11:19:19 <jirutka> skarnet: http://sta.li/ :)
2016-05-11 11:19:23 <ncopa> i say we go for /usr/lib/qemu then
2016-05-11 11:19:38 <ncopa> it has a technical benefit: makes it easier to do multiarch
2016-05-11 11:19:45 <ncopa> in case we�d like to do that in future
2016-05-11 11:19:56 <jirutka> skarnet: I like the idea of statically linked distro
2016-05-11 11:20:28 <ncopa> re /opt/$pkg
2016-05-11 11:20:58 <ncopa> i think the idea is that if you ship in /opt/$pkg/ then all the needed dependencies should be included in that subdir
2016-05-11 11:23:05 <jirutka> ncopa: okay, I’ve updated https://github.com/alpinelinux/aports/pull/73, libexec is gone
2016-05-11 11:23:13 <ncopa> jirutka, thanks!
2016-05-11 11:23:24 <ncopa> barthalion, thanks for your input too
2016-05-11 11:23:33 <barthalion> (y)
2016-05-11 11:24:21 <jirutka> If I remember it corretly, then /opt should be used only for packages not installed by a system package manager.
2016-05-11 11:24:43 <ncopa> thats how i understood it too
2016-05-11 11:25:04 <jirutka> That’s why I’ve used /usr/lib/bundles/$pkgname instead of /opt/$pkgname for Ruby/Python/whatever applications with bundled dependencies of that language
2016-05-11 11:25:08 <jirutka> https://github.com/jirutka/aports#bundles
2016-05-11 11:26:50 <ScrumpyJack> would someone be kind enough to initiate a build of /edge/testing/armhf ? it's been a while, and lots of stuff added since last time
2016-05-11 11:27:36 <skarnet> ncopa: I don't think so, because for instance zlib would be in many subdirectories :)
2016-05-11 11:28:21 <jirutka> barthalion: indeed it counts, I just don’t like your attitude “ncopa is in doubt, I'm not; case solved”… it sounds quite arrogant; but whatever, move on
2016-05-11 11:28:45 <skarnet> the point of /opt installations (and it's an advantage of slashpackage, too) is that it allows you to have several versions of the same software, with one being the default. Also, upgrades are atomic, which is impossible with FHS installations.
2016-05-11 11:29:18 <barthalion> if it would be #alpine-arrogancy, I'd reply it, but it's not
2016-05-11 11:29:49 <jirutka> skarnet: well, that’s not always the case
2016-05-11 11:29:55 <skarnet> jirutka: yes, I know sta.li. I was doing the same thing on my servers before they had the idea. ;)
2016-05-11 11:30:24 <jirutka> skarnet: what do you think about the idea of making Alpine statically linked distro?
2016-05-11 11:30:34 <jirutka> I’ll read the responses later, I have to leave now
2016-05-11 11:31:03 <skarnet> I would love that, but there are objections - for instance, router people are big on saving disk space, and static binaries are generally bigger.
2016-05-11 11:32:09 <skarnet> Also, static linking becomes impractical as soon as you have graphic libraries, which are *really* bloated. You can end up with hundreds of megabytes binaries fairly quickly.
2016-05-11 11:32:51 <skarnet> ISTR MacOS X linked stuff statically at some point, except for the libc or something; the xterm binary was 73 MB.
2016-05-11 11:34:46 <skarnet> so, yeah. Static linking is cool for servers, and I'd love system, low-level software, and servers, to be statically linked. But it's definitely not happening for clients and big apps, and even for servers, I think proponents of static linking are a minority.
2016-05-11 15:07:05 <ncopa> i am changing the build-3-4-x86* builders to halt of first error
2016-05-11 15:16:59 <ncopa> is it just me or is openssl headers messed up? http://build.alpinelinux.org/buildlogs/build-3-4-x86_64/community/py-cryptography/py-cryptography-1.0.2-r0.log
2016-05-11 15:21:52 <fabled> ncopa, py-crypto defines the same name
2016-05-11 15:22:05 <fabled> build/temp.linux-x86_64-2.7/_openssl.c:718:6
2016-05-11 15:22:17 <fabled> i think openssl was fixed to have it 'const'
2016-05-11 15:24:09 <ncopa> bah
2016-05-11 15:37:58 <leo-unglaub> hey :)
2016-05-11 15:38:20 <jirutka> leo: hi
2016-05-11 15:38:54 <jirutka> leo: I’ve trying to compile Rust on Alpine with little cheating, using https://github.com/andyshinn/alpine-pkg-glibc for building stage1
2016-05-11 15:39:29 <jirutka> leo: currently the build failed on http://haste.fit.cvut.cz/raw/uzosulu, do you have some ideas how to fix it?
2016-05-11 15:42:24 <leo-unglaub> puh, sadly no
2016-05-11 15:43:00 <jirutka> I belive that this is not a hard problem, for someone who knows C stuff well
2016-05-11 15:52:44 <skarnet> you need a complete glibc toolchain
2016-05-11 15:52:55 <skarnet> including the shared libgcc compiled with the glibc abi
2016-05-11 15:53:27 <jirutka> so just glibc itself isn’t sufficient? :(
2016-05-11 15:54:02 <skarnet> nope
2016-05-11 15:54:04 <skarnet> it's C++
2016-05-11 15:54:13 <jirutka> that’s bad
2016-05-11 15:54:17 <skarnet> it depends on libgcc
2016-05-11 15:54:53 <skarnet> just grab a glibc-built gcc toolchain, Linaro makes those
2016-05-11 15:55:32 <skarnet> but yes, there's actually a reason why Rust is such a pain to build on Alpine
2016-05-11 15:55:34 <jirutka> it seems that the only option is to build stage1 on glibc system and use it as a snapshot for buidling on Alpine; actually that’s exactly what Rust does, stage0 is downloaded as precompiled binary, but they still doesn’t provide rustc precompiled with musl :(
2016-05-11 15:56:20 <jirutka> > just grab a glibc-built gcc toolchain
2016-05-11 15:56:20 <jirutka> where can I find it?
2016-05-11 15:56:46 <skarnet> as I said, Linaro is one option (look it up).
2016-05-11 15:56:52 <jirutka> hmm, gcc… what if I use clang instead of gcc, can it help?
2016-05-11 15:56:55 <skarnet> I'm sure there are other ones.
2016-05-11 15:57:33 <skarnet> if you use clang, your binaries will depend on the runtime provided by clang.
2016-05-11 15:57:49 <skarnet> and you need to compile libstdc++ with clang, etc.
2016-05-11 15:57:51 <jirutka> ad Linaro, you mean this? http://www.linaro.org/downloads/
2016-05-11 15:58:13 <skarnet> yes.
2016-05-11 15:59:27 <skarnet> or, you can build one yourself - it's gcc, glibc, x86_64 (I assume), native build: all the default options, what gcc people actually test, so it shouldn't bee *too* hard.
2016-05-11 15:59:33 <skarnet> be*
2016-05-11 16:00:22 <jirutka> what exactly do you mean with “build one yourself“? build glibc gcc toolchain?
2016-05-11 16:00:35 <skarnet> Yes. It's not a pleasant experience, but it's very educational.
2016-05-11 16:01:11 <skarnet> And the hard part is fighting the build system's braindeadness and brokenness as soon as you slightly deviate from the default.
2016-05-11 16:01:29 <skarnet> If you're using the default everywhere, which is the case you need, it shouldn't bee *too* hard.
2016-05-11 16:01:45 <skarnet> ... why do I keep writing "bee" ?
2016-05-11 16:02:28 <jirutka> maybe because we have spring, bees are flying all around… :)
2016-05-11 16:02:34 <skarnet> anyway, only do this if you have way too much time on your hands
2016-05-11 16:02:48 <jirutka> that’s exactly what I don’t have :/
2016-05-11 16:03:51 <jirutka> I’ll look at Linaro
2016-05-11 16:03:53 <skarnet> another easy option to get a gcc+glibc toolchain: install your favorite dumb-mainstream distro somewhere, they provide the toolchain you need
2016-05-11 16:03:56 <jirutka> thanks for advice
2016-05-11 16:04:12 <jirutka> and then cross-compile rustc for musl?
2016-05-11 16:05:01 <jirutka> actually, why not, we need to do this just once, then we can use this binary as stage0 for building next stages
2016-05-11 16:05:19 <skarnet> I have no idea what is necessary to compile rustc for musl, I'm unfamiliar with the rustc bootstrap, but you have to start somewhere.
2016-05-11 16:05:38 <jirutka> I wonder why they don’t compile rustc in stage0 statically, so it can be used anywhere
2016-05-11 16:06:30 <jirutka> wait a moment… I don’t have problem running precompiled rustc, this works with precompiled glibc
2016-05-11 16:06:45 <jirutka> the problem is when compiling rustc using this rustc…
2016-05-11 16:06:54 <skarnet> glibc barely supports static binaries, and I'm pretty sure they don't work at all with C++
2016-05-11 16:07:17 <skarnet> that's probably why they don't do that for stage0 :P
2016-05-11 16:07:22 <jirutka> aha
2016-05-11 16:07:43 <skarnet> static C++ binaries work with musl though, I just tested that today
2016-05-11 16:08:07 <skarnet> sure, a Hello World is 700kB (8 MB unstripped), but it works
2016-05-11 16:08:17 <skarnet> gotta love the "lightweight" C++ runtime
2016-05-11 16:08:36 <jirutka> but I still don’t understand the problem, it seems that precompiled rustc works with precompield glibc, I it has succesfully compield part of the rust sources
2016-05-11 16:08:56 <skarnet> what does a ldd on that rustc say ?
2016-05-11 16:10:03 <jirutka> there’s complete log http://haste.fit.cvut.cz/ohewotu.log
2016-05-11 16:10:40 <jirutka> LD_LIBRARY_PATH=/lib:/usr/lib:/usr/glibc-compat/lib ldd x86_64-unknown-linux-musl/stage0/bin/rustc
2016-05-11 16:10:40 <jirutka>         /lib64/ld-linux-x86-64.so.2 (0x29009517000)
2016-05-11 16:10:40 <jirutka>         librt.so.1 => /lib64/ld-linux-x86-64.so.2 (0x29009517000)
2016-05-11 16:10:40 <jirutka>         libpthread.so.0 => /lib64/ld-linux-x86-64.so.2 (0x29009517000)
2016-05-11 16:10:40 <jirutka>         libdl.so.2 => /lib64/ld-linux-x86-64.so.2 (0x29009517000)
2016-05-11 16:10:40 <jirutka>         libm.so.6 => /lib64/ld-linux-x86-64.so.2 (0x29009517000)
2016-05-11 16:10:40 <jirutka>         libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0x290064a5000)
2016-05-11 16:10:41 <jirutka>         libc.so.6 => /lib64/ld-linux-x86-64.so.2 (0x29009517000)
2016-05-11 16:10:41 <jirutka>         ld-linux-x86-64.so.2 => /usr/glibc-compat/lib/ld-linux-x86-64.so.2 (0x29006281000)
2016-05-11 16:10:42 <jirutka> Error relocating x86_64-unknown-linux-musl/stage0/bin/rustc: mallinfo: symbol not found
2016-05-11 16:10:42 <jirutka> Error relocating x86_64-unknown-linux-musl/stage0/bin/rustc: __rawmemchr: symbol not found
2016-05-11 16:10:43 <jirutka> Error relocating x86_64-unknown-linux-musl/stage0/bin/rustc: __register_atfork: symbol not found
2016-05-11 16:11:02 <jirutka> hmm, why it can’t find these symbols?
2016-05-11 16:11:39 <scadu> jirutka: want an account on paste.xinu? :f
2016-05-11 16:11:49 <skarnet> this rustc works by accident
2016-05-11 16:11:58 <skarnet> the libgcc_s is wrong
2016-05-11 16:12:03 <jirutka> scadu: no, why?
2016-05-11 16:13:12 <jirutka> skarnet: because it’s linked against musl and not glibc?
2016-05-11 16:13:16 <skarnet> yes
2016-05-11 16:13:19 <jirutka> aha
2016-05-11 16:13:20 <scadu> jirutka: well, we all love scroll shitty backlog with thousands lines of stacks and other stuff.
2016-05-11 16:13:34 <skarnet> well it's possible that at this low level this makes no difference
2016-05-11 16:13:43 <skarnet> but I wouldn't bet on it
2016-05-11 16:13:45 <jirutka> scadu: you mean these few lines of output from ldd?
2016-05-11 16:14:10 <jirutka> scadu: as you can see just one line above it I’m using http://haste.fit.cvut.cz for longer pastes
2016-05-11 16:15:39 <scadu> cool.
2016-05-11 16:15:39 <jirutka> it probably makes sense, that it “works” for few sources just by accident
2016-05-11 16:16:07 <skarnet> as long as you're not actually calling anything from libgcc_s, it will work
2016-05-11 16:16:43 <skarnet> is the rustc bootstrap written in C or C++?
2016-05-11 16:16:54 <skarnet> (LLVM is C++, right?)
2016-05-11 16:17:12 <jirutka> to be honest, I don’t know
2016-05-11 16:17:37 <jirutka> Rust is written in Rust, but some parts are written in C or C++
2016-05-11 16:17:52 <jirutka> I currently don’t understand how this build process actually work
2016-05-11 16:18:18 <skarnet> that may be the first step if you're actually trying to run that build process :D
2016-05-11 16:20:26 <ncopa> jirutka, apparently you need start with first version of rust, which was implemented in ocaml, then use that to build the second version and then build all 300 releases in between
2016-05-11 16:20:47 <jirutka> ncopa: this would be insane
2016-05-11 16:20:50 <ncopa> it may work with only 7 releases
2016-05-11 16:20:51 <ncopa> yup
2016-05-11 16:21:01 <ncopa> it is insane
2016-05-11 16:21:05 <ncopa> https://internals.rust-lang.org/t/solving-the-bootstrap-problem/3269
2016-05-11 16:21:25 <jirutka> I believe that there’s some simpler option, it shouldn’t be so hard to cross-compile rustc for just another libc, should be?
2016-05-11 16:22:37 <skarnet> Don't you love language developers who take bootstrapping the language exceedingly lightly?
2016-05-11 16:24:41 <ncopa> i think that the sanes option would be a bootstrap rustc implemented in something else, like C, python whatever
2016-05-11 16:24:52 <ncopa> which only purpose would be to build the rust compiler
2016-05-11 16:25:23 <ncopa> urust2c is an interesting idea too
2016-05-11 16:26:37 <skarnet> *every* language should be bootstrappable from C -_-
2016-05-11 16:27:08 <ncopa> that would make things easier yes
2016-05-11 16:27:08 <skarnet> (why C? because if you bootstrap from a higher-level language, at some point there *will* be loops.)
2016-05-11 16:27:10 <jirutka> and how do you bootstrap C? ;)
2016-05-11 16:27:45 <skarnet> people have forgotten how to bootstrap C, unfortunately. You need a gcc to build gcc :/
2016-05-11 16:28:10 <ncopa> we need a C bootstrap compiler in rust!
2016-05-11 16:28:20 <jirutka> ^ +1
2016-05-11 16:28:47 <skarnet> that wouldn't help. A loop doesn't count. :P
2016-05-11 16:28:55 <ncopa> and we need a rust compiler in go
2016-05-11 16:29:03 <jirutka> NOPE!
2016-05-11 16:29:05 <ncopa> rust bootstrap compiler in go
2016-05-11 16:29:13 <ncopa> and a go bootstrap compiler in java
2016-05-11 16:29:20 <ncopa> and a java bootstrap compiler in ocaml
2016-05-11 16:29:27 <ncopa> and an ocaml bootstrap compiler in ada
2016-05-11 16:29:28 <jirutka> and ocaml bootstrap in PHP…?
2016-05-11 16:29:33 <ncopa> and an ada in php
2016-05-11 16:29:36 <ncopa> and php in python
2016-05-11 16:29:38 <jirutka> omfg
2016-05-11 16:29:42 <ncopa> and python in perl
2016-05-11 16:29:51 <jirutka> and perl in brainfuck
2016-05-11 16:29:53 <ncopa> and a perl compiler in awk
2016-05-11 16:29:56 <ncopa> ..in sed
2016-05-11 16:29:58 <jirutka> brainfuck in whitespace
2016-05-11 16:30:15 <ncopa> thats basically what we end up with yes :)
2016-05-11 16:30:25 <skarnet> as long as you root your dependency chain in INTERCAL (implemented directly in machine language) I'm ok with it
2016-05-11 16:30:56 <ScrumpyJack> does an empty http://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/wbar/ mean that it wbar on edge-armhf built correctly?
2016-05-11 16:31:23 <ScrumpyJack> it's dated 18 march
2016-05-11 16:31:42 <ScrumpyJack> yet there is no wbar.apk on any mirror
2016-05-11 16:34:00 <ncopa> i think arm builder choke on something else
2016-05-11 16:34:22 <ScrumpyJack> in fact no mirror has a file in edge/testing/armhf later than dec 2015
2016-05-11 16:35:19 <ScrumpyJack> Dec 16 2015, then nothing
2016-05-11 16:35:57 <ScrumpyJack> dozer's been chocking for a while then ;)
2016-05-11 16:36:46 <ncopa> huh
2016-05-11 16:37:08 <ncopa> probably only testing repo
2016-05-11 16:37:23 <ScrumpyJack> chocking? that sounds delicious. choking even
2016-05-11 16:37:53 <ScrumpyJack> yes, just testing repo
2016-05-11 16:43:27 <przemoc> https://twitter.com/trzmiel/status/689828062855299072  re simple software stack, not bootstrapping in particular
2016-05-11 16:43:45 <przemoc> https://ptrthomas.wordpress.com/2006/06/06/java-call-stack-from-http-upto-jdbc-as-a-picture/  and our beloved java
2016-05-11 16:44:43 <przemoc> (at least it stays within java, I hope...)
2016-05-11 16:45:31 <jirutka> yes, it does and it’s the insane world of enterprise Java :(
2016-05-11 16:45:39 <skarnet> I knew the latter, but not the former. Instant retweet, thanks!
2016-05-11 16:46:11 <skarnet> ... oh wait, I had already retweeted it. So I knew it and had forgotten. :P
2016-05-11 16:46:25 <jirutka> but the worst is that this is one of the better examples, really… there’s not IBM Web Application Shit, EJBs and the whole stack of the hell
2016-05-11 16:46:31 <ncopa> ScrumpyJack, thanks for the unmaintained cleanup patch. i finally figured out how to apply it
2016-05-11 16:46:47 <ncopa> apparently sprunge or something converted dos newlines
2016-05-11 16:46:48 <przemoc> then sorry for bringing up your sclerosis, skarnet :)
2016-05-11 16:46:53 <jirutka> skarnet: what’s your Twitter username?
2016-05-11 16:47:15 <skarnet> przemoc: getting old is heavy on the body *and* the brain!
2016-05-11 16:47:22 <skarnet> jirutka: @laurentbercot
2016-05-11 16:48:15 <ScrumpyJack> ncopa: neat! I'll bug clandmeter for edge/testing/armhf
2016-05-11 16:48:23 <ScrumpyJack> clandmeter: you about? :)
2016-05-11 16:48:58 <przemoc> skarnet: cello player?
2016-05-11 16:49:23 <skarnet> contrabass :) (But I haven't played it for years now.)
2016-05-11 16:49:48 <przemoc> always mixing them up
2016-05-11 16:50:11 <skarnet> the way to tell them apart is the thickness of the strings
2016-05-11 16:50:16 <skarnet> f
2016-05-11 16:50:19 <skarnet> mt
2016-05-11 16:54:54 <chris|> jirutka: good thing 90% if that stack has been obsolete for years
2016-05-11 16:55:17 <jirutka> chris|: ?
2016-05-11 16:55:33 <chris|> enterprise java
2016-05-11 16:55:43 <jirutka> unfortunately not :(
2016-05-11 16:56:19 <przemoc> skarnet: I suggest starting to play it again.  music helps to refocus the mind, even more if you're the one playing.
2016-05-11 16:57:08 <chris|> sounds like you need to change jobs ;)
2016-05-11 16:58:25 <skarnet> przemoc: believe me, I know the virtues of music. It's all a question of prioritization though; I can't devote enough time to an instrument to play it decently, and it frustrates me to have a bass at home and be bad at it, so I stopped.
2016-05-11 17:00:40 <coredumb> ahah
2016-05-11 17:01:16 <coredumb> reminds me my period doing some electronic music on an RMX
2016-05-11 17:01:19 <coredumb> was fun
2016-05-11 17:01:24 <coredumb> was doing music drunk
2016-05-11 17:01:35 <coredumb> already thinking being a badass DJ
2016-05-11 17:01:51 <coredumb> .... look at me now roaming on IRC
2016-05-11 17:03:00 <skarnet> why wouldn't badass DJs roam on IRC
2016-05-11 17:09:17 <chris|> because they are badass
2016-05-11 17:12:18 <przemoc> it's ~2w old news, but read abot it only today.  REAPER is coming to linux, natively.  possibly the only DAW not bloated beyond control and quite cheap - i.e. worth every penny.  haven't tested it yet on linux yet, but there is some rc available.
2016-05-11 19:06:53 <fabled> firefox is crashing on me again :/
2016-05-11 22:22:21 <systWo> hope this is the right place to ask, but what would be an existing acf package I could base a TFTP webui off of for ACF?
2016-05-11 22:22:50 <systWo> tftp server package I'm using is tftp-hpa
2016-05-12 07:49:24 <ScrumpyJack> morning
2016-05-12 08:24:04 <clandmeter> I made some config changes to our nginx reverse proxy, it something related to our infra fails ping me.
2016-05-12 08:25:14 <ScrumpyJack> pardon?
2016-05-12 08:27:15 <ammunta> ScrumpyJack: that means, if everything is on fire, it's clandmeter's fault, I assume
2016-05-12 08:29:12 <mosez> how can i build a golang tool that relies on stringer as a build dependency?
2016-05-12 08:29:20 <mosez> can stringer be part of go-tools?
2016-05-12 08:29:56 <clandmeter> ScrumpyJack: it means i tried to clean up and be smart, which usually leads into issues.
2016-05-12 08:30:02 <mosez> i'm not sure if it makes sense to create a separate package for https://godoc.org/golang.org/x/tools/cmd/stringer
2016-05-12 08:31:12 <mosez> i want to build packages for https://github.com/hashicorp/terraform and https://github.com/cloudfoundry/cli
2016-05-12 08:31:14 <ScrumpyJack> oh, i thought you might have typed in the wrong window :)
2016-05-12 08:33:52 <mosez> clandmeter: you have already built mutliple packages based on go, right?
2016-05-12 08:34:22 <clandmeter> mosez: i tried
2016-05-12 08:34:47 <clandmeter> like caddy in community
2016-05-12 08:34:55 <mosez> i try to find a proper way than symlinking the dir like http://git.alpinelinux.org/cgit/aports/tree/community/godep/APKBUILD because within the makefile of terraform it's crashing :(
2016-05-12 08:35:11 <mosez> yeah, caddy looks nice but i can't create the snapshots :D
2016-05-12 08:35:31 <clandmeter> why not?
2016-05-12 08:35:32 <mosez> and maybe it's related to that why this extended builddir doesnt work
2016-05-12 08:35:43 <mosez> dev.alpinelinux.org:/archive/$pkgname/ :D
2016-05-12 08:35:53 <clandmeter> you can put it anywhere you like.
2016-05-12 08:37:26 <mosez> hum... true
2016-05-12 08:37:42 <mosez> and for the patch to put it into community i can change the url
2016-05-12 08:38:04 <mosez> but first i will create a patch for go/go-tools
2016-05-12 08:49:33 <ammunta> wait
2016-05-12 08:51:34 <ammunta> does go seriously encourage like, generating source code
2016-05-12 08:58:10 <mosez> http://sprunge.us/LhXH is this an acceptable changeset? :)
2016-05-12 08:58:14 <mosez> ammunta: yes
2016-05-12 08:58:36 <mosez> ammunta: https://blog.golang.org/generate
2016-05-12 09:01:53 <ammunta> ... so
2016-05-12 09:02:57 <ammunta> ceterum censeo autem esse go delendum
2016-05-12 09:03:00 <ammunta> I think is how it goes
2016-05-12 09:03:08 <mosez> what?
2016-05-12 09:03:13 <barthalion> mosez: THIS PATCH IS SO TERRIBLE
2016-05-12 09:04:33 <mosez> barthalion: your shift key seems to be broken :P
2016-05-12 09:06:21 <barthalion> hm, for some reason my build env doesn't see go-bootstrap
2016-05-12 09:07:00 <barthalion> nvm
2016-05-12 09:07:08 <barthalion> it's been a long time since I used it
2016-05-12 09:08:09 <mosez> hum... i have subscribed to slpine-aports, but i'm not sure if my patch have been sent oO
2016-05-12 09:08:34 <mosez> Result: 250 from git-email
2016-05-12 09:08:46 <mosez> good to not knowing if that means success :D
2016-05-12 09:09:34 <barthalion> mosez: I'll apply it after it builds
2016-05-12 09:11:00 <mosez> barthalion: cool. can you also confirm if the mail have been sent to the mailinglist?
2016-05-12 09:11:13 <mosez> idk if the archive gets updated in realtime :)
2016-05-12 09:11:47 <barthalion> yep, twice 8)
2016-05-12 09:11:54 <mosez> ops... ok...
2016-05-12 09:12:05 <mosez> so 250 means everything is fine :D
2016-05-12 09:13:59 <mosez> wtf... it failed oO
2016-05-12 09:14:04 <mosez> at least locally
2016-05-12 09:14:27 <mosez> so back to it and fiddling it out :(
2016-05-12 09:16:44 <barthalion> yup
2016-05-12 09:18:15 <mosez> so compiling again because it's so fast on my scaleway server -.-
2016-05-12 09:22:13 <mosez> barthalion: does it fail for you as well? i don't get it why it fails :(
2016-05-12 09:22:39 <mosez> it's part of the tools archive but it looks like it doesnt get built even when it's in the list of tools
2016-05-12 09:23:39 <mosez> # ls src/tools-c887be1b2ebd11663d4bf2fbca508c449172339e/cmd/stringer/
2016-05-12 09:23:40 <mosez> endtoend_test.go  golden_test.go    stringer.go       stringer14.go     testdata          util_test.go
2016-05-12 09:26:15 <barthalion> I'm at work, so I don't really have time to investigate why it failed
2016-05-12 09:27:00 <mosez> sad, ok
2016-05-12 09:27:11 <mosez> who needs to work xD
2016-05-12 09:45:14 <mosez> looks like i have to add a lot of debug output to the build :(
2016-05-12 10:58:44 <mosez> ok i think i found the issue
2016-05-12 10:59:02 <mosez> just building again to verify it
2016-05-12 10:59:35 <mosez> clandmeter: how should i version go tools that don't have tags or releases?
2016-05-12 11:00:29 <clandmeter> use a date
2016-05-12 11:01:00 <mosez> pkgver=2016-05-12?
2016-05-12 11:01:50 <clandmeter> for instance 4.1_git20160315-r1
2016-05-12 11:02:05 <clandmeter> but if there is no official version, you can skip the 4.1
2016-05-12 11:02:25 <ScrumpyJack> clandmeter: can you review #2005
2016-05-12 11:03:36 <mosez> ok, so i will do pkgver=git20160512-r1
2016-05-12 11:05:42 <mosez> just need to ask... on suse we are doing something like 0.0.0+git20160323.27d1c84
2016-05-12 11:13:57 <ScrumpyJack> mosez: go-bootstrap doesn't seem to build at the moment http://build.alpinelinux.org/buildlogs/build-3-4-x86/community/go-bootstrap/go-bootstrap-1.4.3-r1.log
2016-05-12 11:14:36 <ScrumpyJack> fwiw
2016-05-12 11:28:36 <mosez> scrumpyjack: i'm just changing go, not go-bootstrap :)
2016-05-12 11:29:35 <ScrumpyJack> i wonder if the old version of go-bootstrap that go pulls in is broken
2016-05-12 11:30:17 <ncopa> on x86 at least
2016-05-12 11:32:51 <mosez> ah... ncopa is there :D
2016-05-12 11:37:48 <ncopa> mosez: i just came home
2016-05-12 11:38:42 <mosez> and i need a stronger machine for testing alpine packages :(
2016-05-12 11:40:49 <ncopa> definitively helps
2016-05-12 11:40:50 <mosez> looks like i need to launch something locally instead of doing it remote on my mini scaleway machine
2016-05-12 11:43:33 <mosez> compiling go is so damn slooooooooow
2016-05-12 11:55:15 <BastienM> Hello there, still struggling with my custom iso build (https://github.com/BastienM/alpine-server) and I'm stuck with a /sbin/init error when booting up the iso. Does anyone have a clue on that ?
2016-05-12 12:05:10 <ScrumpyJack> mosez: don't scaleway to nice big beefy boxes now?
2016-05-12 12:05:44 <mosez> scrumpyjack: gut the small one with arm
2016-05-12 12:06:08 <ScrumpyJack> ah, me too
2016-05-12 12:06:12 <mosez> so... who wants to merge my latest change for community/go sent a small moment ago? :)
2016-05-12 12:07:29 <ScrumpyJack> mosez: let me know if you find something beefier (hosted or not). fabled uses a wandboard WBQUAD, which I'm contemplating getting
2016-05-12 12:09:39 <clandmeter> somebody killed git.a.o?
2016-05-12 12:10:15 <clandmeter> hmm chrome is trying https...
2016-05-12 12:12:50 <skarnet> video killed the git.a.o. star ?
2016-05-12 12:14:26 <clandmeter> wasnt that something with a radio?
2016-05-12 12:17:45 <skarnet> #ThatsTheJoke
2016-05-12 12:21:33 <ncopa> BastienM: do you have an exact error message?
2016-05-12 12:23:09 <mosez> http://sprunge.us/EMaW yeah...
2016-05-12 12:23:34 <mosez> ncopa: maybe you wanna integrate my packages? :D
2016-05-12 12:23:46 <ncopa> mosez: can you please add it to testing first?
2016-05-12 12:24:16 <mosez> hum.... yeah i can do that
2016-05-12 12:24:23 <ncopa> once you confirmed that the builder actually builds it, and that it actually runs we can move it to community
2016-05-12 12:25:29 <BastienM> ncopa: no errors shows up during the build process
2016-05-12 12:25:36 <mosez> ncopa: http://sprunge.us/Qhfh do i have to send it to the mailinglist with the update again?
2016-05-12 12:26:03 <BastienM> well as far as the make command is concerned
2016-05-12 12:28:23 <ncopa> mosez: nah, i use the sprunge
2016-05-12 12:28:35 <mosez> perfect
2016-05-12 12:29:02 <mosez> i also want to add a package for https://github.com/hashicorp/terraform so this should end in testing as well, right?
2016-05-12 12:29:19 <ncopa> yes
2016-05-12 12:29:28 <mosez> ok, will prepare it for that
2016-05-12 12:29:29 <ncopa> >>> cf-cli: Cleaning temporary build dirs...
2016-05-12 12:29:30 <ncopa>   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
2016-05-12 12:29:30 <ncopa>                                  Dload  Upload   Total   Spent    Left  Speed
2016-05-12 12:29:30 <ncopa>   0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
2016-05-12 12:29:30 <ncopa> curl: (22) The requested URL returned error: 404 Not Found
2016-05-12 12:30:40 <mosez> ah, i have updated the snapshot like it's done for http://git.alpinelinux.org/cgit/aports/tree/community/caddy/APKBUILD
2016-05-12 12:30:56 <mosez> should i keep that on my personal space?
2016-05-12 12:31:14 <mosez> sorry, forgot to mention that :(
2016-05-12 12:31:32 <ncopa> ca we use this tarball? https://github.com/cloudfoundry/cli/archive/v6.18.0.tar.gz
2016-05-12 12:32:29 <mosez> for this package it should work with the symlink workaround used in other go packages, but for terraform that won't work
2016-05-12 12:33:10 <ncopa> http://tpaste.us/2NZl
2016-05-12 12:33:24 <mosez> since the snapshot used by caddy (clandmeter introduced that) tars the entire gopath
2016-05-12 12:33:31 <ncopa> can you test/fix that and create new patch? with commit message: testing/...:
2016-05-12 12:34:10 <mosez> it's more complicated than that
2016-05-12 12:34:49 <mosez> http://git.alpinelinux.org/cgit/aports/tree/community/godep/APKBUILD is an example that works with an github archive, but it requires the symlinking within the builddir
2016-05-12 12:35:05 <jirutka> so for GoShit it’s allowed to bundle dependencies, but for all others it’s not? >_<
2016-05-12 12:35:52 <mosez> but as this approach breaks some go tools like terraform (it's scanning for subpackages) it's resulting in bad behavior because of the src symlink
2016-05-12 12:36:25 <mosez> jirutka: you want to replace the go vendoring?
2016-05-12 12:37:03 <jirutka> mosez: no, I prefer to not use this crap at all ;)
2016-05-12 12:37:24 <mosez> jirutka: so what do you prefer? python? or even node? xD
2016-05-12 12:38:16 <jirutka> Rust, Lua, Python, Ruby, … there are many many options, all of them better
2016-05-12 12:38:36 <mosez> oh... yea... sorry but this is totally crap.
2016-05-12 12:38:51 <mosez> s/this/your trolling/
2016-05-12 12:38:52 <jirutka> but the point is, why it’s allowed to bundle dependencies for some langs and not for some others?
2016-05-12 12:39:23 <jirutka> because some craps doesn’t provide any option? that’s not a good argument imo
2016-05-12 12:40:09 <mosez> packaging python is crap, packaging ruby is crap, packaging node is crap.
2016-05-12 12:40:29 <BastienM> ncopa: if it can help regarding my problem; http://pastebin.com/0xT31J7Q (build log), http://imgur.com/ER0xitp (screenshot when booting the iso with qemu)
2016-05-12 12:40:52 <jirutka> to be clear, I’m not against bundling dependencies at all, it’s the only _sane_ option for any applications other than C/C++, especially for Node/Ruby/Python/…
2016-05-12 12:41:17 <jirutka> but why I can’t push abuild with bundled Python dependencies to the official repo, but Go is okay?
2016-05-12 12:41:32 <mosez> with go vendoring it's simply not possible to keep the deps separate
2016-05-12 12:42:12 <jirutka> with dozen dependencies in lang that itself provides good dependency management tool it’s, well, not impossible, but just totally insane to keep deps strictly separate
2016-05-12 12:42:23 <jirutka> where’s the difference?
2016-05-12 12:42:54 <mosez> but whatever... got to do something.
2016-05-12 12:43:14 <jirutka> you’re out of arguments? ;)
2016-05-12 12:44:25 <mosez> not at all... but not really interested in such discussions. got to build more packages and also got to setup openvpn. that's more important for me ;)
2016-05-12 12:56:42 <ncopa> BastienM: it looks like it failed to extract the packages to the tmpfs root
2016-05-12 12:58:04 <ncopa> possible reasons: it did not find the boot repository, if faild to read the packages from boot repository, it has not the needed apk keys in initramfs
2016-05-12 12:58:16 <ncopa> my guess is that it does not have the keys
2016-05-12 12:59:30 <ncopa> BastienM: can you from the emergency shell do: find /media -name .boot_repository
2016-05-12 12:59:53 <ncopa> it should find a file /media/cdrom/apks/x86_64/.boot_repositories
2016-05-12 13:00:13 <ncopa> if it finds that, then it indicates that if managed to mount the iso and find the boot repo
2016-05-12 13:00:32 <ncopa> then also in emergency shell verify that the pub key is in /etc/apk/keys
2016-05-12 13:00:40 <ncopa> that you used to sign the repo index with
2016-05-12 13:12:40 <BastienM> ncopa: .boot_repository is found but /etc/apk doesn't exists ... I may have excluded it from the blu :/
2016-05-12 13:13:11 <BastienM> lbu*
2016-05-12 13:14:18 <BastienM> well, the pub keys are not excluded ... lbu list does return them
2016-05-12 13:15:52 <ncopa> the keys needs to be in the generated initramfs image
2016-05-12 13:16:03 <ncopa> apk will need it to extract the packages from boot repository
2016-05-12 13:16:17 <ncopa> oh
2016-05-12 13:16:31 <ncopa> we could make it use the keys in apkovl
2016-05-12 13:16:44 <ncopa> but then it means it cannot boot without the apkovl
2016-05-12 13:17:01 <BastienM> I got the alpine keys, but by the wiki a root*.pub must be there too
2016-05-12 13:17:23 <mosez> ncopa: should i name it cf-cli or cloudfoundry-cli?
2016-05-12 13:17:35 <mosez> homebrew named it cf-cli, but the repo is named cloudfoundry-cli
2016-05-12 13:17:40 <ncopa> the key from ~/.abuild/*.pub
2016-05-12 13:17:44 <mosez> cloudfoundry/cli
2016-05-12 13:18:14 <ncopa> mosez: cloudfoundry-cli is clearer i think
2016-05-12 13:18:21 <ncopa> cf-cli does not tell me anything
2016-05-12 13:18:50 <ncopa> BastienM: APK_KEYS        ?= /etc/apk/keys
2016-05-12 13:19:08 <ncopa> i think you need to copy the ~/.abuild/*.pub into /etc/apk/keys
2016-05-12 13:19:30 <ncopa> or to another directory, and run make APK_KEYS=/path/to/keys
2016-05-12 13:21:39 <BastienM> ho ok, I'll try that
2016-05-12 13:27:38 <mosez> ncopa: and can you already integrate my community/go patch? :)
2016-05-12 13:28:35 <ncopa> http://patchwork.alpinelinux.org/project/aports/list/ i see 3 of them?
2016-05-12 13:29:02 <ncopa> which one should i take?
2016-05-12 13:29:24 <mosez> ncopa: the last one... the the others are anyway duplicates because i didn't realized that git send-mail succeded on first run
2016-05-12 13:29:33 <mosez> http://patchwork.alpinelinux.org/patch/2007/
2016-05-12 13:31:18 <mosez> you can reject the 1st and 2nd
2016-05-12 13:34:46 <BastienM> ncopa: all is ok :) it does boot properly now
2016-05-12 13:34:49 <BastienM> thx
2016-05-12 13:34:54 <ncopa> nice!
2016-05-12 13:35:22 <BastienM> gonna write this down, to not forget about the damn keys
2016-05-12 13:43:59 <mosez> ncopa: http://sprunge.us/XdSG what do you think about that?
2016-05-12 13:51:10 <mosez> http://sprunge.us/NAJa damn checksums. now it's fine
2016-05-12 13:51:45 <mitchty> so ghc 8.0.1 got source released, i'm going to update my patch queue for ghc to just use that as i have it building with arm to validate as we speak
2016-05-12 13:52:11 <mitchty> my arm box is slow though so i've got like a day to wait for it
2016-05-12 13:52:42 <mitchty> but a question on names and llvm
2016-05-12 13:53:37 <mitchty> so ghc will be ghc, but ghc depends on realistically singular major versions of llvm, i built a ghc-llvm that just includes llc/opt-$major in /usr/lib/ghc/$program and build ghc against that
2016-05-12 13:54:04 <mitchty> but would it be better if we had old/new(er) llvm's in community that could be referenced maybe?
2016-05-12 13:54:30 <mitchty> that way if the system llvm cannot be used for $reason, the more appropriate one can be selected?
2016-05-12 13:55:32 <mitchty> then last question promise, ghc has 2 build tools (long story), cabal and stack, do I name those haskell-{cabal|stack} or just stack cabal?
2016-05-12 14:00:04 <BastienM> ncopa: got another strange issue, /etc/apk/world doesn't seems to be used on the live session. Once booted the file only contains alpine-base ...
2016-05-12 14:02:33 <BastienM> well, the lbu archive doesn't seems to be loaded at all ...
2016-05-12 14:09:05 <ScrumpyJack> I would like to add a Package update to the issues category, for when Package request doesn't make sense? Any objections? (It's useful for filtering etc)
2016-05-12 14:10:03 <ScrumpyJack> (on redmine)
2016-05-12 15:22:39 <mosez> ncopa: anything bad with my patch for go and my new package? otherwise i would continue with the next one :)
2016-05-12 15:59:37 <przemoc> IIRC ncopa wanted to do some more thorough revamp regarding issue categories, but details have not been laid yet
2016-05-12 16:06:23 <barthalion> folks
2016-05-12 16:06:28 <barthalion> this is real opportunity
2016-05-12 16:06:55 <barthalion> we can have unlimited private repositories on github for just 516$
2016-05-12 16:07:00 <barthalion> /s
2016-05-12 16:07:16 <jirutka> “just”?
2016-05-12 16:07:40 <poptart> shhh no tears barthalion
2016-05-12 16:07:45 <poptart> only dreams
2016-05-12 16:08:04 <barthalion> and that's for 7 people we have listed under organization
2016-05-12 16:08:21 <kaniini> LOL
2016-05-12 16:09:24 <jirutka> well, if you wanna close your code into a safe, then you should pay :P
2016-05-12 16:09:49 <jirutka> more important is that it’s still free for open projects :)
2016-05-12 16:10:35 <poptart> it would be cheaper to buy a dedi and host your own git server
2016-05-12 16:11:06 <barthalion> it would be cheaper to print the code, buy a safe and close it there
2016-05-12 16:11:19 <barthalion> and that would be one time cost, not yearly one
2016-05-12 16:11:20 <jirutka> yeah, but corporates just love to pay a lot of money for nothing :P
2016-05-12 16:11:57 <kaniini> i just consider that a sign that github is swirling the drain and VCs are requiring them to meet targets
2016-05-12 16:12:03 <kaniini> time to move ;)
2016-05-12 16:12:07 <jirutka> well, not exactly nothing, but for shitty and/or useless services; GH is at least good service
2016-05-12 16:12:11 <poptart> they have been hemoraging money for years
2016-05-12 16:12:17 <barthalion> travis is killer feature
2016-05-12 18:18:29 <arnaudsj> Hi, could somebody point me to the kernel sources for alpine? I would like to attempt to compile nvidia kernel extensions for CUDA/GPU
2016-05-12 18:48:50 <poptart> http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec
2016-05-12 19:17:26 <arnaudsj> poptart: thank you!
2016-05-12 19:24:28 <poptart> also thoughts on making a GID for TPE restrictions? I'm getting tired of compiling it in ;)
2016-05-12 19:39:34 <mosez> 16:06 < barthalion> we can have unlimited private repositories on github for just 516$ ... and gitlab sent a newsletter today where thay say they will stay for free forever :D
2016-05-12 19:40:15 <jirutka> free for open repositories, not private ones
2016-05-12 19:41:06 <mosez> https://about.gitlab.com/2016/05/11/git-repository-pricing/
2016-05-12 19:41:17 <mosez> gitlab is free for public and private repos
2016-05-12 19:41:38 <mosez> you just have to pay for better support like next business day or for enterprise edition/on premise
2016-05-12 19:41:49 <jirutka> aha, you’re talking about GitLab
2016-05-12 19:42:34 <mosez> GitLab.com: Free SaaS version featuring private repos, contributors, and CI runners (powered by Digital Ocean).
2016-05-12 19:42:37 <mosez> GitLab Community Edition: Free on-premises version featuring unlimited users and CI.
2016-05-12 19:42:40 <mosez> GitLab Enterprise Edition: $39/user/year offering an on-premises version with additional enterprise-specific features, security, and support.
2016-05-12 19:42:43 <mosez> GitHost.io: Paid depending on the server size you want us to host for you.
2016-05-12 19:42:44 <jirutka> I’m curious to see how they solve their performance problems…
2016-05-12 19:43:29 <mosez> that's a general gitlab issue
2016-05-12 19:43:52 <jirutka> I now GitLab very well, I’m administrating four instances, including faculty instance, but for open-source project, I prefer GitHub
2016-05-12 19:43:56 <mosez> i'm still hosting a ce on my cluster and it's also so damn slow... i will switch to gogs soon
2016-05-12 19:44:31 <jirutka> it’s not slow for hundreds of users, but for public hosted service, it’s slow
2016-05-12 19:44:33 <mosez> or maybe sourcegraph
2016-05-12 19:45:14 <jirutka> try gitlab abuild from https://github.com/jirutka/aports; their OmniBus package is total crap
2016-05-12 19:45:40 <jirutka> use 2 Unicorn works at minimum, 3 or more is better
2016-05-12 19:45:41 <mosez> gitbucket is also promising, even when it's written with scala... but they already received a post from github that they are forced to change their ui :D
2016-05-12 19:45:59 <mosez> i generally dislike this omnibus crap
2016-05-13 07:34:27 <clandmeter> ncopa: i think i found our new arm builder
2016-05-13 07:34:34 <clandmeter> http://b2b.gigabyte.com/products/product-page.aspx?pid=5422
2016-05-13 07:50:50 <scadu> clandmeter: cool, but almost 600 GBP? :c
2016-05-13 07:58:38 <barthalion> clandmeter knows their security system
2016-05-13 07:59:32 <clandmeter> I can try to get a sample
2016-05-13 09:51:13 <ncopa> clandmeter: wow, thats a nice server indeed
2016-05-13 09:52:00 <ncopa> http://b2b.gigabyte.com/products/product-page.aspx?pid=5423
2016-05-13 09:52:56 <ncopa> is that cpu arm64?
2016-05-13 09:53:57 <ncopa> yup
2016-05-13 10:21:42 <ScrumpyJack> ncopa: did you suggest as some point the we should stop using _vars in APKBUILD?
2016-05-13 10:27:07 <clandmeter> ncopa: we could also try to get http://b2b.gigabyte.com/products/product-page.aspx?pid=5460
2016-05-13 10:27:20 <clandmeter> it only has 384 cores :)
2016-05-13 10:28:38 <skarnet> Ì like that baby
2016-05-13 10:30:37 <ncopa> +1
2016-05-13 10:30:48 <ncopa> how do we get one of those?
2016-05-13 10:31:10 <skarnet> the obvious answer is kinda by forking the dough
2016-05-13 10:31:25 <skarnet> but I'm not sure it's reasonable :)
2016-05-13 10:31:55 <clandmeter> you could try to find somebody who works for gigabyte and be nice to him.
2016-05-13 10:32:22 <skarnet> clandmeter with the pimp mentality already :D
2016-05-13 10:33:15 <clandmeter> maybe nice wouldnt do it for such hw. :)
2016-05-13 10:36:12 <ncopa> each node has 48 cores? isnt that 192 cores?
2016-05-13 10:37:19 <coredumb> ncopa: 2x 48 per node
2016-05-13 10:37:23 <ncopa> ah
2016-05-13 10:37:38 <coredumb> how much is that ? and how much power does it eat ?
2016-05-13 10:38:10 <skarnet> "how much is that" is exactly what the site doesn't say...
2016-05-13 10:39:09 <skarnet> woooo 38 kg
2016-05-13 10:48:09 <scadu> >it only has 384 cores :)
2016-05-13 10:48:13 <scadu> i'm in love.
2016-05-13 10:50:52 <ScrumpyJack> are all core accessible to the same bus? or is it like 4 buses, one per tray/node
2016-05-13 10:53:06 <skarnet> one of my "when I'm very rich" fantasies is to have an AC'ed cellar with a huge 'net connection, and my own rack bay with that kind of server in it. :)
2016-05-13 10:54:20 <scadu> #nsfw
2016-05-13 10:55:00 <skarnet> ... no, not *that* kind of fantasies.
2016-05-13 10:58:29 <leo-unglaub> hey :)
2016-05-13 11:00:47 <scadu> skarnet: well, rack with 384 cores could be IT porn :p
2016-05-13 11:01:53 <skarnet> sure, but that kind is totally sfw. If your manager catches you, he'll just sit and watch with you. :)
2016-05-13 11:03:34 <scadu> just kidding.
2016-05-13 11:03:43 <scadu> but you are probably right :>
2016-05-13 12:34:00 <fabled> clandmeter, domoticz fails on armhf still
2016-05-13 12:34:10 <fabled> you seemed to enable it on armhf when moving to community
2016-05-13 12:34:46 <clandmeter> fabled: i was looking at it
2016-05-13 12:34:59 <clandmeter> not sure i can try a local build.
2016-05-13 12:35:29 <fabled> probably missing #include <asm/ioctl.h> but not sure if that causes other issues
2016-05-13 12:35:33 <clandmeter> it was as simple as a missing include
2016-05-13 12:35:46 <clandmeter> right
2016-05-13 12:36:05 <clandmeter> you want me to disable it for now?
2016-05-13 12:36:11 <fabled> unless you want to fix it
2016-05-13 12:36:18 <fabled> otherwise yes
2016-05-13 12:36:29 <clandmeter> not really, i no device currently to support it.
2016-05-13 12:36:48 <fabled> ok disable it for now
2016-05-13 12:38:28 <fabled> ncopa, firefox46.0-r1 (@testing) is crashing on me again at times
2016-05-13 12:38:49 <fabled> gdb sais it's libxul calling some mem/str function
2016-05-13 12:38:54 <fabled> but no debug info
2016-05-13 12:46:30 <clandmeter> do we support inverted arch selection?
2016-05-13 12:52:01 <fabled> mmm... probably not
2016-05-13 12:52:16 <fabled> would be nice to have syntax like "all !armhf"
2016-05-13 12:52:28 <jirutka> we currently support just three platforms, so why you need it? o.O
2016-05-13 12:52:49 <fabled> we've had plans to add armv7, and there was request for couple of others too
2016-05-13 12:52:58 <jirutka> okay
2016-05-13 12:54:24 <fabled> ncopa, wish list for abuild: support ! notation in arch; make mqtt notify say (failed) in status if build failed and it stopped
2016-05-13 13:55:38 <ScrumpyJack> fabled: has buildozer stopped on firefox or is it just taking ages?
2016-05-13 13:55:54 <fabled> ScrumpyJack, on arm? it's just taking ages.
2016-05-13 13:56:22 <fabled> even on 24-cpu x86_64 it takes ten minutes+
2016-05-13 14:07:09 <jirutka> I’m thinking about some academic project that would defent building of ARM cluster. :P
2016-05-13 14:07:40 <jirutka> Like doing science on paper, building Alpine packages in reality. :P
2016-05-13 14:30:05 <skarnet> jirutka: the dream of many a student, and the nightmare of many an academic science department steward. :)
2016-05-13 14:30:12 <clandmeter> nice weekend climbers!
2016-05-13 14:30:43 <skarnet> climbers? because Alpine?
2016-05-13 14:31:50 <jirutka> skarnet: I work at IT department :)
2016-05-13 14:32:25 <skarnet> heh, then you must know how hard it is to get funding!
2016-05-13 14:32:34 <jirutka> yeah, I know…
2016-05-13 15:06:01 <ncopa> fabled: you mean so build.alpinelinux.org will show package failed?
2016-05-13 15:21:06 <ScrumpyJack> ncopa: yes, and also on the mqtt feed. at the moment, the sub just shows the last package being built
2016-05-13 15:22:49 <ScrumpyJack> armhf domoticz failed to build, but the mqtt feed doesn't show that. i just sits there, like this: build/build-edge-armhf 22/23 167/169 community/domoticz 3.4834-r1
2016-05-13 15:23:03 <ncopa> i've seen it
2016-05-13 15:23:03 <ScrumpyJack> (as an example)
2016-05-13 15:23:06 <ncopa> ok
2016-05-13 15:23:10 <ScrumpyJack> does that make sense?
2016-05-13 15:23:15 <ncopa> yes
2016-05-13 15:23:18 <ScrumpyJack> neat :)
2016-05-13 15:23:59 <ncopa> actually
2016-05-13 15:24:03 <ncopa> we post errors
2016-05-13 15:24:06 <ncopa> on mqtt
2016-05-13 15:24:16 <ncopa> to different topic
2016-05-13 15:24:52 <ScrumpyJack> aha, cool
2016-05-13 15:25:07 <ncopa> "build/%/errors"
2016-05-13 15:25:09 <ncopa> iirc
2016-05-13 15:25:58 <ScrumpyJack> how do you sub to more than one?
2016-05-13 15:26:38 <ncopa> add another -t ?
2016-05-13 15:27:55 <ScrumpyJack> hmm, wouldn't build/build-edge-armhf give me build/build-edge-armhf/errors ?
2016-05-13 15:28:23 <ScrumpyJack> maybe build/build-edge-armhf/+
2016-05-13 15:28:58 <fabled> ncopa, or just when error comes and it stops, just make status "community/domoticz 3.4834-r1 (failed)"
2016-05-13 15:29:12 <fabled> so it shows the package where it stopped with failure indication
2016-05-13 15:29:20 <ncopa> yes
2016-05-13 15:30:09 <fabled> i posted http://bugs.alpinelinux.org/issues/5559 about the firefox crash
2016-05-13 15:30:29 <fabled> dunno if we should file it upstream in firefox or brotli
2016-05-13 15:30:36 <fabled> probably ff since it contains embedded brotli
2016-05-13 15:30:49 <ncopa> fabled: i reported it upstream
2016-05-13 15:30:52 <fabled> oh, cool
2016-05-13 15:30:54 <fabled> thx
2016-05-13 15:30:57 <ncopa> they fixed it very quick
2016-05-13 15:31:07 <ncopa> they forwarded it to upstream brotli devs
2016-05-13 15:31:34 <fabled> can you update the ticket with the upstream ref?
2016-05-13 15:31:42 <ncopa> and i think they said they will pull in next upstream brotli to their clone in firefox
2016-05-13 15:36:19 <ScrumpyJack> have a nice weekend
2016-05-13 15:36:45 <fabled> ncopa, i think that's different
2016-05-13 15:37:32 <fabled> maybe the build portable might have a difference
2016-05-13 15:37:35 <fabled> but it's not fortify
2016-05-13 15:37:43 <fabled> this time it's going to real memcpy, and crashing there
2016-05-13 15:37:47 <fabled> segfault
2016-05-13 15:37:53 <fabled> memcpy called with freed memory or similar
2016-05-13 15:43:20 <ncopa> hum
2016-05-13 15:43:21 <ncopa> ok
2016-05-13 15:43:29 <ncopa> then its not reported upstream
2016-05-13 15:43:37 <ncopa> i have not had my ff crash though
2016-05-13 16:04:22 <ncopa> icinga2 provides libyajl.so.2
2016-05-13 16:04:38 <ncopa> which breaks libvirt-glib
2016-05-13 16:06:36 <ncopa> i think we need to force icinga2 to build with system libs
2016-05-13 16:06:42 <ncopa> and make it fail if those are missing
2016-05-13 18:26:07 <tdtrask> systWo: just noticed your question about ACF
2016-05-13 18:26:34 <tdtrask> what are you hoping to have the ACF do? just basic start/stop/status and edit config file?
2016-05-13 18:32:35 <tdtrask> acf-lighttpd is simple example with multiple config files
2016-05-13 18:32:51 <tdtrask> since tftp-hpa doesn't seem to have any config, you can just yank out those portions
2016-05-13 19:27:44 <ncopa> kodi fails to build
2016-05-13 19:27:53 <ncopa> i have a patch for musl platform support
2016-05-13 19:28:21 <ncopa> but it still fails on missing libdcadec
2016-05-13 19:28:27 <ncopa> i wonder if i can move it to unmaintained?
2016-05-13 19:31:45 <systWo> tdtrask, I'm hoping to have it start/stop, show working dir, and list files
2016-05-13 19:35:08 <tdtrask> systWo: then the acf-lighttpd example is good
2016-05-13 19:35:21 <systWo> TY
2016-05-13 19:35:32 <tdtrask> just need to modify the config option to show the working dir instead
2016-05-13 19:36:16 <tdtrask> if you want the option to add files/directories, that will take some extra coding
2016-05-13 19:36:40 <systWo> nah, the root of the tftp dir should be fine
2016-05-13 19:36:57 <tdtrask> k
2016-05-13 19:37:02 <tdtrask> let me know how it goes
2016-05-13 19:37:28 <systWo> will do.  I Just want a project to get my feet wet, and maybe contribute something useful
2016-05-13 19:42:28 <jirutka> ncopa: could you please merge upgrade of jruby? https://github.com/alpinelinux/aports/pull/75.patch
2016-05-13 19:59:49 <barthalion> there you go
2016-05-13 20:00:12 <jirutka> thanks!
2016-05-13 21:06:36 <ncopa> In any case someone wants to continue with kodi-16.1: http://tpaste.us/Aj5n
2016-05-13 21:06:36 <ncopa> http://tpaste.us/2DNX
2016-05-13 21:07:05 <ncopa> it currently fails on missing crossguid
2016-05-13 22:06:37 <mitchty> so, finally built ghc 8.0.1 on arm https://gist.githubusercontent.com/mitchty/f6466f89f6f5fbfd0a25fb3e492194b0/raw/42fc3a1931c5ed58fdbc2e1b96cd1d5d5468c21f/gistfile1.txt
2016-05-13 22:07:00 <mitchty> tonight i'll clean this up a bit and submit a new patch series, technically 8.0.1 isn't out but the source release is
2016-05-14 00:38:37 <mitchty> question for everyone, since the bootstrap can happen on arm and x64 but uses different sources, should the ghc-bootstrap package download both sources and extract them for both arches?
2016-05-14 05:23:09 <coredumb> mitchty: shouldn't it detect arch and extract the correct source ?
2016-05-14 10:24:53 <leo-unglaub> hey :)
2016-05-14 10:24:58 <jirutka1> hi
2016-05-14 10:25:23 <jirutka> hi
2016-05-14 10:43:25 <mitchty> coredumb: it does, but if you ran it on arm/x86_64 you'd get differnt sums https://gist.githubusercontent.com/mitchty/e7b809d4b0c12b8d92bb7704189fd647/raw/40f171eecbd776d25204703be851bd05800c5c0c/gistfile1.txt
2016-05-14 20:29:45 <saintdev> Is there something special I need to do for a suid binary in a package? I'm getting an error when I build the package "Found executable files with SUID bit set:"
2016-05-14 20:37:58 <jirutka1> add `options="suid"`
2016-05-14 20:39:07 <jirutka> like here https://github.com/alpinelinux/aports/blob/master/main/qemu/APKBUILD#L42
2016-05-14 20:40:07 <saintdev> ahh, thanks :)
2016-05-14 20:40:25 <saintdev> missed that in the APKBUILDs I looked at
2016-05-16 07:25:04 <fcolista> ncopa, you around?
2016-05-16 07:25:24 <fcolista> letsencrypt has been renamed from letsencrypt to certbot.
2016-05-16 07:26:12 <fcolista> How should i handle it in order to allow a smooth upgrade and/or downgrade?
2016-05-16 07:27:22 <ncopa> the project/package name?
2016-05-16 07:27:27 <fcolista> yes
2016-05-16 07:27:28 <ncopa> or the binary
2016-05-16 07:27:30 <ncopa> ok
2016-05-16 07:27:30 <fcolista> both
2016-05-16 07:27:43 <ncopa> did we ship letsencrypt to any stabe branch?
2016-05-16 07:27:48 <ncopa> stable*
2016-05-16 07:27:49 <fcolista> community
2016-05-16 07:27:54 <ncopa> v3.3?
2016-05-16 07:27:55 <fcolista> v3.3
2016-05-16 07:27:57 <ncopa> ok
2016-05-16 07:27:59 <fcolista> yes iirc
2016-05-16 07:28:04 <fcolista> gotta check though
2016-05-16 07:28:05 <ncopa> we can not rename it in v3.3
2016-05-16 07:28:15 <ncopa> but we can rename it in v3.4/edge
2016-05-16 07:28:30 <fcolista> it's enough git mv then?
2016-05-16 07:28:41 <fcolista> and of course the pkgname in the apkbuild
2016-05-16 07:28:47 <ncopa> i think so yes
2016-05-16 07:29:01 <ncopa> we might want add a provides="letsencrypt"
2016-05-16 07:29:11 <fcolista> that's what i was thinking
2016-05-16 07:29:27 <ncopa> and possibly a symlink /usr/bin/letsencrypt
2016-05-16 07:29:52 <fcolista> ok
2016-05-16 07:30:19 <skarnet> if someone automatically renamed "letsencrypt" to "certbot" on my machine, I would *freak out*
2016-05-16 07:30:35 <fcolista> +1 skarnet
2016-05-16 07:30:50 <ncopa> we cannot do it in stable branch as said
2016-05-16 07:31:11 <fcolista> but the rename is done according with an upgrade
2016-05-16 07:31:12 <ScrumpyJack> morning
2016-05-16 07:31:15 <fcolista> package upgrade
2016-05-16 07:31:22 <fcolista> so we don't backport that
2016-05-16 07:31:47 <przemoc> what would be the procedure to deprecate old name (letsencrypt) for future relases (like 3.5)?
2016-05-16 07:32:35 <fcolista> we can do /usr/bin/letsencrypt a wrapper for certbot
2016-05-16 07:32:36 <ncopa> first we check if we provide it for v3.3 at all
2016-05-16 07:32:52 <przemoc> maybe instead of symlink there could be simple shell script issueing a warning
2016-05-16 07:33:03 <ncopa> wrapper
2016-05-16 07:33:04 <ncopa> ues
2016-05-16 07:33:15 <ncopa> good idea
2016-05-16 07:33:17 <fcolista> a wrapper that says letencrypt is deprecated
2016-05-16 07:33:28 <fcolista> and from the next version will be removed
2016-05-16 07:33:52 <skarnet> when in doubt, inform users. :)
2016-05-16 07:34:47 <fcolista> letsencrypt is in 3.3/community
2016-05-16 07:35:16 <fcolista> skarnet, yes. At least you can always say "we announced that :)"
2016-05-16 07:36:35 <przemoc> symlinks are nice until they're gone.  that's why they're not the best for such rename compatibility.  (unless your binary does argv[0] check, of course)
2016-05-16 07:38:15 <przemoc> in future apk tools should handle that kind of rename out-of-the-box, generating wrapper by itself
2016-05-16 07:38:59 <przemoc> when requested, that is
2016-05-16 07:46:32 <coredumb> my chromium from community can't load any page.... anyone else sees that ?
2016-05-16 07:47:26 <fcolista> coredumb, yes
2016-05-16 07:47:28 <fcolista> me too
2016-05-16 07:48:31 <coredumb> fcolista: ahhhh
2016-05-16 07:48:35 <coredumb> nice to not be alone
2016-05-16 07:48:42 <fcolista> mainly https
2016-05-16 07:50:03 <coredumb> think mine can't open anything
2016-05-16 07:56:38 <coredumb> ah it load my static web site
2016-05-16 07:57:16 <coredumb> seems more like it fails on js
2016-05-16 08:20:29 <fcolista> ncopa:
2016-05-16 08:20:30 <fcolista> http://sprunge.us/QNHP
2016-05-16 08:20:40 <fcolista> and the wrapper:
2016-05-16 08:20:40 <fcolista> http://sprunge.us/LOAW
2016-05-16 08:20:44 <fcolista> what do you think?
2016-05-16 08:33:30 <ncopa> send the echo to stderr
2016-05-16 08:33:40 <ncopa> echo ... >&2
2016-05-16 08:34:27 <ncopa> otherwise i think it looks good
2016-05-16 08:37:29 <coredumb> I'd love for alpine to provide chromium as a supported browser
2016-05-16 08:37:41 <coredumb> firefox have become a mess these days
2016-05-16 08:46:29 <coredumb> and I think we can agree that firefox is far from having the best results at every year's pwn2own
2016-05-16 09:03:15 <skarnet> "these days"
2016-05-16 09:03:41 <skarnet> I struggle to recall a time when firefox wasn't a mess :P
2016-05-16 09:04:14 <skarnet> to their credit, a browser is hard. Probably the most complex piece of real-life software there is.
2016-05-16 09:17:07 <coredumb> skarnet: well I remember that - apart from the security standpoint - before they started to get these high version number it wasn't so bad
2016-05-16 09:18:03 <ncopa> would be nice if we could upstream the patches we have for both firefox and chromium
2016-05-16 09:18:33 <coredumb> yeah but first chromium would have to work :D
2016-05-16 09:23:32 <ncopa> i think there is an upsteram bug
2016-05-16 09:23:43 <ncopa> that makes build fail with system icu
2016-05-16 09:23:44 <ncopa> or similar
2016-05-16 09:23:56 <ncopa> it also failed with the forked icu
2016-05-16 09:24:37 <ncopa> but
2016-05-16 09:24:39 <ncopa> right now
2016-05-16 09:24:47 <ncopa> all i want is get 3.4-rc1
2016-05-16 09:25:02 <ncopa> we added too much new stuff too late
2016-05-16 09:25:13 <ncopa> so now we have to clean up broken stuff
2016-05-16 09:35:08 <coredumb> kommmmmm7nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnj
2016-05-16 09:35:21 <coredumb> $
2016-05-16 09:35:21 <skarnet> à tes souhaits
2016-05-16 09:35:55 <coredumb> ¨¨
2016-05-16 09:36:32 <coredumb> oooops 1yo kid ...
2016-05-16 09:36:38 <coredumb> merci skarnet ;)
2016-05-16 09:37:24 <skarnet> do you also leave your 1yo kid near open terminals with root shells ?
2016-05-16 09:37:44 <coredumb> yeah
2016-05-16 09:38:08 <skarnet> so *that* explains the outage last week! XD
2016-05-16 09:38:24 <coredumb> lol
2016-05-16 09:38:44 <skarnet> (your kid would still do better than a few "sysadmins" I used to know)
2016-05-16 09:38:45 <coredumb> I like taking risks :D
2016-05-16 09:50:25 <ScrumpyJack> ncopa: it looks like libvirt-glib is failing to build again
2016-05-16 09:50:51 <ScrumpyJack> can you review and push #2009?
2016-05-16 09:51:04 <ScrumpyJack> oops, i mean http://patchwork.alpinelinux.org/patch/2009/
2016-05-16 09:53:34 <coredumb> skarnet: that's what I say to some of my colleagues - I'm a sysadmin - "Yeah my kid could do that"
2016-05-16 09:53:36 <coredumb> :D
2016-05-16 09:59:22 <ncopa> ScrumpyJack: i'm not convinced its the correct fix
2016-05-16 10:02:56 <ScrumpyJack> ok cool
2016-05-16 10:06:49 <ScrumpyJack> ncopa: you mean adding the yajl-dev build dependancy isn't correct?
2016-05-16 10:06:59 <ncopa> yes
2016-05-16 10:08:54 <ncopa> it pulled in an old version ov icinga2
2016-05-16 10:09:24 <ScrumpyJack> ld seems to want that libyajl.so.2, can't find it, and exits. the fixed worked for me :/
2016-05-16 10:09:42 <ncopa> but
2016-05-16 10:10:01 <ncopa> its not libvirt-glib that wans libyajl
2016-05-16 10:10:03 <ncopa> its libvirt
2016-05-16 10:10:42 <ncopa> so the question is why does not libvirt pull in libyajl.so.2?
2016-05-16 10:10:59 <ncopa> the thing is that it acutally did pull in the libyajl.so.2
2016-05-16 10:11:08 <ncopa> but it was provided by icinga2
2016-05-16 10:11:19 <ncopa> and icinga2 installed it in wrong location
2016-05-16 10:11:29 <ncopa> so apk was happy, but linker was not
2016-05-16 10:11:48 <ncopa> the problem was that icinga2 should not ship a clone of libyajl
2016-05-16 10:12:41 <ScrumpyJack> i see
2016-05-16 10:12:56 <ncopa> i fixed that earlier
2016-05-16 10:13:13 <ncopa> but since the build of the repo failed, the old package was not purged
2016-05-16 10:13:20 <ncopa> so the old, bad package was still there
2016-05-16 10:13:34 <ncopa> and libvirt-glib ended up picking that up
2016-05-16 10:13:52 <ncopa> so i manually removed it from builder
2016-05-16 10:14:01 <leo-unglaub> hey :)
2016-05-16 10:14:04 <ncopa> hi
2016-05-16 10:17:22 <ScrumpyJack> ncopa: libvirt seems to pull in (61/172) Installing yajl-dev (2.1.0-r0), where does icinga2 come in?
2016-05-16 10:18:10 <ncopa> icinga2 was fixed some time ago
2016-05-16 10:18:32 <ncopa> http://git.alpinelinux.org/cgit/aports/commit/?id=0db7695c7e366c92f7943bce6d69b7a54a2f1dde
2016-05-16 10:27:45 <ScrumpyJack> what i meant was that I don't see libvirt pulling in icinga2 at build time, but i do see libvirt pulling in yajl-dev
2016-05-16 10:29:41 <ScrumpyJack> you say that: "why does not libvirt pull in libyajl.so.2? it acutally did pull in the libyajl.so.2, but it was provided by icinga2." -I don't see that
2016-05-16 10:30:22 <ScrumpyJack> maybe I'm not looking in the right place - can you enlighten me?
2016-05-16 10:30:33 <ncopa> because its not easy to reproduce anymore
2016-05-16 10:30:56 <ncopa> its only reproduceable if you have an old icinga2 package in some of your repos
2016-05-16 10:31:02 <ncopa> which was what the builder had
2016-05-16 10:31:09 <ncopa> but it no longer has
2016-05-16 10:32:05 <ncopa> the exact explanation:
2016-05-16 10:32:30 <ncopa> icinga2-2.4.7-r1 had a file /usr/lib/icinga/libyajl.so.2 (or similar)
2016-05-16 10:32:53 <ncopa> so the icinga2-2.4.7-r1 got a provides=so:libyajl.so.2
2016-05-16 10:33:15 <ncopa> libvirt has a depends=so:libyajl2.so.2
2016-05-16 10:33:27 <ncopa> apk found out that icinga2 provides that
2016-05-16 10:33:37 <ncopa> so it installed icinga2
2016-05-16 10:33:40 <ncopa> but
2016-05-16 10:33:53 <ncopa> that file was in /usr/lib/icinga/ directory
2016-05-16 10:34:26 <ncopa> so when building libvirt-glib, the linker found that libvirt was broken
2016-05-16 10:34:34 <ncopa> and errored
2016-05-16 10:34:45 <ncopa> i fixed it with: http://git.alpinelinux.org/cgit/aports/commit/?id=0db7695c7e366c92f7943bce6d69b7a54a2f1dde
2016-05-16 10:34:56 <ncopa> which basically make icinga2 build with the system yajl, as it should
2016-05-16 10:35:25 <ncopa> however, the old icinga2-2.4.7-r1 package was still in the index on the builder
2016-05-16 10:35:39 <ncopa> because the builder will not purge the old package til all packages are built
2016-05-16 10:36:17 <ncopa> so apk would find the old, broken icinga2 that still provided the so:libyajl.so.2
2016-05-16 10:36:55 <ncopa> to fix builder, i simply manually deleted the old, broken icinga2
2016-05-16 10:37:15 <ncopa> and apk would instead use libyajl's provides=libyajl.so.2
2016-05-16 10:37:23 <ncopa> which is the correct
2016-05-16 10:42:45 <ScrumpyJack> "its only reproduceable if you have an old icinga2 package in some of your repos" - gotchya
2016-05-16 10:45:44 <ncopa> wow
2016-05-16 10:45:47 <ncopa> builders are done!
2016-05-16 10:45:49 <ncopa> \o/
2016-05-16 10:54:53 <ScrumpyJack> awesome
2016-05-16 11:16:35 <ncopa> leo-unglaub: http://blog.rust-lang.org/2016/05/13/rustup.html
2016-05-16 13:56:30 <fcolista> ncopa, shouldn't we move arch="" packages in "unmaintained"?
2016-05-16 13:57:01 <mitchty> ncopa: hopefully my email reply hits the high points, at work now so won't really be able to reply in email but feel free to ping me here for more info
2016-05-16 13:58:24 <ncopa> mitchty: thanks
2016-05-16 13:58:32 <ncopa> i think i'll wait til after v3.4 release though
2016-05-16 13:58:46 <ncopa> what i woudl have hoped for is to make it work with system llvm
2016-05-16 13:58:52 <ncopa> so we dont need a second llvm
2016-05-16 13:59:03 <mitchty> note, i'm only compiling llc/opt
2016-05-16 13:59:06 <mitchty> not all of llvm
2016-05-16 13:59:29 <ncopa> llvm is big and painful enough so i would really prefer not double the trouble
2016-05-16 13:59:46 <ncopa> i will try look at it after v3.4 release
2016-05-16 14:00:01 <mitchty> the greater issue is llvm isn't a very stable target at the ir level
2016-05-16 14:00:18 <mitchty> the setup i have is what ghc itself will be doing
2016-05-16 14:00:27 <mitchty> for other platforms
2016-05-16 14:00:46 <mitchty> mostly because of trying to use system llvm and supporting system llvm problems in the past
2016-05-16 14:01:37 <mitchty> and if we wanted 7.10.3 we'd need to have 3.5.2 somehow in alpine linux 3.4
2016-05-16 14:01:50 <mitchty> as an example of future pain that will happen
2016-05-16 14:06:08 <ncopa> :-/
2016-05-16 14:06:18 <mitchty> https://ghc.haskell.org/trac/ghc/wiki/ImprovedLLVMBackend please read through that to understand my port is mostly embodying what the ghc folks are going to do on other platforms
2016-05-16 14:06:29 <ncopa> ok
2016-05-16 14:06:37 <mitchty> honestly if the goal is ghc only with system llvm, i'll just keep this port off distribution
2016-05-16 14:07:07 <mitchty> i've spent a year fighting with llvm and ghc and i'd rather not spend that much more time
2016-05-16 14:08:23 <mitchty> getting the ghc-bootstrap process to work like go for 7.10.3/8.0 already puts me in the "only user of ghc regularly using it" category according to the ghc devs
2016-05-16 14:08:25 <ncopa> understand
2016-05-16 14:09:25 <ncopa> hm
2016-05-16 14:09:37 <ncopa> also, this is what others do too, i suppose
2016-05-16 14:09:51 <ncopa> the stick a given clang version against specific llvm
2016-05-16 14:09:57 <mitchty> not trying to be stubborn about it, but i'd rather not dig around llvm codegen again
2016-05-16 14:10:14 <ncopa> its ok
2016-05-16 15:10:27 <ScrumpyJack> ncopa: what was http://patchwork.alpinelinux.org/patch/2005/ superseded by?
2016-05-16 15:11:06 <ncopa> bb01d9bfb6fccfb03a297176ed88dcb4972b353a
2016-05-16 15:11:30 <ncopa> basically same thing
2016-05-16 15:12:05 <ncopa> just excluding commit message which say we maybe will do this or that in future
2016-05-16 15:12:46 <ScrumpyJack> ah ok, cool.
2016-05-16 15:15:12 <ScrumpyJack> if you're in the mood, could you review patch 1902 to 1906 ?
2016-05-16 15:17:17 <ncopa> why is it not called libftdi1 ?
2016-05-16 15:17:22 <ncopa> thats what the tarball call it
2016-05-16 15:17:38 <ncopa> what the tarball is called i mean
2016-05-16 15:21:36 <ncopa> and i would prefer a patch over that sed line
2016-05-16 16:51:00 <ncopa> i finally applied the dovecot stuff
2016-05-16 16:51:21 <ncopa> we should do the testing/dovecot-antispam- thingy too
2016-05-16 17:20:58 <BitL0G1c> is a problem with lxc@edge - "Error relocating /usr/lib/liblxc.so.1: utmpxname: symbol not found" (when trying to use lxc-console) - v.1.1.5-r2 is ok
2016-05-16 17:28:27 <ncopa> you mix edge && v3.x?
2016-05-16 17:49:47 <BitL0G1c> I've been using lxc@edge for about a year
2016-05-16 17:50:32 <BitL0G1c> am running 3.3.3
2016-05-16 17:52:51 <ncopa> only thing for us to fix it for you is revert some recent musl update
2016-05-16 17:53:27 <ncopa> v3.3 has lxc-1.1.5
2016-05-16 17:53:44 <ncopa> why do you need it from edge?
2016-05-16 17:54:01 <BitL0G1c> 1.1.5-r2 is ok - it's r5 that has problems
2016-05-16 17:54:12 <BitL0G1c> it's ok - i downgraded to r2
2016-05-16 17:54:37 <ncopa> cd ..
2016-05-17 06:36:16 <fcolista> rnalrd, there's a reason why bacula is still to v7.0.5 ?
2016-05-17 06:36:56 <fcolista> i've a working APKBUILD for bacula 7.4.0
2016-05-17 06:37:01 <fcolista> it needs a patch to build
2016-05-17 06:37:40 <fcolista> http://sprunge.us/GSfV <--- APKBUILD
2016-05-17 06:37:52 <fcolista> http://sprunge.us/GSfV <--- patch
2016-05-17 06:38:32 <fcolista> http://sprunge.us/ZOhU
2016-05-17 06:38:32 <fcolista>  <-- patch (sorry, wrong paste before)
2016-05-17 07:32:18 <scadu> fcolista: ping
2016-05-17 07:33:13 <fcolista> echo reply
2016-05-17 07:45:29 <ScrumpyJack> SYN ACK
2016-05-17 09:50:39 <fcolista> rnalrd, probably i miss your comment since i was disconneted
2016-05-17 09:50:47 <fcolista> What do you think about this: http://sprunge.us/Njgb ?
2016-05-17 09:51:38 <rnalrd> k
2016-05-17 10:36:04 <vitronic> ncopa, hi
2016-05-17 10:36:08 <vitronic> ncopa, there any easy way to download the full package repository?
2016-05-17 10:36:44 <vitronic> you me know please
2016-05-17 10:42:58 <leo-unglaub> hey friends :)
2016-05-17 10:43:47 <leo-unglaub> a company called me they cannot open there outlook anymore, it hangs on startup because there mailbox is to big. i asked what adress it is and they told me noreply@ ....
2016-05-17 10:44:11 <leo-unglaub> they send out a newsletter with all adresses they cound find and then got 1,4 million "Undelivered Mail Returned to Sender" mails back *g*
2016-05-17 10:44:29 <leo-unglaub> thats the kind of lulz an admin likes to wake up to *g*
2016-05-17 10:44:46 <leo-unglaub> but thanks to alpine the mail server did not complain about the 1,4 million mails in one night *g*
2016-05-17 11:12:59 <ncopa> vitronic: rsync
2016-05-17 11:13:29 <vitronic> ncopa, rsync over http ??
2016-05-17 11:14:11 <ncopa> rsync over rsync protocol
2016-05-17 11:14:45 <ncopa> rsync://rsync.alpinelinux.org/alpine/v3.3
2016-05-17 11:15:02 <vitronic> :-0
2016-05-17 11:15:30 <vitronic> excellent thank you very much
2016-05-17 11:39:58 <clandmeter> vitronic: there is also a faster rsync mirror for init sync at leaseweb.
2016-05-17 12:03:01 <Guest_85755> Allah is doing
2016-05-17 12:04:03 <Guest_85755> sun is not doing Allah is doing
2016-05-17 12:04:13 <Guest_85755> moon is not doing Allah is doing
2016-05-17 12:04:25 <Guest_85755> stars are not doing Allah is doing
2016-05-17 12:05:14 <Guest_85755> planets are not doing Allah is doing
2016-05-17 12:05:30 <Guest_85755> galaxies are not doing Allah is doing
2016-05-17 12:06:15 <Guest_85755> oceans are not doing Allah is doing
2016-05-17 12:06:32 <Guest_85755> mountains are not doing Allah is doing
2016-05-17 12:06:42 <Guest_85755> trees are not doing Allah is doing
2016-05-17 12:06:55 <Guest_85755> mom is not doing Allah is doing
2016-05-17 12:07:06 <Guest_85755> dad is not doing Allah is doing
2016-05-17 12:07:17 <Guest_85755> boss is not doing Allah is doing
2016-05-17 12:07:27 <Guest_85755> job is not doing Allah is doing
2016-05-17 12:07:39 <Guest_85755> dollar is not doing Allah is doing
2016-05-17 12:07:51 <Guest_85755> degree is not doing Allah is doing
2016-05-17 12:08:04 <Guest_85755> medicine is not doing Allah is doing
2016-05-17 12:08:04 <nmeum> …
2016-05-17 12:08:50 <Guest_85755> customers are not doing Allah is doing
2016-05-17 12:09:07 <leo-unglaub> :)
2016-05-17 12:09:14 <Guest_85755> you can not get a job without the permission of allah
2016-05-17 12:09:34 <nmeum> clandmeter kaniini ncopa rnalrd: *highlight*
2016-05-17 12:10:06 <clandmeter> i forgot how to ban :)
2016-05-18 10:54:50 <leo-unglaub> ncopa: there is currently a huge discussion about alpine-linux-docs on the OpenBSD mailinglist :)
2016-05-18 10:57:00 <leo-unglaub> http://thread.gmane.org/gmane.os.openbsd.misc/231014
2016-05-18 11:00:09 <leo-unglaub> someone defendet alpine there :)
2016-05-18 11:00:25 <leo-unglaub> and even Theo de Raadt got involved himself *g*
2016-05-18 11:02:07 <rnalrd> :-o
2016-05-18 11:03:42 <leo-unglaub> i think what they dont understand is that alpine is not responsible for bad man pages coming from upstream
2016-05-18 11:05:16 <fabled> yes, BSD works different
2016-05-18 11:05:26 <fabled> BSD is a fork/clone of everything, kernel+userland+docs
2016-05-18 11:12:17 <leo-unglaub> brb reboot
2016-05-18 12:11:26 <ncopa> :)
2016-05-18 12:11:28 <ncopa> http://thread.gmane.org/gmane.os.openbsd.misc/231014
2016-05-18 12:12:12 <ncopa> i agree that we can improve docs
2016-05-18 12:14:07 <kunkku> our documentation is a bunch of random and often outdated wiki pages
2016-05-18 12:14:38 <kunkku> the reason: writing doc is boring and no one is paying for it
2016-05-18 12:18:24 <ncopa> yes
2016-05-18 12:18:45 <ncopa> not only boring
2016-05-18 12:18:48 <ncopa> its also difficult
2016-05-18 12:19:25 <leo-unglaub> well, yeah but the comparisn is still not that fair
2016-05-18 12:19:39 <leo-unglaub> because they have there own docs and alpine ships the docs it gets from the vendor
2016-05-18 12:22:52 <kunkku> what do you think of the other claim about us being notably unresponsive to user questions?
2016-05-18 12:23:16 <leo-unglaub> it happens
2016-05-18 12:23:34 <leo-unglaub> its irc, if he asks at a time on no one is around its possible
2016-05-18 12:23:46 <leo-unglaub> or its a question that is simply not answerable
2016-05-18 12:24:05 <leo-unglaub> if he asks: "why is my firmware x not working" there is not much alpine can do about that
2016-05-18 12:24:07 <leo-unglaub> its a kernel issue
2016-05-18 12:25:04 <skarnet> why are you guys fretting over that thread? it's just OpenBSD people circlejerking and saying bad things about Linux, that's how they get their rocks off.
2016-05-18 12:25:20 <skarnet> Just a normal day among OpenBSD people. Nothing to see, move along.
2016-05-18 12:26:34 <leo-unglaub> hmm, i disagree
2016-05-18 12:26:54 <leo-unglaub> most complains from openbsd users have a real background ... but yes, sometimes there tone is of
2016-05-18 12:26:59 <ncopa> i think we can and should improve the alpine docs
2016-05-18 12:27:08 <ncopa> i dont know how though
2016-05-18 12:27:55 <coredumb> usually it's as "simple" as hiring a documentation writter
2016-05-18 12:28:01 <coredumb> :)
2016-05-18 12:28:11 <coredumb> got a couple of indians here for that :D
2016-05-18 12:28:19 <ncopa> i suspect its more difficult than that
2016-05-18 12:28:26 <coredumb> not sure I can make them work for Alpine though
2016-05-18 12:28:40 <ncopa> i mean, i want a doc writer genious :)
2016-05-18 12:28:48 <coredumb> I hear you
2016-05-18 12:29:20 <coredumb> so far I've found the alpine's doc quite good
2016-05-18 12:31:00 <leo-unglaub> yeah, the alpine docs are good. not one bit of rust on them *g*
2016-05-18 12:31:08 <ncopa> lol
2016-05-18 12:31:13 <jirutka> XD
2016-05-18 12:31:29 <barthalion> did you check rustup?
2016-05-18 12:31:49 <jirutka> rustup doesn’t solve our problem
2016-05-18 12:32:11 <jirutka> it’s all about cross-compiling packages, but not rustc itself
2016-05-18 12:32:47 <barthalion> looks like I missed this discussion
2016-05-18 12:33:14 <coredumb> ncopa: is there any plan to have friendlier lvm/raid/crypt setup for alpine's install?
2016-05-18 12:33:48 <ncopa> coredumb: no specific plan atm
2016-05-18 12:34:32 <coredumb> I'd be interested in that
2016-05-18 12:35:30 <leo-unglaub> coredumb: its a pain in the ass to get raid,crypto setups running in alpine ... there are some scripts from BitL0G1c for that that work most of the time
2016-05-18 12:35:34 <leo-unglaub> maybe worth having a look
2016-05-18 12:36:23 <coredumb> yeah it's PITA
2016-05-18 12:36:46 <coredumb> actually something to set up LVM correctly and a choice in FS would be great
2016-05-18 12:37:08 <coredumb> maybe I could find time to write one
2016-05-18 12:37:37 <leo-unglaub> coredumb: look at the BitL0G1c one
2016-05-18 12:37:46 <ncopa> i dont think adding support for cryptsetup would be that difficult
2016-05-18 12:37:59 <ncopa> probably easier that installing it manually :)
2016-05-18 12:38:18 <ncopa> btw, i noticed that installing on 3 disks is broke
2016-05-18 12:38:48 <BitL0G1c> cryptsetup & partition scripts https://github.com/itoffshore/alpine-linux-scripts
2016-05-18 12:39:05 <ncopa> i think current setup-disk does not add raid5 kernel module or similar
2016-05-18 12:39:16 <coredumb> thx BitL0G1c
2016-05-18 12:39:20 <leo-unglaub> https://it-offshore.co.uk/linux/alpine-linux/25-alpine-linux-luks-encrypted-installations
2016-05-18 12:39:36 <BitL0G1c> ;-)
2016-05-18 12:40:22 <coredumb> damn you even put a video
2016-05-18 12:40:29 <coredumb> that's so neat
2016-05-18 12:41:35 <leo-unglaub> coredumb: if you setup alpine so often as i have you can start to type BitL0G1c's website blind ....
2016-05-18 12:41:56 <leo-unglaub> at first the fucking .co. throw me off but now i even got that one *g*
2016-05-18 12:42:39 <leo-unglaub> coredumb: but one hint ... quit often you have to restart the computer after doing the partitions because you will get an error because it does not re-read the diskstructure again
2016-05-18 12:42:43 <leo-unglaub> so remember to reboot it
2016-05-18 12:43:14 <ncopa> i would also like support for efi
2016-05-18 12:45:49 <leo-unglaub> hmm, while we are add it .. i would love to have the OpenBSD bootloader in Alpine as a default because its the only working bootloader out there that also allows a encrypted /boot :)
2016-05-18 12:46:59 <coredumb> leo-unglaub: kpartx to make reread partitions
2016-05-18 12:47:25 <coredumb> so now that I find this it's obvious I have nothing to do
2016-05-18 12:47:28 <coredumb> thx BitL0G1c
2016-05-18 12:47:29 <coredumb> :)
2016-05-18 13:02:34 <jirutka1> huh, abuild segfaulted on travis: https://travis-ci.org/alpinelinux/aports/builds/131126367#L390
2016-05-18 13:07:37 <jirutka> ncopa: could you please merge https://github.com/alpinelinux/aports/pull/77 ? the upgrade has been requested by user via pkgs.a.o and since there are some ppl who use this package, I think that we can move it to the community repo
2016-05-18 13:08:05 <jirutka> and also https://github.com/alpinelinux/aports/pull/78.patch, it’s just upgrade of ruby-bundler
2016-05-18 13:09:15 <ncopa> ok
2016-05-18 13:09:49 <ncopa> i wonder if we can upgrade ruby too while at it
2016-05-18 13:10:08 <ncopa> maybe rename it to ruby2.2 and add ruby2.3=
2016-05-18 13:10:11 <ncopa> ?
2016-05-18 13:10:48 <ncopa> it is technically too late
2016-05-18 13:28:15 <jirutka> well, we should definitely upgrade ruby, but it’s de facto a major version bump, so I think that it will require some testing
2016-05-18 13:28:33 <jirutka> I’ll try to look at it today or maybe tomorrow
2016-05-18 13:29:21 <ncopa> problem is that we already built various things with the old ruby
2016-05-18 13:29:37 <barthalion> ncopa: didn't you complain yesterday that too many things were pushed close to freeze and now much of this is broken?
2016-05-18 13:30:24 <ncopa> i did, which is why i dont think we can uprade ruby
2016-05-18 13:30:39 <ncopa> but i think we might geet away with adding ruby2.3
2016-05-18 13:31:16 <ncopa> problem is also that if we dont upgrade ruby-2.2 to 2.3 then we will have to support ruby 2.2 2 years
2016-05-18 13:31:24 <ncopa> which i doubt that we want
2016-05-18 13:32:29 <ncopa> http://tpaste.us/GmKj
2016-05-18 13:32:41 <ncopa> *if* we upgrade ruby, those will need rebuild
2016-05-18 13:33:08 <ncopa> webkit is a pain
2016-05-18 13:35:36 <fcolista> and you are not sure that those will works
2016-05-18 13:35:45 <ncopa> exactly
2016-05-18 13:35:54 <fcolista> that's a pain
2016-05-18 13:36:00 <fcolista> ruby it's a pain to maintain
2016-05-18 13:36:52 <ncopa> ideally, we should: upgrade ruby to 2.3, add ruby2.2 (for backwards compat), rebuild the list with ruby-2.3, verify that things works, and then push it all in one go
2016-05-18 13:37:03 <fcolista> yes
2016-05-18 13:37:15 <jirutka> I can take care of asciidoctor (I’m contributor of Asciidoctor project) and ruby-bundler (I’m already listed as a maintainer of this)
2016-05-18 13:37:17 <ncopa> its one days work approx
2016-05-18 13:37:29 <jirutka> how abuild depends on ruby?
2016-05-18 13:37:31 <fcolista> i can work on xapian-bindings
2016-05-18 13:37:46 <ncopa> jirutka: we have a tool that would check the ruby-* package deps
2016-05-18 13:37:51 <ncopa> but now we purged those
2016-05-18 13:37:58 <ncopa> so we dont really need it anymore
2016-05-18 13:42:23 <ncopa> whee!
2016-05-18 13:42:26 <ncopa> raid5 works now
2016-05-18 13:42:40 <ncopa> and i also tested using xfs as the rootfs
2016-05-18 13:42:45 <ncopa> it works!
2016-05-18 13:43:04 <ncopa> before starting setup-alpine, simply: export ROOTFS=xfs
2016-05-18 15:23:57 <coredumb> ncopa: cool!
2016-05-18 15:30:12 <fcolista> what's the purpose of "mask" option in APKBUILD file?
2016-05-18 16:05:22 <ScrumpyJack> fabled: when are we likely to see the armhf build in the repo?
2016-05-18 16:05:41 <fabled> ScrumpyJack, what build?
2016-05-18 16:06:13 <ScrumpyJack> sorry, builds (with an s). when will we see them in the package repos?
2016-05-18 16:07:13 <fabled> the packages should be in rsync master
2016-05-18 16:07:20 <ScrumpyJack> hmm
2016-05-18 16:07:54 <ScrumpyJack> LeaseWeb mirror service?
2016-05-18 16:08:40 <ScrumpyJack> clandmeter: suggested we pull from rsync://mirror.ams1.nl.leaseweb.net now. I guess it hasn't been synced to there yet
2016-05-18 16:27:07 <ScrumpyJack> how can i make libfetch ignore timestamps in the package repos?
2016-05-18 16:28:32 <ScrumpyJack> libfetch sends If-Modified-Since: Wed, 18 May
2016-05-18 16:29:02 <ScrumpyJack> can i stop that? (i want my updates now! ;)
2016-05-18 17:25:48 <ScrumpyJack> there is a alpine/v3.3/armhf directory. a mistake?
2016-05-18 17:27:23 <jkwright> I asked in #alpine-linux but someone in here might know better: Has anyone had any luck getting Bluetooth working on the RPi3? I spent about five hours fighting with it last night and have made very little progress.
2016-05-18 17:28:33 <ScrumpyJack> i'm using an RPi3 here, but no bt device to connect
2016-05-18 17:31:38 <jkwright> The service starts but any time I ever have to interact with the BT stack, for example "power on" in bluetoothctl, I get a message along the lines of "No default controller available". I've pretty much exhausted Google with it. I know that Raspbian has some platform-specific BT tools but I don't know much about what they do.
2016-05-18 17:32:15 <jkwright> Information specific to Alpine on the Pi is pretty scarce
2016-05-18 17:32:22 <jkwright> Especially the Pi3
2016-05-18 21:22:51 <coredumb> there seems to have an issue with deluge dep in testing
2016-05-18 21:23:12 <coredumb> ERROR: unsatisfiable constraints:
2016-05-18 21:23:14 <coredumb>   so:libboost_python.so.1.58.0 (missing):
2016-05-18 21:23:16 <coredumb>     required by: libtorrent-rasterbar-0.16.19-r2[so:libboost_python.so.1.58.0]
2016-05-18 21:23:18 <coredumb>                  libtorrent-rasterbar-0.16.19-r2[so:libboost_python.so.1.58.0]
2016-05-19 07:17:39 <clandmeter> ScrumpyJack: i suggest to use leaseweb for init sync because its faster.
2016-05-19 07:18:05 <clandmeter> then later use another which syncs more frequent
2016-05-19 07:42:29 <ScrumpyJack> still not able to upgrade edge 12 hours later. wish i remembered how to override If-Modified-Since:
2016-05-19 07:42:46 <ScrumpyJack> ncopa mentioned it once, but i didnt write it down
2016-05-19 07:46:39 <clandmeter> ScrumpyJack: ?
2016-05-19 07:47:14 <fabled> ScrumpyJack, apk update --force
2016-05-19 07:47:36 <ScrumpyJack> i tried that, nothing. i'll try again
2016-05-19 07:48:26 <ScrumpyJack> how does libfetch set "If-Modified-Since:
2016-05-19 07:48:50 <ScrumpyJack> it's now "If-Modified-Since: Tue, 14 Jun 2"
2016-05-19 07:50:37 <ScrumpyJack> i messed with my clock and ran apk update
2016-05-19 07:50:57 <ScrumpyJack> clock is fine now, but "If-Modified-Since: still sets old time
2016-05-19 07:56:19 <ScrumpyJack> ok, "If-Modified-Since:" back to today
2016-05-19 07:56:46 <Sachiru> Is there any way to check what dependencies are there for each APK? I'd like to install Redmine but not use lighttpd
2016-05-19 07:56:51 <Sachiru> Rather use Unicorn and nginx
2016-05-19 07:57:35 <clandmeter> Sachiru: we suggest no to use apk for redmine anymore, instead use bundler.
2016-05-19 07:58:04 <Sachiru> So... Just install a standard ruby+rails stack, then bundler install for redmine?
2016-05-19 07:59:01 <clandmeter> Sachiru: you can follow the redmine wiki.
2016-05-19 07:59:09 <clandmeter> currently we use puma on ours
2016-05-19 07:59:18 <clandmeter> before we used unicorn
2016-05-19 07:59:50 <Sachiru> Thanks!
2016-05-19 07:59:57 <clandmeter> Sachiru: make sure you have installed build-base and some headers needed for some ruby pkgs.
2016-05-19 08:00:15 <clandmeter> i dont know exactly which one you need.
2016-05-19 08:00:46 <Sachiru> I see. Will do, thanks. So the idea is to run redmine using Unicorn, with nginx as proxy. Is that feasible?
2016-05-19 08:01:42 <clandmeter> sure
2016-05-19 08:02:01 <clandmeter> we use nginx to proxy all containers
2016-05-19 08:02:16 <clandmeter> atleast my part of infra we do
2016-05-19 08:06:36 <Sachiru> Thanks.
2016-05-19 08:06:55 <Sachiru> Following this procedure to get a minimal image: http://blog.kontena.io/building-minimal-docker-image-for-rails/
2016-05-19 08:07:16 <Sachiru> With the exception that I am replacing postgresql-dev with mariadb-dev
2016-05-19 08:07:38 <clandmeter> right
2016-05-19 08:07:49 <clandmeter> im not a docker user, so i cannot help you in that direction
2016-05-19 08:07:54 <clandmeter> but it looks sane
2016-05-19 08:17:35 <ScrumpyJack> there is a typo in a pkgname var in an APKBUILD. if i fix the typo, should i bump the pkgrel?
2016-05-19 08:18:46 <ScrumpyJack> are there issues in renaming a package?
2016-05-19 08:19:11 <ScrumpyJack> (what happens to the old apk in the repos)
2016-05-19 08:29:55 <clandmeter> ScrumpyJack: check if anything depends on it
2016-05-19 08:45:28 <ScrumpyJack> nothing does, no
2016-05-19 08:47:05 <ScrumpyJack> it's py-unicornhat - the typo in the APKBUILD file is pkgname=py-unicorni-hat (probably a vi insert slip up)
2016-05-19 08:49:21 <ScrumpyJack> also, we have two armhf python raspberry pi gpio libs: py-rpi-gpio and py-rpigpio
2016-05-19 09:10:02 <leo-unglaub> hey :)
2016-05-19 09:10:10 <leo-unglaub> is someone of you a hardware expert?
2016-05-19 09:10:20 <leo-unglaub> i have some hardware trouble i cannot figure out
2016-05-19 09:15:24 <clandmeter> 
ACTION hides

2016-05-19 09:17:00 <leo-unglaub> last night it got my new http://www.bequiet.com/en/powersupply/477
2016-05-19 09:17:34 <leo-unglaub> i put it in my computer like 100 times before and now it does not start anymore :(
2016-05-19 09:17:37 <leo-unglaub> but i have no idea why
2016-05-19 09:17:59 <leo-unglaub> i measured every power source i could find, the powersuply holds the corrcet power
2016-05-19 09:18:26 <clandmeter> you do not have another psu?
2016-05-19 09:18:41 <leo-unglaub> my pc starts for one secound, then shutdowns for 3 secounds and then starts again but i never see something on my monitor and usb devices have no power as well
2016-05-19 09:19:10 <leo-unglaub> i have my old psu with 750w but after trying the new one the old one also does not work :(
2016-05-19 09:19:20 <leo-unglaub> and without an error message i dont see anything
2016-05-19 09:19:43 <clandmeter> you didnt mix the vga power connector with the cpu power connector?
2016-05-19 09:20:21 <leo-unglaub> no, i did not ... but let me check
2016-05-19 09:21:00 <leo-unglaub> its correct
2016-05-19 09:21:07 <leo-unglaub> but let me try another cable
2016-05-19 09:21:15 <clandmeter> some psu dont mention its vga or cpu.
2016-05-19 09:21:41 <clandmeter> which mb do you have?
2016-05-19 09:23:42 <leo-unglaub> its labeled vga1
2016-05-19 09:57:43 <ScrumpyJack> neither power works? perhaps they both work, but you've blown a voltage regulator on your board and you're tripping a fuse
2016-05-19 17:14:40 <jomat> Ok... busybox readlink differs from coreutils readlink, thats why mkinitfs fails for me
2016-05-19 17:55:33 <fabled> urgh. seems dhcpcd-6.11.0 is broken still
2016-05-19 18:21:12 <fabled> i'm gonna revert dhcpcd to 6.10.x
2016-05-20 02:33:27 <Sachiru> Is RVM recommended on Alpine Linux for Ruby management?
2016-05-20 06:31:42 <ScrumpyJack> morning
2016-05-20 07:07:19 <clandmeter> morning climbers
2016-05-20 07:20:46 <vkris> ncopa, any reason for not having shutdown cmd in /sbin/, maybe for sake of FHS if not an issue ?
2016-05-20 07:24:44 <vkris> I think I would create a bugsy for it
2016-05-20 07:25:34 <vkris> for current FHS 3.0 related
2016-05-20 07:29:53 <Sachiru> Can anyone help with redmine installation? I'm getting the following error when running "rake generate_secret_token": LoadError: Error relocating /usr/lib/ruby/gems/2.2.0/extensions/x86_64-linux/2.2.0/therubyracer-0.12.2/v8/init.so: __vfprintf_chk: symbol not found - /usr/lib/ruby/gems/2.2.0/extensions/x86_64-linux/2.2.0/therubyracer-0.12.2/v8/init.so
2016-05-20 07:30:04 <Sachiru> The therubyracer-0.12.2 gem is installed
2016-05-20 07:31:18 <ncopa> vkris: i only use poweroff, not shutdown
2016-05-20 07:34:11 <vkris> it can be a wrapper script around poweroff+halt
2016-05-20 07:34:33 <vkris> #5604
2016-05-20 07:39:58 <clandmeter> Sachiru: probalby its some crappy ruby module that includes binaries
2016-05-20 07:40:05 <clandmeter> spank the developer
2016-05-20 07:40:11 <Sachiru> Sigh
2016-05-20 07:40:22 <Sachiru> Redmine, why are you giving me such a hard time
2016-05-20 07:40:30 <clandmeter> check if you can compile it yourself
2016-05-20 07:40:33 <Sachiru> bundle install builds, but executing rake on anything
2016-05-20 07:40:41 <Sachiru> Causes that error to be spat out
2016-05-20 07:41:15 <clandmeter> therubyracer is needed for redmine?
2016-05-20 07:41:25 <clandmeter> i dont remember installing it.
2016-05-20 07:41:56 <Sachiru> It's required in bundle install for Redmine 3.0
2016-05-20 07:42:58 <clandmeter> Embed the V8 Javascript Interpreter into Ruby
2016-05-20 07:43:01 <clandmeter> jickes
2016-05-20 07:46:10 <Sachiru> Ah, found the dependeny
2016-05-20 07:46:13 <Sachiru> *dependency
2016-05-20 07:46:21 <Sachiru> One plugin requires it
2016-05-20 07:46:27 <Sachiru> And said plugin is now nuked from existence
2016-05-20 07:47:59 <clandmeter> Sachiru: what depends on it?
2016-05-20 07:49:02 <Sachiru> redmine_login_audit
2016-05-20 07:49:08 <Sachiru> The plugin for login auditing
2016-05-20 07:50:26 <clandmeter> ah you installed some wacky plugins. your support is now canceled :p
2016-05-20 07:55:50 <clandmeter> unicorn day at github...
2016-05-20 08:03:43 <vkris> seems over
2016-05-20 08:12:02 <clandmeter> dell is really a pita
2016-05-20 08:12:52 <clandmeter> i want my new xps15! i dont want to wait 2 freaking weeks when shipping is 3-5 days :|
2016-05-20 08:14:47 <Sachiru> Finally got Redmine to work!
2016-05-20 08:14:49 <Sachiru> Thanks very much!
2016-05-20 08:16:14 <barthalion> I thought Dell is a company
2016-05-20 08:16:17 <barthalion> not a pita
2016-05-20 08:16:19 <barthalion> https://dbgcbnch6yz43.cloudfront.net/images/pictures/files/000/036/131/xlarge/lebanese_pita_bread_776x600.jpg?1427363159
2016-05-20 08:16:43 <skarnet> this kind of pita is better than the other one
2016-05-20 08:17:09 <Sachiru> It's nice to be able to consolidate three 1GB RAM VMs into a single 640MB RAM Alpine Linux VM
2016-05-20 08:17:17 <clandmeter> i dont care, give me my notebook!
2016-05-20 08:17:17 <Sachiru> Thanks to how light musl and Alpine is.
2016-05-20 08:17:24 <clandmeter> :)
2016-05-20 08:17:31 <skarnet> 640MB should be enough for everyone
2016-05-20 08:17:40 <fcolista> rotfl
2016-05-20 08:17:47 <barthalion> lol
2016-05-20 08:17:56 <ncopa> lol
2016-05-20 08:26:59 <vkris> yeah ! till we ready mars and make all opensource proprietary, then it would be 6400MB
2016-05-20 08:27:14 <vkris> reach Mars*
2016-05-20 08:32:55 <ScrumpyJack> what's the minial amount of ram anyone's got away with for AL in a VM? I think it failed for me on 32Mb and booted on 40~
2016-05-20 08:33:09 <ScrumpyJack> not tried recently
2016-05-20 08:40:37 <barthalion> you need at least the size of initram with kernel iirc
2016-05-20 08:41:36 <Sachiru> Does anyone have any idea on how to enable the APC for PHP-FPM on Alpine?
2016-05-20 08:56:27 <clandmeter> did you install php-apcu?
2016-05-20 09:27:12 <Sachiru> Yes, and it's configured
2016-05-20 09:27:48 <Sachiru> Load time for the page is 0.15 seconds on the Turnkey VM, and 0.18 seconds on Alpine. Granted, small difference, but I'd like to be able to justify all the effort to my boss.
2016-05-20 09:32:08 <vkris> Sachiru: I noticed AL uses  slightly more mem against couple of other setup I tested
2016-05-20 09:32:27 <vkris> did you try "ab" to benchmark the mem usage/time too ?
2016-05-20 09:33:13 <vkris> that would show a little more realistic picture
2016-05-20 09:39:04 <vkris> then you can send ncopa beer if boss likes it
2016-05-20 10:41:22 <vkris> or maybe do https://wiki.debian.org/X32Port for AL for 30%+ mem save
2016-05-20 11:26:04 <ScrumpyJack> 3.3.3 doesn't boot on an rpi zero. did a previous release ever boot?
2016-05-20 11:29:48 <ncopa> ScrumpyJack: i dont knwo
2016-05-20 11:32:52 <ScrumpyJack> i can't figure out what wrong with 3.3.3 on rpi zero
2016-05-20 13:22:30 <AmatCoder> hi
2016-05-20 13:22:37 <AmatCoder> can somebody merge this: https://github.com/alpinelinux/aports/pull/79 ?
2016-05-20 13:31:01 <jirutka1> huh, today will be v3.4 release?
2016-05-20 13:36:14 <ncopa> jirutka: no
2016-05-20 13:36:18 <ncopa> but i hoped for rc2
2016-05-20 13:36:26 <ncopa> seems like arm will not catch up
2016-05-20 13:36:27 <jirutka> aha, okay
2016-05-20 15:26:03 <AmatCoder> new pull requests on github: libtorrent-raster and deluge
2016-05-20 15:27:08 <AmatCoder> deluge fails on travis because it needs the updated libtorrent-raster
2016-05-20 16:08:56 <jirutka> AmatCoder: I did a code-review and wrote you some suggestions.
2016-05-21 13:32:03 <leo-unglaub> hey :)
2016-05-21 14:08:02 <ScrumpyJack> hey
2016-05-21 14:08:42 <ScrumpyJack> anyone know anything about "linux" on mobile phones?
2016-05-21 14:08:51 <ScrumpyJack> non-android
2016-05-21 14:38:09 <leo-unglaub> ScrumpyJack: there is the ubuntu phone
2016-05-21 14:38:15 <leo-unglaub> and i think the debian guys also have something
2016-05-21 14:38:23 <leo-unglaub> but it works horrible from what i have heared
2016-05-21 16:00:15 <ScrumpyJack> the debian phone is horrible? i can beleive that
2016-05-21 18:41:18 <ammunta> https://github.com/tombh/texttop
2016-05-22 06:17:55 <scadu> ping alpinelinux.org
2016-05-22 06:17:57 <scadu> ping: unknown host alpinelinux.org
2016-05-22 06:17:59 <scadu> woohoo
2016-05-22 06:20:56 <ncopa> i'm on it
2016-05-22 06:21:02 <ncopa> something has wend badly wrong
2016-05-22 06:23:45 <ncopa> i think its ns1.a.o that is broke
2016-05-22 06:33:55 <scadu> seems that all alpine domains are accessible now.
2016-05-22 06:34:18 <ncopa> yes
2016-05-22 06:35:55 <ncopa> it should be back now
2016-05-22 06:36:23 <scadu> good. what happened?
2016-05-22 06:36:51 <ncopa> i need to ask some people. i have theory, but i need to talk with someone first
2016-05-22 07:40:16 <ernierasta_> hi, I will repeat myself here too - there is problem with DNS for alpinelinux.org - entries are disappearing from world dns servers
2016-05-22 07:41:21 <scadu> ncopa: ^
2016-05-22 07:42:16 <ernierasta_> i tested google and opendns, A entry is there (cached probably) but rest is missing including mirrors entries
2016-05-22 07:42:45 <scadu> ernierasta_: our team is working on this.
2016-05-22 07:43:13 <ernierasta_> thanx, great - good luck
2016-05-22 07:43:56 <ernierasta_> I can provide DNS servers if needed (located in europe)
2016-05-22 07:48:36 <scadu> ernierasta_: ncopa is decision maker. he was here a while ago.
2016-05-22 08:00:53 <ernierasta_> of course, just declaring i am ready to help, i will go now, but if I am needed contact me at: ernierasta at zori.cz ( I am talking about ns1-3.virtualzone.eu), the only disadvantage is that currently ns1 is always (and the only) master for them, but i see no problem to make syncer from other master server if needed
2016-05-22 08:42:00 <GarethAdams> Maybe the /topic should be changed to point out that you know about the DNS issues ;)
2016-05-22 09:54:00 <leo-unglaub> hey :)
2016-05-22 10:02:15 <ScrumpyJack> o/
2016-05-22 10:11:47 <leo-unglaub> ScrumpyJack: did you get alpine to run on your phone yet?
2016-05-22 10:24:16 <ScrumpyJack> heh - i don't have a phone yet :)
2016-05-22 10:26:18 <leo-unglaub> well, that could be a limiting factor ...
2016-05-22 10:35:37 <jirutka> well, we have good emulators nowdays :P
2016-05-22 10:35:54 <jirutka> so this is no excuse :P
2016-05-22 10:36:49 <jirutka> leo-unglaub: any news about rustc for musl?
2016-05-22 10:38:35 <jirutka> leo-unglaub: upstream already realeses stdlib compiled for musl, all dependencies of rustc can be provided externally (i.e. llvm is not bundled-only), so the only problem is how to (cross)-compile rustc for musl without binary snapshot…?
2016-05-22 10:39:43 <leo-unglaub> jirutka: i have not looked into it recently, i have a vm with OpenBSD in it an i run rust there corrently
2016-05-22 10:40:59 <jirutka> :/
2016-05-22 10:41:17 <leo-unglaub> i just ssh mount the vm in my alpine
2016-05-22 10:41:22 <leo-unglaub> and then i can work
2016-05-22 10:43:25 <jirutka> interesting, part of the Internet is currently inaccessible for me; I thought that it’s some outage of transit connectivity from my country, but IRC feenode.net works o.O
2016-05-22 10:43:45 <leo-unglaub> propobly your isp having routing issues
2016-05-22 10:44:04 <leo-unglaub> or you are located in a shit country that sometimes cencours the internet randomly
2016-05-22 10:44:32 <jirutka> don’t even talk about this… no, we don’t have censored Internet (yet?)
2016-05-22 10:44:42 <leo-unglaub> where are you from?
2016-05-22 10:46:03 <jirutka> but our fucking parlament has endorsed a law that will allow officers from Ministry of Finance to maintain list of blacklisted sites… you now, becase gambling sites etc.
2016-05-22 10:47:22 <jirutka> luckily it must be also approved by a senate and there’s a huge iniciative from local companies to stop this bullshit
2016-05-22 10:47:34 <jirutka> Czech Republic
2016-05-22 10:48:46 <jirutka> ha, Internet is fully working again, uff
2016-05-22 10:49:03 <scadu> well, we will have something similar in Poland.
2016-05-22 10:49:06 <scadu> https://en.panoptykon.org/articles/new-polish-anti-terrorism-law-every-foreigner-potential-threat
2016-05-22 10:50:19 <scadu> it is already in parliament.
2016-05-22 10:50:46 <leo-unglaub> i am in austria and we are going to have something similar as well
2016-05-22 10:50:57 <leo-unglaub> but i love the czech republic
2016-05-22 10:51:06 <jirutka> this is terrifying :(
2016-05-22 10:51:09 <leo-unglaub> some of the most awesome paramedics i met are from there!
2016-05-22 10:51:27 <jirutka> :)
2016-05-22 10:52:12 <leo-unglaub> https://fbcdn-sphotos-b-a.akamaihd.net/hphotos-ak-xta1/v/t1.0-9/12189850_894336190621851_5303877243661324309_n.jpg?oh=915d50ec6d5315208ac80f936dc604ef&oe=57D23FC4&__gda__=1469716813_245f466d4e2526acaa937b7bd575dbf2
2016-05-22 10:52:22 <leo-unglaub> me with the paramedics from czech republic
2016-05-22 10:54:58 <jirutka> nice
2016-05-22 10:55:01 <jirutka> who are you on this photo?
2016-05-22 10:55:12 <leo-unglaub> i am on the left
2016-05-22 11:35:32 <dxb_> how to add lsb_release in alpine?
2016-05-22 11:36:57 <dxb_> any idea?
2016-05-22 11:46:01 <leo-unglaub> why do you need lsb_release in alpine?
2016-05-22 11:46:29 <leo-unglaub> cat /etc/alpine-release does the same
2016-05-23 07:08:14 <ncopa> hum firefox dies on me once in a while
2016-05-23 07:10:03 <ScrumpyJack> ncopa: which arch?
2016-05-23 07:10:57 <ncopa> x86_64
2016-05-23 07:12:02 <ncopa> i think its #5559
2016-05-23 07:12:18 <ScrumpyJack> ah.
2016-05-23 07:18:30 <fcolista> happens to me also
2016-05-23 07:19:06 <ncopa> it seems like it happens in memcpy in a ringbuffer
2016-05-23 07:19:21 <ncopa> my guess is that the memory region is overlapping or similar
2016-05-23 07:22:29 <ScrumpyJack> do i need bluez-firmware for get the bt device running on the rpi3?
2016-05-23 07:23:23 <ScrumpyJack> and if so, what's the preferered method of getting the files from bluez-firmware into modloop?
2016-05-23 07:23:40 <ncopa> i don tknow if its needed
2016-05-23 07:23:58 <ncopa> i think you can cp the firmware to a dir /firmware on boot media
2016-05-23 07:25:04 <ScrumpyJack>  /firmware on the root of the boot media? i'll try that
2016-05-23 07:35:20 <ScrumpyJack> how do i change the root install for a package with apk?
2016-05-23 07:38:09 <ScrumpyJack> (i want to install bluez-firware into /firmware intead of /
2016-05-23 07:38:28 <ScrumpyJack> apk -p?
2016-05-23 07:40:58 <clandmeter> isnt firmware readonly?
2016-05-23 07:41:25 <ScrumpyJack> ncopa suggested i install it "to a dir /firmware on boot media"
2016-05-23 07:41:56 <ScrumpyJack> trying apk add -p fails with ERROR: Unable to lock database: No such file or directory
2016-05-23 07:42:09 <clandmeter> you can add it to the firmware dir on sdcard
2016-05-23 07:42:24 <ScrumpyJack> that's what i'm trying to do :)
2016-05-23 07:43:16 <clandmeter> openrc has some script to take care of moving it in the right place
2016-05-23 07:43:52 <clandmeter> i think what i did was just extract the apk to /media/sdcard/firmware
2016-05-23 07:44:44 <ScrumpyJack> no way to relocate files in package with with apk?
2016-05-23 07:44:57 <clandmeter> no
2016-05-23 07:45:06 <clandmeter> well
2016-05-23 07:45:10 <ScrumpyJack> np. i'll extract
2016-05-23 07:45:15 <clandmeter> depends if it depens on soemthing
2016-05-23 07:45:38 <clandmeter> with -p in an empty dir, apk will install all deps
2016-05-23 07:46:10 <ScrumpyJack> from earlier: trying apk add -p fails with ERROR: Unable to lock database: No such file or directory
2016-05-23 07:46:20 <clandmeter> right
2016-05-23 07:46:26 <clandmeter> apk doesnt have a db yet
2016-05-23 07:46:30 <clandmeter> so it needs init-db
2016-05-23 07:46:39 <clandmeter> you are bootstrapping
2016-05-23 07:46:56 <clandmeter> but you dont want that.
2016-05-23 07:47:00 <ScrumpyJack> i'll just untar
2016-05-23 07:47:01 <clandmeter> just just want some files
2016-05-23 07:47:09 <clandmeter> so tar is your tool
2016-05-23 07:48:58 <clandmeter> i think modloop(initd file) has the magic to mount the sdcard firmware dir to its correct location (when found).
2016-05-23 08:13:54 <ScrumpyJack> indeed ... modloop found /firmware and i now have /lib/firmware/ with the files from bluez-firmware
2016-05-23 08:20:05 <ScrumpyJack> clandmeter: did you need to do anything else? how do you get the bluetooth kernel module to load the firmware?
2016-05-23 08:21:05 <clandmeter> kernel should automatically find it
2016-05-23 08:21:14 <clandmeter> did you check dmesg?
2016-05-23 08:25:14 <ScrumpyJack> nothing about bt
2016-05-23 08:25:42 <clandmeter> you dont get an error?
2016-05-23 08:25:49 <clandmeter> do you load the module manually?
2016-05-23 08:29:51 <ScrumpyJack> looks like bluetoothd loads the bluetooth module
2016-05-23 08:31:06 <ScrumpyJack> hang on, which module?
2016-05-23 08:31:21 <ScrumpyJack> can you show me your list?
2016-05-23 08:32:39 <ScrumpyJack> there are several in kernel/drivers/bluetooth
2016-05-23 08:32:52 <ScrumpyJack> bcm203?
2016-05-23 08:34:35 <ByteFlinger> Hello. Is there a way to configure MAX_UID when using adduser? I keep hitting some limit of 60000
2016-05-23 08:35:08 <ByteFlinger> And if there is not, have anybody here compile alpine linux before, I wouldn't midn doing it but have had a hard time finding information on where to get started (Or even where the source code is)
2016-05-23 08:41:25 <ncopa> ByteFlinger: might be you can use the package shadow to do it.
2016-05-23 08:41:31 <ncopa> i think its busybox that give you that limit
2016-05-23 08:48:06 <ByteFlinger> ncopa: Yeah I think it is but is there a way to configure it somewhere to allow for a bigger range?
2016-05-23 08:48:39 <ScrumpyJack> clandmeter: can you show me your rpi3 output of lsmod ?
2016-05-23 08:49:01 <clandmeter> ScrumpyJack: i dont have one with alpine
2016-05-23 08:49:29 <ByteFlinger> I found a lot of mentioning that it is possible to configure busybox with some MAX_UID parameter however being new to busybox I cannot find where such a thing can be configured (Maybe it is on compilation time only)?
2016-05-23 08:50:47 <ncopa> i found it
2016-05-23 08:50:55 <ncopa> what would you need to be the max_uid?
2016-05-23 08:51:04 <ncopa> its a compile option
2016-05-23 08:51:08 <ByteFlinger> Damn
2016-05-23 08:51:20 <ncopa> i will fix it for 3.4.0_rc2
2016-05-23 08:51:27 <ByteFlinger> The issue is that I need it to be the same as my host id and I have a pretty big uid
2016-05-23 08:51:50 <ByteFlinger> So basically a support of 150000 would be nice
2016-05-23 08:52:27 <ByteFlinger> Hey that sounds nice. It would be great if it would be a configurable option if possible otherwise setting a high number would cover most people I guess
2016-05-23 08:53:00 <ByteFlinger> Make it 300000 if you can, why not :)
2016-05-23 08:53:12 <ncopa> i was thinking 256000
2016-05-23 08:53:19 <ByteFlinger> Yeah I guess that would suffice
2016-05-23 08:53:37 <ScrumpyJack> it's the new kernel limit something like 4 billion? :)
2016-05-23 08:53:53 <ncopa> i think there might be old lmit of 16 bit
2016-05-23 08:53:57 <ByteFlinger> Is there some guide out there on how to compile Alpine from scratch?
2016-05-23 08:54:18 <ncopa> compile all packages?
2016-05-23 08:54:22 <ByteFlinger> Cause apart from a bunch of the git repos I couldnt find much
2016-05-23 08:54:33 <ncopa> `buildrepo main`
2016-05-23 08:54:36 <ncopa> should beuild it all
2016-05-23 08:54:39 <ncopa> build*
2016-05-23 08:54:42 <ByteFlinger> Well compile the base alpine, the same which is supplied as a docker image today
2016-05-23 08:55:07 <ByteFlinger> But that's only the packages no?
2016-05-23 08:55:25 <ncopa> i think you'd have to manually compile the list  of specific packages
2016-05-23 08:55:38 <ByteFlinger> Hmm OK.
2016-05-23 08:56:31 <ncopa> in main dir in aports repo: ap builddirs $package | while read pkg; do (cd $pkg && abuild -r)|| break; done
2016-05-23 09:00:10 <ByteFlinger> I see. So basically clone aports main, perform whatever changes, build it all and use https://github.com/gliderlabs/docker-alpine to build a rootfs might do the job I guess
2016-05-23 09:02:52 <przemoc> what's `ap`?  where it comes from?
2016-05-23 09:04:26 <ncopa> lua-aports
2016-05-23 09:04:37 <ncopa> its a tool to parse aports
2016-05-23 09:06:14 <przemoc> I see
2016-05-23 09:06:25 <ncopa> to calculate the build time deps
2016-05-23 09:06:28 <ncopa> build order
2016-05-23 09:07:14 <przemoc> BTW in cgit:  1) abuild description is stale (alpine 1.9.x)  2) lua-aports lacks any description
2016-05-23 09:19:13 <przemoc> what's the reason behind  aports-build  being kept literally in aports instead of its own repo?
2016-05-23 09:29:51 <jirutka> hi
2016-05-23 09:32:55 <jirutka> ncopa: there are few patches waiting to be merged:
2016-05-23 09:32:56 <jirutka> * https://github.com/alpinelinux/aports/pull/80 – testing/libtorrent-rasterbar: update (by AmatCoder, I’ve already done code-review)
2016-05-23 09:32:56 <jirutka> * https://github.com/alpinelinux/aports/pull/81 – testing/deluge: update (by AmatCoder, I’ve already done code-review)
2016-05-23 09:32:56 <jirutka> * https://github.com/alpinelinux/aports/pull/82 – testing/R: update (by me)
2016-05-23 09:32:56 <jirutka> * https://github.com/alpinelinux/aports/pull/83 – testing/qemu-init-script: new abuild (by me)
2016-05-23 12:41:42 <vkris> ncopa, can I push github changes to msg.a.o ?
2016-05-23 12:42:04 <vkris> any thoughts on this format ?
2016-05-23 12:42:33 <vkris> I mean for AL related repo
2016-05-23 12:47:33 <ncopa> what kind of github changes?
2016-05-23 12:52:18 <vkris> code commits on AL related repo like github.com/insteps/aport-api
2016-05-23 12:52:32 <barthalion> I don't see the point
2016-05-23 14:30:16 <ScrumpyJack> hmm, i'm just trying to add some kernel modules to linux-rpi2, but the diff tells mt i'm loading 4.4.0 and not 4.4.10
2016-05-23 14:30:30 <ScrumpyJack> /s/mt/me
2016-05-23 14:31:28 <ScrumpyJack> apk search shows me 4.4.11, which is what's in the APKBUILD file
2016-05-23 14:32:07 <fcolista> vkris, do you have a working example on how aports-api can be used?
2016-05-23 14:32:53 <ScrumpyJack> arg, i'm confising myself.
2016-05-23 14:36:13 <ScrumpyJack> why do i see this: grep Kernel\ Configuration config-rpi2.armhf -> Linux/arm 4.4.0 Kernel Configuration
2016-05-23 14:36:44 <ScrumpyJack> grep Kernel\ Configuration config-rpi2.armhf.old -> Linux/arm 4.4.10 Kernel Configuration
2016-05-23 15:18:46 <ncopa> i wonder why the arm images was not built?
2016-05-23 15:22:20 <jirutka1> ncopa: you’ve closed https://github.com/alpinelinux/aports/pull/83 (testing/qemu-init-script) as merged, but it is not in the repo
2016-05-23 15:23:01 <ncopa> oh sorry
2016-05-23 15:23:09 <ncopa> i though it merged it
2016-05-23 15:28:01 <jirutka1> np
2016-05-23 15:44:44 <ncopa> \o/
2016-05-23 15:45:01 <jirutka1> \o/
2016-05-23 15:45:05 <ncopa> arm release 3.4.0_rc2 finally got uploaded
2016-05-23 15:45:22 <skarnet> Grats!
2016-05-23 15:45:44 <ncopa> whats left now for the 3.4.0 release?
2016-05-23 15:47:09 <GarethAdams> \o/
2016-05-23 15:47:11 <jirutka> ncopa: but I still don’t see testing/qemu-init-script in the repo; have you forgotten for the second time? :)
2016-05-23 15:47:21 <ncopa> not forgotten
2016-05-23 15:47:29 <ncopa> i just have my hands full
2016-05-23 15:47:33 <ncopa> and wife is waiting...
2016-05-23 15:47:39 <jirutka> okay, no problem
2016-05-23 15:48:34 <ncopa> pushed
2016-05-23 15:48:36 <przemoc> no problem, but he has his knife sharp and ready... :)
2016-05-23 15:48:47 <przemoc> oh, it won't be needed (for now)
2016-05-23 15:48:57 <skarnet> those Eastern Europe types
2016-05-23 15:49:03 <ncopa> probably good idea to keep it shart for next time :)
2016-05-23 15:49:06 <ncopa> bbl
2016-05-23 16:17:31 <ScrumpyJack> is this just a mockup or is there a larger version of this desktop image somewhere? http://osworld.pl/wp-content/uploads/alpine-326x235.png
2016-05-23 16:17:42 <ScrumpyJack> would make a cool lockscreen
2016-05-23 16:21:07 <przemoc> ScrumpyJack: http://osworld.pl/wp-content/uploads/alpine.png  enjoy!  (but it's not as HQ as it should be, IMHO)
2016-05-23 16:21:53 <vkris> fcolista__: try curl https://api.alpinelinux.org/packages | jq . | less
2016-05-23 16:22:37 <jirutka> wow!
2016-05-23 16:22:44 <jirutka> where can we find documentation of this API?
2016-05-23 16:23:44 <vkris> jirutka: its too early (wip) but from this url you can follow relationships url and other links
2016-05-23 16:23:51 <jirutka> i see
2016-05-23 16:23:55 <jirutka> where’s the source code?
2016-05-23 16:24:11 <vkris> https://github.com/insteps/aport-api
2016-05-23 16:24:23 <jirutka> omg, why PHP?!
2016-05-23 16:24:39 <vkris> what would you prefer ?
2016-05-23 16:25:00 <ScrumpyJack> przemoc: we could make another high quality one without the silly applemac frame. is that mountain range image somewhere?
2016-05-23 16:25:09 <vkris> there is also an lua-turbo based early version
2016-05-23 16:25:10 <jirutka> Alpine uses Lua for most many utilities aetc.
2016-05-23 16:25:20 <jirutka> Python may be also sane option
2016-05-23 16:25:47 <vkris> somewhere in clandmeter repo,
2016-05-23 16:25:58 <jirutka> anyway, it’s definitely better than nothing! :)
2016-05-23 16:26:00 <vkris> don't know if its still active
2016-05-23 16:28:26 <przemoc> ScrumpyJack: most likely, but I'm not a mountain expert, so cannot provide you any further link (I wasn't the author of the mock image you linked originally)
2016-05-23 16:30:12 <jirutka> vkris: how can I find package by its name?
2016-05-23 16:31:21 <vkris> currenlty only pid,id being checked
2016-05-23 16:31:58 <vkris> would add it soon, need to formalize on all restApiUri
2016-05-23 16:33:44 <przemoc> ScrumpyJack: but I think there is other good photo in g+ available header photos (I know about it, because I've chosen it for my [mostly dormant] g+ profile :> - it was way before I even heard about AL).  only licensing of it should be checked.
2016-05-23 16:33:56 <vkris> and together with jsonapi as close as possible
2016-05-23 16:34:36 <przemoc> https://lh3.googleusercontent.com/-beMAvVGcF4s/Udlqgy1k5bI/AAAAAAAAAL8/c7lqwHuql4kNGys10kPlAKnmrPl8k9Q6ACL0B/s1000-fcrop64=1,00003f89ffffffff/8E0gvAL.jpg
2016-05-23 16:35:14 <przemoc> it could be flipped horizontally for better effect :)
2016-05-23 16:37:58 <przemoc> or maybe it wasn't in the standard set...
2016-05-23 16:40:00 <ScrumpyJack> i'll use google, it has a license filter
2016-05-23 16:40:16 <ScrumpyJack> but at work, it's too obvious :)
2016-05-23 16:43:18 <przemoc> I found where I took the image from:  http://i.imgur.com/8E0gvAL.jpg
2016-05-23 16:44:07 <przemoc> > This was taken during a flight in a light aircraft
2016-05-23 16:45:14 <przemoc> it was done by  https://www.reddit.com/user/TangoYankee
2016-05-23 16:55:38 <ScrumpyJack> anyone use awall?
2016-05-23 16:55:49 <jirutka> what is it?
2016-05-23 16:56:23 <ScrumpyJack> http://git.alpinelinux.org/cgit/awall
2016-05-23 16:56:24 <przemoc> I would like to, but it doesn't support custom rules, which limits its usability greatly
2016-05-23 16:56:54 <jirutka> tl;dr?
2016-05-23 16:57:12 <jirutka> aha, firewall
2016-05-23 16:57:13 <przemoc> firewall JSON front-end
2016-05-23 16:57:32 <jirutka> uh, inspired by Shorewall… uh
2016-05-23 16:59:28 <przemoc> I talked about it a bit with kunkku, but he's mostly busy, so improvements in this department won't land soon
2016-05-23 17:00:59 <przemoc> awall inflexibility makes me think about shorewall removal in extended flavor kind of dubious
2016-05-23 17:03:19 <przemoc> I haven't used shorewall yet, though :)
2016-05-23 17:07:52 <ScrumpyJack> how about the acf, does anyone use that?
2016-05-23 17:08:52 <jirutka> nope, I know that it exists, but I prefer terminal and editing plain old text configs
2016-05-23 17:12:20 <ScrumpyJack> speaking of which, i was looking for a basic ncurses/terminal intrastructive monitoring system this morning to monitor a bunch of hosts. any advice?
2016-05-23 17:13:43 <jirutka> ^ me too
2016-05-23 17:14:55 <jirutka> I know only big complex systems (like Icinga) and/or horrible shits (like nagios or zabbix)
2016-05-23 17:15:55 <jirutka> this one https://github.com/arachnys/cabot is quite interesting, not so insanly complex like others
2016-05-23 17:16:23 <jirutka> but I don’t know about any ncurses/terminal interface for any of these
2016-05-23 17:32:27 <ScrumpyJack> jirutka: https://github.com/yaronn/blessed-contrib
2016-05-23 17:33:04 <jirutka> I know this one :)
2016-05-23 17:33:19 <jirutka> but it’s not a monitoring software,
2016-05-23 17:35:32 <ScrumpyJack> no, but it would be awesome to use the framework to display monitoring data
2016-05-23 17:35:39 <jirutka> yeah
2016-05-23 17:36:34 <przemoc> I was thinking about such CLI monitoring system myself too.  and some CLI dashboard showing what's going on in system (possibly run 24/7 on tty8 for instance)
2016-05-23 17:37:24 <ScrumpyJack> right, hometime
2016-05-23 17:57:34 <vkris> jirutka: you can try `curl https://api.alpinelinux.org/packages/firefox | jq . | less`
2016-05-23 17:59:00 <jirutka> this is not a good API – for number, you will get a single resource, but for non-number, you will get a collection…
2016-05-23 17:59:58 <jirutka> and what if some package would have a numeric name? however, I don’t know if it’s even allowed to have just numbers in a package name
2016-05-23 18:00:30 <vkris> this is not formalized /packages/{name
2016-05-23 18:00:41 <vkris> maybe /packages/name/<name>
2016-05-23 18:01:52 <vkris> note ids/pids would be nums only
2016-05-23 18:02:09 <jirutka> OMG f*cking IRC
2016-05-23 18:02:30 <jirutka> this would be more clear imo: /packages?name={name}
2016-05-23 18:02:32 <vkris> but name would be mix, so even if does not start with alpha it might just work
2016-05-23 18:03:17 <vkris> but be a problem if name is fully numbers
2016-05-23 18:03:29 <vkris> with I have not found yet in AL
2016-05-23 18:03:33 <vkris> which*
2016-05-23 18:03:52 <jirutka> you can even integrate RSQL <https://github.com/jirutka/rsql-parser> and allow more complex queries like ?query=name==firefox;arch=in=(x86_64,x86) etc.
2016-05-23 18:04:15 <vkris> yes FILTER is not yet done
2016-05-23 18:05:12 <vkris> what expect full API in wip process ;)
2016-05-23 18:05:15 <barthalion> …in java
2016-05-23 18:05:46 <jirutka> the grammar is quite simple, you can reimplement it in Lua or any other lang
2016-05-23 18:07:18 <vkris> does it support clean uri ?
2016-05-23 18:07:33 <jirutka> what do you mean by it?
2016-05-23 18:08:21 <vkris> ?name/<name>/arch/x86_64,x86/foo/blah
2016-05-23 18:08:28 <vkris> kinda key/value pair
2016-05-23 18:09:05 <jirutka> what? RSQL? no, this are two quite different things
2016-05-23 18:09:13 <vkris> ah ok
2016-05-23 18:09:25 <jirutka> > ?name/<name>/arch/x86_64,x86/foo/blah
2016-05-23 18:09:25 <jirutka> this is VERY limited
2016-05-23 18:11:46 <jirutka> a little more expressive would be to use matrix parameters, but these are very rarely used
2016-05-23 18:12:15 <jirutka> however, it may be quite sufficient for this API
2016-05-23 18:15:17 <vkris> hmm.. look ok
2016-05-23 18:16:33 <vkris> we might not need expression like id>123
2016-05-23 18:17:04 <jirutka> but we might need exression like version>4.2
2016-05-23 18:17:13 <vkris> though there is version but we do not have multiple entries of same pkg
2016-05-23 18:17:28 <vkris> only single for all archs
2016-05-23 18:17:32 <jirutka> actuallly, we have…
2016-05-23 18:17:38 <jirutka> because we have multiple respositories
2016-05-23 18:18:01 <vkris> how ?
2016-05-23 18:18:14 <jirutka> e.g. https://pkgs.alpinelinux.org/packages?name=qemu&repo=all&=&arch=x86_64&=&maintainer=all&=
2016-05-23 18:18:20 <jirutka> see?
2016-05-23 18:18:39 <jirutka> you might also need to filter packages by build date
2016-05-23 18:18:53 <vkris> yes different branches
2016-05-23 18:19:00 <jirutka> there are always use cases for operators
2016-05-23 18:19:53 <vkris> edge will mostly like have updated(ge) version, that would not be very logical to search
2016-05-23 18:21:22 <vkris> need to focus more on branch/arch/repo(main|community)
2016-05-23 18:21:45 <jirutka> yes
2016-05-23 18:22:18 <jirutka> filtering is not a replacement for direct paths, it’s a supplement
2016-05-23 18:51:31 <jirutka> ncopa: good for merge, fixes security vulnerability https://github.com/alpinelinux/aports/pull/86 (by AmatCoder)
2016-05-23 19:04:28 <Shamanoid> Hi all, quick noob question, I am actually trying to package a binary into a .apk like I can do for .deb .rpm is there a simple way to do that using abuild maybe ? Thanks
2016-05-23 20:18:39 <AmatCoder> build-edge-armhf is failing to build libtorrent-rasterbar: http://build.alpinelinux.org/buildlogs/build-edge-armhf/testing/libtorrent-rasterbar/libtorrent-rasterbar-1.1.0-r0.log
2016-05-23 20:19:58 <AmatCoder> so I going to downgrade it: https://github.com/alpinelinux/aports/pull/87
2016-05-23 21:03:49 <faffolter> please git pull git://git.alpinelinux.org/user/fab/aports
2016-05-23 21:04:32 <faffolter> what is the process to apply for commit access?
2016-05-23 21:05:11 <poptart> lol
2016-05-23 21:06:13 <faffolter> my updates are usually sitting around for weeks and till i ask again for a pull
2016-05-23 21:07:38 <przemoc> I guess it's better to send such requests to ML.
2016-05-23 21:07:58 <jirutka> as I know, the standard way is to send patches to alpine-aports@alpinelinux.org (these goes to http://patchwork.alpinelinux.org/) or open pull request on GitHub https://github.com/alpinelinux/aports, not to request pulling from your repo
2016-05-23 21:08:54 <faffolter> well, i'm still doing it the way we agreed some years ago.
2016-05-23 21:09:46 <jirutka> however, based on my experience, you must still remind it on IRC to speed up things; but then it’s quite quick
2016-05-23 21:09:51 <faffolter> seems that i missed the change away from git.alpinelinux.org to github
2016-05-23 21:10:11 <jirutka> GitHub is just a mirror of git.a.o
2016-05-23 21:10:34 <jirutka> even pull requests are not merged directly on GitHub, but pulled to git.alpinelinux.org
2016-05-23 21:10:47 <przemoc> I think, faffolter, that PR from own git repo is fine, it's just that requesting it on IRC while most people are absent won't help you.  that's why ML seems better.
2016-05-23 21:11:49 <jirutka> I’m quite new here (few months), so I don’t know about personal repos on git.alpinelinux.org; as I said, what I’ve seen here is just alpine-aports list or GH
2016-05-23 21:12:49 <poptart> I haven't had many issues. Mine normally only take like a day or two after posting on the ML
2016-05-23 21:18:32 <vkris> jirutka: try now, pkgname can now begin with digits
2016-05-23 21:18:52 <faffolter> i will talk with ncopa about this
2016-05-23 21:18:53 <vkris> but AL still does not have any such pkg
2016-05-24 10:04:26 <ScrumpyJack> afternoon
2016-05-24 10:19:09 <jirutka> hi
2016-05-24 10:19:14 <ncopa> hey
2016-05-24 10:20:01 <jirutka> ncopa: I’ve opened two PRs for abuild:
2016-05-24 10:20:01 <jirutka> * https://github.com/alpinelinux/abuild/pull/1
2016-05-24 10:20:01 <jirutka> * https://github.com/alpinelinux/abuild/pull/2
2016-05-24 10:20:01 <jirutka> The first one may be a little controversial, I’m not quite sure if this is a good approach to solve it (using find).
2016-05-24 10:57:31 <ScrumpyJack> ncopa: about #5620
2016-05-24 10:58:28 <ScrumpyJack> perhaps we should have a non-free repo of stuff that can't be distributed along side GPL'ed software?
2016-05-24 11:00:18 <ScrumpyJack> (or for whatever reason that they can't be conveyed)
2016-05-24 11:02:33 <nmeum> we have a non-free repo
2016-05-24 11:02:42 <nmeum> http://git.alpinelinux.org/cgit/aports/tree/non-free
2016-05-24 11:03:00 <nmeum> but package in it are not built currently, are they?
2016-05-24 11:05:40 <ncopa> packages are not built because we are not allowed to redistribute those
2016-05-24 11:05:55 <ncopa> jirutka: they look both ok to me
2016-05-24 11:21:46 <jirutka> ncopa: great! so are you gonna merge them? :)
2016-05-24 11:26:00 <ncopa> jirutka: i'm not happy
2016-05-24 11:26:10 <ncopa> why should this delay the 3.4 release?
2016-05-24 11:26:23 <ncopa> i already spent time troubleshooting this testing thing
2016-05-24 11:26:27 <jirutka> ncopa: I dunno, why it should delay it?
2016-05-24 11:26:46 <ncopa> i am currently working on getting v3.4 out
2016-05-24 11:26:51 <ncopa> do we need this for 3.4?
2016-05-24 11:27:07 <jirutka> okay, so hold these patches until v3.4 is merged
2016-05-24 11:27:31 <ncopa> i can either continue work on getting v3.4 out or i can continue work on the stuff people want for v3.5 with the result of v3.4 getting delayed
2016-05-24 11:27:46 <jirutka> I vote for the first option
2016-05-24 11:27:55 <jirutka> these patches can wait :)
2016-05-24 11:28:18 <ncopa> yes
2016-05-24 11:28:21 <ncopa> lets fix bugs now
2016-05-24 11:28:35 <ncopa> so v3.4 is good when it is shiped
2016-05-24 13:17:30 <ScrumpyJack> is patchwork working? i can't see any recent patches from the mailing list. perhaps just mine :)
2016-05-24 13:19:25 <ScrumpyJack> nothing since that dns issue on sunday (could be a coincidence)
2016-05-24 13:23:10 <scadu> ScrumpyJack: i have noticed a delay on patchwork before problem with dns.
2016-05-24 13:24:08 <stwa> ScrumpyJack: i submitted patches too on sunday evening which are not in patchwork :/
2016-05-24 13:26:23 <barthalion> clandmeter: ^
2016-05-24 16:15:17 <ncopa> jirutka1:  what do you think about rename qemu-init-script to qemu-openrc
2016-05-24 16:15:38 <ncopa> and add a binfmt misc eumlation script
2016-05-24 16:17:24 <jirutka1> ncopa: Well, the project is actually named qemu-openrc, but both Alpine and Gentoo uses  -init-script(s) for such things
2016-05-24 16:17:34 <jirutka1> there’s https://pkgs.alpinelinux.org/package/v3.3/main/x86_64/udev-init-scripts
2016-05-24 16:17:49 <ncopa> we normally follow upstream package name
2016-05-24 16:18:16 <jirutka1> https://packages.gentoo.org/packages/search?q=init-script
2016-05-24 16:18:19 <ncopa> i need this too: http://tpaste.us/3qqO
2016-05-24 16:18:38 <jirutka1> OMG, what’s that?!
2016-05-24 16:18:51 <jirutka1> wtf is qemu-mipsel:M:0:\x7f\x45\x4c\x46\x01\x01\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x02\x00\x08\x00:\xff\xff\xff\xff\xff\xff\xff\x00\xfe\xff\xff\xff\xff\xff\xff\xff\xfe\xff\xff\xff:/usr/bin/qemu-mipsel-static:OC?
2016-05-24 16:19:00 <ncopa> you can register the binary format of arch
2016-05-24 16:19:27 <ncopa> so a mipsel compiled binary can be executed by kernel
2016-05-24 16:19:50 <ncopa> using qemu as "interpreter"
2016-05-24 16:19:59 <jirutka1> aha
2016-05-24 16:20:13 <ncopa> i think we need change those -static though
2016-05-24 16:20:59 <ncopa> actually
2016-05-24 16:21:05 <jirutka1> what exactly do these binary data mean?
2016-05-24 16:21:48 <ScrumpyJack> 
w
2016-05-24 16:21:50 <ncopa> https://en.wikipedia.org/wiki/Binfmt_misc
2016-05-24 16:22:59 <ncopa> its the magic, the id in the binary that tells what arch it is
2016-05-24 16:23:11 <ncopa> so you start correct qemu binary for the correct arch
2016-05-24 16:23:40 <jirutka1> interesting
2016-05-24 16:24:10 <jirutka1> is it Alpine specific or can this works even on Gentoo?
2016-05-24 16:24:16 <jirutka1> *work
2016-05-24 16:24:19 <ncopa> works on all linux
2016-05-24 16:24:32 <ncopa> https://wiki.debian.org/QemuUserEmulation
2016-05-24 16:24:42 <jirutka1> okay, I’m gonna add it to https://github.com/jirutka/qemu-openrc
2016-05-24 16:25:02 <ncopa> https://wiki.gentoo.org/wiki/Cross_Container_Support_Project
2016-05-24 16:25:53 <ncopa> the reason that they point to a -static is because you need the x86_64 static binary in an arm container on x86_64 kernel
2016-05-24 16:26:37 <ncopa> so, on x86_64 host, extract arm binaries to container dir, add qemu-arm-static too, register the binfmt_misc
2016-05-24 16:27:00 <ncopa> and then you can run an arm container on an x86_64 host
2016-05-24 16:28:24 <jirutka1> so every arm binary is automatically executed via Qemu emulator?
2016-05-24 16:28:46 <ncopa> yes
2016-05-24 16:29:41 <jirutka1> how big overhead does it have?
2016-05-24 16:29:51 <ncopa> performance or size?
2016-05-24 16:29:52 <jirutka1> comparing to running VM
2016-05-24 16:29:55 <jirutka1> performance
2016-05-24 16:30:12 <ncopa> you need to start up the qemu every exec
2016-05-24 16:30:22 <ncopa> so it is slower in theory
2016-05-24 16:31:31 <jirutka1> so what are the benefits?
2016-05-24 16:31:54 <ncopa> you dont need install an OS to execute a single binary
2016-05-24 16:32:10 <ncopa> which will need a kernel and drivers etc
2016-05-24 16:32:37 <jirutka1> but you will still need the whole rootfs…?
2016-05-24 16:33:00 <ncopa> depends on what you use it for
2016-05-24 16:33:13 <ncopa> if you forexample crosscompile a static binary
2016-05-24 16:33:23 <ncopa> then you only need the binary
2016-05-24 16:34:12 <ncopa> you dont need full rootfs either
2016-05-24 16:34:38 <ncopa> you can also extract the needed libs into a dir and set LD_LIBRARY_PATH
2016-05-24 16:35:06 <ncopa> on debian i think they support full multiarch
2016-05-24 16:35:20 <ncopa> install the libs in /usr/$arch/lib
2016-05-24 16:35:38 <jirutka1> interesting, maybe we can it to compile Rust packages! XD
2016-05-24 16:36:23 <ncopa> should not make any difference, since with rust, the arch is not the problem
2016-05-24 16:36:38 <ncopa> (my guess is dynamic linker is)
2016-05-24 16:37:01 <jirutka1> the problem is how to compile rustc for x86_64 musl
2016-05-24 16:37:33 <jirutka1> because you need rustc to compile rustc… and there are not prebuild binaries for x86_64 musl yet
2016-05-24 16:37:42 <ncopa> so you need crosscompile
2016-05-24 16:37:49 <jirutka1> yeah
2016-05-24 16:38:31 <jirutka1> I *think* that it may not be a big problem, but my knowledge of C/C++ tools is quite limited and I don’t know much about cross-compiling
2016-05-24 16:39:51 <jirutka1> anyway, back to qemu-openrc… so you want to add http://tpaste.us/3qqO and rename abuild to qemu-openrc?
2016-05-24 16:42:51 <ncopa> i think so yes
2016-05-24 16:42:55 <ncopa> yes
2016-05-24 16:43:16 <jirutka1> how should be that runscript named?
2016-05-24 16:43:38 <ncopa> i think binfmt_misc
2016-05-24 16:44:02 <ncopa> or qemu-binfmt_misc
2016-05-24 16:44:24 <jirutka1> should I keep -static suffix?
2016-05-24 16:44:30 <ncopa> not sure
2016-05-24 16:44:40 <ncopa> maybe have a config option for that
2016-05-24 16:44:45 <ncopa> in conf.d
2016-05-24 16:44:54 <ncopa> QEMU_SUFFIX="-static"
2016-05-24 16:45:04 <ncopa> or similar
2016-05-24 16:45:17 <jirutka1> okay
2016-05-24 16:45:41 <ncopa> need to go
2016-05-24 16:45:48 <ncopa> see u l8r
2016-05-24 16:45:49 <jirutka1> see ya :)
2016-05-25 01:29:47 <Sachiru> Query: Are there other ruby webserver engines (Puma/Passenger) available in Alpine Linux?
2016-05-25 07:16:38 <ScrumpyJack> clandmeter: morning dude. you about? can you take a look at patchwork?
2016-05-25 07:34:38 <clandmeter> morning
2016-05-25 07:34:45 <clandmeter> whats the problem?
2016-05-25 07:39:18 <clandmeter> looks like emails do not get send to pw
2016-05-25 07:39:27 <clandmeter> ncopa: can you check mail.a.o?
2016-05-25 07:52:26 <ncopa> clandmeter: it moved to new host last weekend
2016-05-25 08:41:17 <ncopa> clandmeter: i think i fixed the config for patchwork
2016-05-25 08:41:33 <ncopa> this one is a bit confusing: http://www.postfix.org/postconf.5.html#relay_domains
2016-05-25 08:42:24 <vkris> hi, morning
2016-05-25 08:42:24 <vkris> anyone building a frontend ui based on api.a.o, pls let me know of any requirements
2016-05-25 08:42:27 <vkris> starting point for accessing all data is api.a.o/packages
2016-05-25 08:42:30 <vkris> I have added api.a.o/packages/<name>
2016-05-25 08:42:30 <vkris> will be adding api.a.o/search , for basic search.
2016-05-25 08:42:30 <vkris> jsonapi has provisions for [filters] that I have not formalized yet, still thinking its need.
2016-05-25 08:42:33 <vkris> Any suggestion to more URI is welcomed
2016-05-25 08:42:36 <vkris> code is here, https://github.com/insteps/aport-api
2016-05-25 08:43:22 <ncopa> need to be per branch
2016-05-25 08:43:40 <clandmeter> ncopa: can you resend the missing patches?
2016-05-25 08:43:40 <ncopa> so api.a.o/packages/v3.2/<name>
2016-05-25 08:44:14 <clandmeter> ncopa: thats what the filter is for.
2016-05-25 08:45:23 <vkris> current db has "latest-stable, v3.3, edge"
2016-05-25 08:45:29 <vkris> for all arch
2016-05-25 08:47:09 <vkris> db master ^^ any thoughts on adding more version (that is still active) ?
2016-05-25 08:47:30 <vkris> not reached EOL
2016-05-25 08:47:55 <vkris> btw, what is "latest-stable"
2016-05-25 08:49:31 <clandmeter> vkris: latest-stable should be v3.3
2016-05-25 08:49:59 <clandmeter> where did you see latest-stable?
2016-05-25 08:51:25 <vkris> in current db - aports.db.20160523
2016-05-25 08:51:52 <vkris> 'select branch from repoversion'
2016-05-25 08:53:16 <clandmeter> ah i think its an old entry
2016-05-25 08:53:28 <clandmeter> before i had latest-version
2016-05-25 08:53:36 <clandmeter> but i changed it to v3.3
2016-05-25 08:53:52 <clandmeter> latest-version is a symlink on our mirror
2016-05-25 08:54:17 <clandmeter> err latest-stable..
2016-05-25 08:54:48 <vkris> ok
2016-05-25 08:55:47 <vkris> so currently only edge+v3.3
2016-05-25 08:56:12 <clandmeter> yes, when 3.4 is released ill add it to the config
2016-05-25 08:56:17 <ncopa> hum
2016-05-25 08:56:29 <ncopa> how do i inject a raw message
2016-05-25 08:57:11 <ncopa> i have a raw message that i want inject to a given mail server
2016-05-25 08:57:49 <ncopa> ah
2016-05-25 08:57:54 <ncopa> its my stupid isp
2016-05-25 08:58:01 <ncopa> that blocks tcp 25
2016-05-25 08:58:01 <clandmeter> sendmail?
2016-05-25 08:59:21 <ncopa> whee
2016-05-25 09:00:09 <clandmeter> looks like its working
2016-05-25 09:02:15 <ncopa> cool
2016-05-25 09:06:58 <vkris> hmm.. uri like api.a.o/packages/v3.2 can make splitting of db easy
2016-05-25 09:10:59 <vkris> if db is being updated (not re-created) then less changes to stable dbs , less chance of corrupting
2016-05-25 09:13:25 <clandmeter> if you have a currupt db, you are doing something else wrong.
2016-05-25 09:14:11 <clandmeter> vkris: you are free to generate your own db, and adapt your api.
2016-05-25 09:14:45 <clandmeter> you can have several api entry points based on branches like ncopa suggested.
2016-05-25 09:19:08 <ScrumpyJack> ncopa: clandmeter: thanks for fixing patchwork :)
2016-05-25 09:30:52 <leo-unglaub> hey :)
2016-05-25 09:31:19 <leo-unglaub> my desktop crashed and i am reinstalling it currently ... so expect some "why is X not starting questions" *g*
2016-05-25 11:00:36 <Sachiru> Hi! Are there plans to support other ruby app engines (other than Unicorn) like Puma, Phusion Passenger, etc.?
2016-05-25 11:01:23 <ncopa> Sachiru: can you install those with gem?
2016-05-25 11:02:15 <Sachiru> Haven't tried. I was just wondering since we have ruby-unicorn, maybe we can have ruby-passenger/puma too
2016-05-25 11:03:34 <Sachiru> Additionally, Phusion Passenger requires a bit more work compared to just gem install: https://www.phusionpassenger.com/library/install/nginx/install/oss/rubygems_norvm/
2016-05-25 11:08:11 <ncopa> we have removed the ruby-* packages for v3.4
2016-05-25 11:08:20 <ncopa> its just too much work to maintain them
2016-05-25 11:08:28 <ncopa> and upstream only support installing via gem
2016-05-25 11:09:21 <Sachiru> Ah, I see. So the recommended procedure is to just do the manual installation process for Passenger?
2016-05-25 11:09:31 <Sachiru> And gem install for puma + fix breakage as it occurs?
2016-05-25 11:09:42 <ncopa> yes
2016-05-25 11:12:16 <Sachiru> I see, thanks!
2016-05-25 11:12:58 <Sachiru> BTW: Is Unicorn still the recommended webapp engine for Redmine, or is there another engine recommended (Puma/Passenger/Thin/Some other engine)?
2016-05-25 11:18:21 <GarethAdams> ncopa: you mean the gem packages, and not ruby-dev (I hope)
2016-05-25 11:19:21 <ncopa> GarethAdams: yes, the gem packages, not ruby-dev
2016-05-25 11:20:47 <GarethAdams> tbf that was obvious from context, I shouldn't have needed to clarify
2016-05-25 11:20:51 <GarethAdams> my bad
2016-05-25 13:06:08 <leo-unglaub> ncopa: what do you think about using busybox as the default ntp daemon?
2016-05-25 13:06:26 <leo-unglaub> äh i meant ntp client
2016-05-25 13:06:49 <leo-unglaub> busybox, opensmtpd, chrony ... i think the smallest is busybox
2016-05-25 13:07:43 <jirutka> IIRC busybox ntp client is not a real NTP client
2016-05-25 13:08:03 <jirutka> I mean it’s used only to one-time sync of clock
2016-05-25 13:08:18 <jirutka> I personally prefer OpenNTPd
2016-05-25 13:10:39 <leo-unglaub> i dont mean ditching the other ones, but i was just woundering about the default
2016-05-25 13:10:55 <jirutka> I understand that
2016-05-25 13:11:08 <leo-unglaub> for a normal workstation, do you really need the time syncing daemon? or is it enought at startup
2016-05-25 13:11:16 <jirutka> if Busybox NTP isn’t a real NTP client, then it shouldn’t be the default IMO
2016-05-25 13:11:29 <jirutka> yes I need
2016-05-25 13:11:30 <leo-unglaub> but on the other hand, if alpine is used more on servers then maybe my ideas is bad
2016-05-25 13:11:47 <jirutka> even on workstation, how often do you restart?
2016-05-25 13:11:52 <jirutka> Me, once a month?
2016-05-25 13:12:28 <leo-unglaub> jirutka: every day ... because my girlfriend kills me if i keep the computer running *G*
2016-05-25 13:12:33 <leo-unglaub> but maybe i am a bad reference
2016-05-25 13:12:57 <jirutka> I just close lid
2016-05-25 13:13:03 <jirutka> I use notebook as my primary computer
2016-05-25 13:16:27 <leo-unglaub> okay, thats different
2016-05-25 15:18:48 <ncopa> i think i have figured out why ff crashes
2016-05-25 15:20:57 <ScrumpyJack> how about chomeium? :)
2016-05-25 15:21:06 <ScrumpyJack> (with an r)
2016-05-25 15:21:55 <ncopa> it seems that they allocate 128kb buffer on stack
2016-05-25 15:25:43 <ncopa> re chromium, this bug hits us: https://crbug.com/584920
2016-05-25 15:29:17 <leo-unglaub> ncopa: what ff crashes?
2016-05-25 15:29:25 <ncopa> ff 46
2016-05-25 15:29:41 <ncopa> https://bugs.alpinelinux.org/issues/5559
2016-05-25 15:38:04 <leo-unglaub> ncopa: thanks
2016-05-25 15:38:08 <leo-unglaub> interresting bug
2016-05-25 15:38:18 <leo-unglaub> but the one that anoys me even more is the gtk3 theme they use
2016-05-25 15:38:27 <leo-unglaub> checkboxes are now like 10cm large ...
2016-05-25 15:38:37 <ncopa> screenshot?
2016-05-25 15:38:42 <ncopa> looks okish here
2016-05-25 15:40:34 <ncopa> whee!
2016-05-25 15:40:44 <ncopa> i have a workaround/patch
2016-05-25 15:43:45 <clandmeter> patchwork and django... what a nightmare.
2016-05-25 15:44:35 <leo-unglaub> ncopa: all my gtk3 apps have that problem
2016-05-25 15:45:00 <leo-unglaub> but i think thats on purpose from those gtk3 hippsters
2016-05-25 15:45:53 <ncopa> wouldnt surprise me :)
2016-05-25 15:47:50 <scadu> leo-unglaub: looks cool. in case when you have fullhd and 13-15' screen :f
2016-05-25 15:50:54 <leo-unglaub> so, finally reinstalled my desktop
2016-05-25 15:51:04 <leo-unglaub> most stuff should work again :)
2016-05-25 16:53:37 <leo-unglaub> since a couple days i have strange green coloring in my terminal
2016-05-25 16:53:37 <leo-unglaub> http://i.imgur.com/EutcKM8.png
2016-05-25 16:53:42 <leo-unglaub> do you have the same?
2016-05-25 16:54:04 <jirutka> nope
2016-05-25 20:03:39 <vilhelmgray> Hello, I'm trying to get a better understanding of the default Linux kernel configuration of distributions (see http://unix.stackexchange.com/questions/285222/how-do-linux-kernel-modules-get-enabled-in-a-linux-distros-kernel-package). I see the git log for Alpine's linux-vanilla package shows changes to the default config file used (http://git.alpinelinux.org/cgit/aports/log/main/linux-vanilla/con
2016-05-25 20:03:40 <vilhelmgray> fig-vanilla.x86_64). What determines whether a module or kernel feature is enabled in the default config file?
2016-05-25 20:12:47 <barthalion> Long story short, whatever is enabled in grsec flavor, makes it to vanilla
2016-05-25 20:13:19 <barthalion> And at least in past, we tried to keep kernel not tainted
2016-05-25 20:13:59 <barthalion> Not sure how it looks now as it's been a while since I used Alpine on bare metal
2016-05-25 20:14:45 <barthalion> But if something is usable to one of developers or users, why not enable it?
2016-05-25 20:16:30 <barthalion> Pretty much it, no magical procedures take place
2016-05-25 20:16:38 <barthalion> vilhelmgray: see above
2016-05-25 20:17:45 <vilhelmgray> Thanks barthalion, I understand now. :)
2016-05-25 20:23:39 <barthalion> No problem
2016-05-25 20:26:23 <vilhelmgray> I'm trying to get a better understanding of the default Linux kernel configuration of distributions (see http://unix.stackexchange.com/questions/285222/how-do-linux-kernel-modules-get-enabled-in-a-linux-distros-kernel-package).
2016-05-25 20:26:32 <leo-unglaub> hmm, is the busybox tar broken when it comes to xz support?
2016-05-25 20:26:32 <vilhelmgray> oops sorry about,I typo'd
2016-05-25 20:28:05 <leo-unglaub> can someone please try the following command
2016-05-25 20:28:19 <leo-unglaub> tar cfvJ test.tar.xz /some/folder
2016-05-25 20:54:35 <ncopa> leo-unglaub: i dont think there are any unxz in busybox?
2016-05-26 06:17:16 <clandmeter> good morning climbers
2016-05-26 06:47:46 <ScrumpyJack> o/
2016-05-26 08:33:29 <leo-unglaub> good morning :)
2016-05-26 08:49:40 <leo-unglaub> (20:54:35) ncopa: leo-unglaub: i dont think there are any unxz in busybox?
2016-05-26 08:49:51 <leo-unglaub> accordig to the tar --help there is xz support
2016-05-26 08:56:09 <ncopa> it likely depend on external helper program?
2016-05-26 08:56:41 <ncopa> does it work if you apk add xz?
2016-05-26 08:57:02 <leo-unglaub> i have to look it up in the source
2016-05-26 08:57:08 <leo-unglaub> how they implemented it
2016-05-26 08:57:50 <leo-unglaub> clandmeter: did you remove the cert again? http://pkgs.alpinelinux.org
2016-05-26 09:17:42 <clandmeter> leo-unglaub: ?
2016-05-26 09:33:19 <leo-unglaub> it does not redirect to tls anymore
2016-05-26 09:33:42 <leo-unglaub> and there is not htst header anymore
2016-05-26 09:33:50 <leo-unglaub> i thought you did set the header
2016-05-26 09:49:14 <clandmeter> htswhat?
2016-05-26 09:49:30 <clandmeter> yes i removed it
2016-05-26 09:49:47 <clandmeter> cause it was also redirecting subdomains
2016-05-26 09:50:08 <leo-unglaub> HSTS -> https://en.wikipedia.org/wiki/HTTP_Strict_Transport_Security
2016-05-26 10:30:59 <ScrumpyJack> fabled: hi dude. did the linux-rpi kernel get rebuilt with #5619 ?
2016-05-26 10:31:46 <fabled> ScrumpyJack, yes, it's in edge/main
2016-05-26 10:31:50 <fabled> also v3.4
2016-05-26 10:33:02 <ScrumpyJack> thank you
2016-05-26 10:40:43 <clandmeter> what do you call a daemon which has 713 options/switches listed in --help?
2016-05-26 10:44:06 <royger> git?
2016-05-26 10:44:15 <leo-unglaub> ncopa: does the /etc/init.d/mariadb script work for you since the lastest upgrade?
2016-05-26 10:44:43 <leo-unglaub> https://paste.debian.net/plain/705358
2016-05-26 10:51:31 <clandmeter> royger: uwsgi
2016-05-26 11:00:48 <ncopa> leo-unglaub: looks like its already running
2016-05-26 11:00:57 <ncopa> ps xa | grep mariadb
2016-05-26 11:05:28 <leo-unglaub> yeah, stop says it stopped it but then its still running
2016-05-26 11:06:48 <leo-unglaub> ncopa: i think the problem is because he tryes to start mysqld_safe but the running daemon is mysql
2016-05-26 11:06:56 <leo-unglaub> so he maybe does not close the correct one
2016-05-26 11:07:04 <ncopa> sounds like that
2016-05-26 11:10:14 <leo-unglaub> hmm, on the other hand, mysqld_safe is the correct daemon to start
2016-05-26 11:38:23 <jirutka> hi
2016-05-26 11:39:06 <leo-unglaub> hi jirutka
2016-05-26 12:02:36 <jirutka> ncopa, are you here?
2016-05-26 12:04:30 <clandmeter> leo-unglaub: the initd script is broken
2016-05-26 12:05:03 <clandmeter> ncopa: you broke it when you tried to fix it.
2016-05-26 12:07:43 <ncopa> ugh
2016-05-26 12:08:20 <leo-unglaub> clandmeter: while we are at the topic, could you please remove "need firewall" from the hiawatha initd file as well
2016-05-26 12:08:22 <ncopa> with this? 13d027613dc9bc51202eb338821c558580c031a4
2016-05-26 12:08:30 <leo-unglaub> its not needed and breaks stuff
2016-05-26 12:09:03 <clandmeter> ncopa: soembody tried to be smart with a new intid, but it would be better to revert to orignal
2016-05-26 12:09:19 <jirutka> cladnmeter: which initd?
2016-05-26 12:09:21 <ScrumpyJack> fabled: can you remind me how to install the new modules in the rpi?
2016-05-26 12:09:39 <clandmeter> jirutka: mariadb
2016-05-26 12:10:25 <jirutka> what exactly is wrong in it?
2016-05-26 12:10:52 <clandmeter> http://git.alpinelinux.org/cgit/aports/commit/main/mariadb/mariadb.initd?id=3b88157511585b012b9b698fe3a7f3f607a1cfa2
2016-05-26 12:11:13 <jirutka> that’s not answer
2016-05-26 12:11:30 <jirutka> hmm, I think that this https://github.com/alpinelinux/aports/blob/master/main/mariadb/mariadb.initd#L11 will not work
2016-05-26 12:12:03 <clandmeter> yes thats the fix ncopa made
2016-05-26 12:12:48 <clandmeter> it fixes the initial issue, but introduces another.
2016-05-26 12:12:53 <jirutka> it’s easy to fix it, not needed to revert to the previous one IMHO
2016-05-26 12:13:17 <jirutka> this https://github.com/alpinelinux/aports/blob/master/main/mariadb/mariadb.initd#L11 is after the “fix” or before?
2016-05-26 12:13:43 <clandmeter> the problem is that openrc will read all init files for deps
2016-05-26 12:13:50 <jirutka> I know
2016-05-26 12:13:57 <clandmeter> so all logic should be enclosed in fucntions
2016-05-26 12:13:58 <ncopa> http://tpaste.us/2Kag
2016-05-26 12:14:37 <jirutka> ncopa: this is also wrong, because pidfile is needed also for stopping
2016-05-26 12:14:48 <jirutka> just add this to pre_stop function too
2016-05-26 12:14:53 <jirutka> *stop_pre
2016-05-26 12:15:15 <jirutka> ah, pardon
2016-05-26 12:15:22 <ncopa> stop_pre already grab the pid file
2016-05-26 12:15:43 <jirutka> nope, it’s indeed in start_pre only
2016-05-26 12:16:05 <ncopa> stop_pre() {
2016-05-26 12:16:05 <ncopa>         pidfile=$(getconf pid-file "/run/mysqld/mysqld.pid")
2016-05-26 12:16:05 <ncopa> }
2016-05-26 12:16:33 <jirutka> aha, and command_args is not needed for stop()
2016-05-26 12:16:43 <jirutka> I should read it more carefully
2016-05-26 12:16:45 <ncopa> should not be needed
2016-05-26 12:16:54 <jirutka> so where’s the problem now?
2016-05-26 12:17:02 <ncopa> clandmenter can you test this: http://tpaste.us/2Kag
2016-05-26 12:20:09 <jirutka> ncopa: I think that it’d be better to `return 1` in https://github.com/alpinelinux/aports/blob/master/main/mariadb/mariadb.initd#L29 to halt the starting
2016-05-26 12:20:53 <clandmeter> ncopa: i dont have it installed here, maybe leo-unglaub can test it?
2016-05-26 12:21:21 <ncopa> i think we need pidfile for status too?
2016-05-26 12:23:18 <jirutka> ncopa: do you know that `required_dirs` is a “special” variable?
2016-05-26 12:24:26 <ncopa> what does it do?
2016-05-26 12:24:38 <jirutka> required_dirs : A list of directories which must exist for the service to start.
2016-05-26 12:24:40 <jirutka> man…
2016-05-26 12:24:47 <jirutka> there’s also required_files
2016-05-26 12:25:37 <jirutka> so you tells openrc-run to check this directory and simultaneously you check if by hand
2016-05-26 12:26:01 <jirutka> I’m not sure if it’s an intention or not
2016-05-26 12:26:11 <ncopa> probably not
2016-05-26 12:26:15 <ncopa> but it should not hurt
2016-05-26 12:26:39 <jirutka> then it’d be IMO better to rename it to e.g. `datadir` and use `local`
2016-05-26 12:27:35 <jirutka> it should not hurt, but it’s confusing
2016-05-26 12:34:44 <jirutka> ncopa: why you want let mariadb to create a pidfile instead of start-stop-daemon anyway?
2016-05-26 12:36:22 <jirutka> I personally always prefer to create pidfile by runscript instead of relaying to the program being started to do that, it’s often more complicated and less reliable
2016-05-26 12:38:48 <leo-unglaub> ncopa: what do you need tested?
2016-05-26 12:40:51 <jirutka> leo: this http://tpaste.us/2Kag
2016-05-26 12:41:47 <jirutka> whether start, status, and stop works correctly
2016-05-26 12:47:00 <leo-unglaub> ncopa: jirutka: the patch works
2016-05-26 12:47:09 <leo-unglaub> but only for newly started instances
2016-05-26 12:47:23 <leo-unglaub> i could not stop the already running instance
2016-05-26 12:48:41 <jirutka> that’s expected
2016-05-26 12:48:54 <jirutka> have you tested status?
2016-05-26 12:49:08 <jirutka> when mariadb is running and when it’s stopped?
2016-05-26 12:49:37 <leo-unglaub> yes
2016-05-26 12:49:38 <jirutka> and to be sure, also when it’s crashed (kill it manully)
2016-05-26 12:49:40 <leo-unglaub> works both
2016-05-26 12:50:23 <leo-unglaub> status is crashed
2016-05-26 12:50:29 <leo-unglaub> when i kill it manually
2016-05-26 12:50:37 <jirutka> interesting, the status function in OpenRC is still quite a mystery for me
2016-05-26 12:51:59 <leo-unglaub> hmm, wait
2016-05-26 12:52:05 <leo-unglaub> after it crashed i cannot start it anymore
2016-05-26 12:52:15 <jirutka> it seems that it doesn’t use a pidfile, so it must somehow tracks processes started by start-stop-daemon… but if it’s correct, then why openrc use pidfile at all?
2016-05-26 12:52:27 <jirutka> use zap
2016-05-26 12:52:27 <leo-unglaub> batdesk:/etc/hiawatha# /etc/init.d/mariadb status
2016-05-26 12:52:27 <leo-unglaub>  * status: crashed
2016-05-26 12:52:27 <leo-unglaub> batdesk:/etc/hiawatha# /etc/init.d/mariadb start
2016-05-26 12:52:27 <leo-unglaub> batdesk:/etc/hiawatha# /etc/init.d/mariadb start
2016-05-26 12:52:27 <leo-unglaub>  * WARNING: mariadb has already been started
2016-05-26 12:52:27 <leo-unglaub>  * WARNING: mariadb has already been started
2016-05-26 12:52:27 <leo-unglaub> batdesk:/etc/hiawatha# /etc/init.d/mariadb status
2016-05-26 12:52:28 <leo-unglaub>  * status: crashed
2016-05-26 12:52:39 <jirutka> aha, just stop it
2016-05-26 12:53:15 <jirutka> it should work… also you can use `zap` to just switch the status to stopped
2016-05-26 12:56:03 <leo-unglaub> works
2016-05-26 12:56:08 <leo-unglaub> i think the patch can be mergerd
2016-05-26 12:56:45 <jirutka> thanks!
2016-05-26 13:12:10 <leo-unglaub> clandmeter: i am really happy with your pkgs site
2016-05-26 13:12:13 <leo-unglaub> i really like it
2016-05-26 13:12:36 <leo-unglaub> i think your next project should be http://bugs.alpinelinux.org/projects/alpine/issues :)
2016-05-26 13:27:07 <jirutka> me too, Alpine sites are very neat!
2016-05-26 13:49:44 <fabled> ncopa, ok to cherry-pick http://lists.busybox.net/pipermail/busybox/2016-May/084197.html ?
2016-05-26 13:55:56 <ncopa> fabled: yes
2016-05-26 14:00:23 <jirutka> ncopa: have you looked at https://github.com/jirutka/qemu-openrc/commit/1728e404e4e5df9afd156ffc640cacd9e383e2c4 ?
2016-05-26 14:10:27 <ncopa> i looked at it
2016-05-26 14:10:29 <ncopa> it looks ok
2016-05-26 14:10:32 <ncopa> but i havent tested it
2016-05-26 14:21:38 <leo-unglaub> ncopa: what are you thoughts on making php more safe by default?
2016-05-26 14:25:37 <ncopa> how?
2016-05-26 14:26:49 <leo-unglaub> limit "max_input_vars, post_max_size" and disable "expose_php, display_startup_errors, display_errors"
2016-05-26 14:28:27 <leo-unglaub> we could ship a /etc/php/conf.d/alpine.ini with some good defaults
2016-05-26 14:29:21 <ncopa> i like the idea of tightening up the default config
2016-05-26 14:29:32 <ncopa> but i dont know if like alpine.ini
2016-05-26 14:30:22 <ncopa> i think a patch to modify the default config would be nice
2016-05-26 14:31:10 <leo-unglaub> hmmm, maybe
2016-05-26 14:31:18 <leo-unglaub> the problem is that this php.ini is a huge file
2016-05-26 14:31:27 <leo-unglaub> and finding changes in there is always hard
2016-05-26 14:31:49 <leo-unglaub> i am going to search some settigns we should set and  then we can desice how we apply them
2016-05-26 14:51:58 <ncopa> oh no
2016-05-26 14:52:02 <ncopa> someone pushed make 4.2
2016-05-26 14:53:19 <ncopa> nmeum: i'll revert make 4.2 upgrade til after v3.4 release
2016-05-26 14:53:29 <ncopa> we have build the entire world and the release is right around the corner
2016-05-26 14:53:49 <ncopa> i *dont* want spen another week rebuilding world, fixing all potensial build issues that make 4.2 introduced
2016-05-26 14:53:54 <ncopa> i want release
2016-05-26 14:54:22 <ncopa> make is nasty
2016-05-26 14:54:41 <ncopa> i need to manually downgrade on build servers
2016-05-26 14:56:08 <ncopa> please no more updates of build tools, like binutils, gcc make
2016-05-26 14:56:12 <ncopa> til after v3.4
2016-05-26 14:56:43 <skarnet> what do you mean, make is nasty?
2016-05-26 14:56:50 <ncopa> i mean
2016-05-26 14:56:59 <ncopa> we built world with make 4.1
2016-05-26 14:57:05 <ncopa> make 4.2 may change behaviour
2016-05-26 14:57:17 <skarnet> yeah, sure
2016-05-26 14:57:19 <ncopa> so if we push make 4.2 after world is built
2016-05-26 14:57:23 <skarnet> stick with 4.1 for now
2016-05-26 14:57:26 <ncopa> then we dont know if things builds
2016-05-26 14:57:48 <ncopa> its not like updating a minor version of some runtime lib
2016-05-26 14:58:24 <ncopa> the only reason i discovered this was becase git failed to build
2016-05-26 14:58:31 <ncopa> luckily it did
2016-05-26 14:58:36 <ncopa> or i wouldnt have noticed
2016-05-26 14:58:38 <skarnet> is it vital that you ship the same version of dev tools you used to build the production apks?
2016-05-26 14:58:47 <ncopa> well
2016-05-26 14:58:56 <ncopa> what happens if i need push a security fix in future
2016-05-26 14:59:03 <ncopa> and the package fails to build for other reasons?
2016-05-26 14:59:25 <skarnet> that doesn't answer my question
2016-05-26 14:59:49 <ncopa> it is important that build tools are similar enough
2016-05-26 14:59:54 <skarnet> the implication of your answer is that you build on a production machine
2016-05-26 15:00:15 <skarnet> prod environment and dev environment don't have to be the same
2016-05-26 15:00:40 <skarnet> it's obviously best if you dogfood, but it's not an absolute requirement
2016-05-26 15:00:50 <ncopa> i do build sec fixes for stable branches on a machine with that alpine version
2016-05-26 15:01:02 <ncopa> so for v3.3 sec fixes i use a v3.3 dev machine
2016-05-26 15:01:16 <skarnet> which could by definition have been compromised by the sec issue :P
2016-05-26 15:02:02 <ncopa> if sec issue was in the build tools or base tools
2016-05-26 15:02:05 <ncopa> most sec issues are not
2016-05-26 15:02:10 <skarnet> true
2016-05-26 15:02:58 <ncopa> ok
2016-05-26 15:03:07 <ncopa> downgrading make to 4.1 make git build again
2016-05-26 15:03:27 <skarnet> this should be reported to make
2016-05-26 15:03:33 <ncopa> or git
2016-05-26 15:03:38 <skarnet> make
2016-05-26 15:03:47 <skarnet> bumping a minor version number should not break compatibility
2016-05-26 15:04:02 <skarnet> if stuff that builds with 4.1 breaks with 4.2, then it should be named 5.0, not 4.2
2016-05-26 15:04:07 <ncopa> i dont know if they consider 4.1 -> 4.2 minor
2016-05-26 15:04:11 <ncopa> its GNU...
2016-05-26 15:04:17 <ncopa> so i dont have expectations :)
2016-05-26 15:04:27 <skarnet> then they should call it 4.1.0 and 4.2.0
2016-05-26 15:04:27 <ncopa> nmeum: care to file bug to upstream gnu make?
2016-05-26 15:04:50 <ncopa> nmeum: to reproduce try compile git on a machine with many cores
2016-05-26 15:05:04 <ncopa> with make 4.2 it fails and with make 4.1 it works
2016-05-26 15:14:42 <nmeum> ncopa: Sorry, I wasn't sure if it would be a good idea to upgrade GNU make before the 3.4 release, shouldn't have done it in the first place
2016-05-26 15:15:33 <nmeum> regrading the bug you found: I will try to reproduce it and will report it upstream
2016-05-26 15:18:20 <ncopa> what is left now for the v3.4 release?
2016-05-26 15:18:31 <ncopa> leo-unglaub: thank you for the mariadb report
2016-05-26 15:18:46 <ncopa> that kind of thing is exactly what i want focus on at this point
2016-05-26 15:18:51 <ncopa> fixing bugs
2016-05-26 15:18:55 <ncopa> finding and fixing bugs
2016-05-26 15:19:29 <leo-unglaub> sure, thanks for fixing it
2016-05-26 15:20:24 <ncopa> http://bugs.alpinelinux.org/projects/alpine/issues?c[]=project&c[]=tracker&c[]=status&c[]=priority&c[]=subject&c[]=assigned_to&c[]=updated_on&f[]=status_id&f[]=fixed_version_id&f[]=&group_by=&op[fixed_version_id]=%3D&op[status_id]=o&per_page=100&set_filter=1&utf8=%E2%9C%93&v[fixed_version_id][]=102
2016-05-26 15:21:06 <ncopa> any volunteer to work on: http://bugs.alpinelinux.org/issues/5462?
2016-05-26 15:21:13 <leo-unglaub> ncopa: could you please apply this patch as well? https://paste.debian.net/plain/705954
2016-05-26 15:21:20 <leo-unglaub> it makes hiawatha start again
2016-05-26 15:21:31 <ncopa> when upgrading from v3.3 to v3.4, we want cron work properly
2016-05-26 15:21:55 <ncopa> i think cron will not auto start after reboot after upgrading to v3.4
2016-05-26 15:22:09 <nmeum> ncopa: I will take a look at it
2016-05-26 15:22:21 <ncopa> leo-unglaub: can we change it to "after firewall"?
2016-05-26 15:23:08 <leo-unglaub> i have no "firewall" on this computer
2016-05-26 15:23:14 <leo-unglaub> so it fails if it does not find it
2016-05-26 15:23:17 <ncopa> http://tpaste.us/AjJn
2016-05-26 15:23:25 <ncopa> should not with 'after'
2016-05-26 15:24:04 <ncopa> it means, if you have a firewall service and its enabled, then make sure hiawatha is started after
2016-05-26 15:24:29 <ncopa> leo-unglaub: can you please test http://tpaste.us/AjJn
2016-05-26 15:24:41 <ncopa> we should also consider moving it to community
2016-05-26 15:25:43 <leo-unglaub> ncopa: the patch worls
2016-05-26 15:25:45 <leo-unglaub> works
2016-05-26 15:28:12 <leo-unglaub> i have no problem with that
2016-05-26 15:46:44 <leo-unglaub> the best way to find bugs is to install alpone on your girlfriends laptop .. because then you see all the bugs you normally ignore because you can work around it
2016-05-26 15:47:09 <nmeum> ncopa: here is a quick and dirty solution for #5462 http://sprunge.us/HNMK
2016-05-26 15:47:33 <nmeum> didn't test it yet but will do so if you like the dirty /tmp hack to do communication between the post- and pre-upgrade files
2016-05-26 15:50:03 <ncopa> nmeum: i was thinking something like: mv /run/openrc/started/cron /run/openrc/started/crond
2016-05-26 15:50:40 <ncopa> and:
2016-05-26 15:51:04 <ncopa> for i in /etc/runlevels/*/cron; do mv $i ${i}d; done
2016-05-26 15:51:10 <ncopa> testing if it exists first
2016-05-26 15:52:21 <nmeum> is it really a good idea to tamper with files in /run/openrc manually?
2016-05-26 15:58:02 <nmeum> if you implement it your way you will at least also have to move crons pidfile in /run and maybe other files I don't know about as well…
2016-05-26 17:56:44 <mattaitchison> ncopa: Seems like  edge:make-4.1r0 has a bad sig. https://github.com/gliderlabs/docker-alpine/issues/177
2016-05-26 19:20:41 <ncopa> mattaitchison: yes. i had to downgrade it
2016-05-26 19:20:54 <ncopa> so it got rebuilt
2016-05-26 19:25:55 <barthalion> why the sig is wrong then?
2016-05-26 19:26:07 <ncopa> i downgraded to 4.1
2016-05-26 19:26:14 <ncopa> this led to a rebuild of the package
2016-05-26 19:26:23 <ncopa> the fastly cache had the old binary cached
2016-05-26 19:26:32 <ncopa> apkindex had new checksum
2016-05-26 19:27:17 <ncopa> i suppose some timestamp in the apk was changed
2016-05-26 19:27:35 <ncopa> oh, we also store the git commit
2016-05-26 19:27:56 <ncopa> the correct fix is to bump pkgrel, which i just did
2016-05-26 20:22:27 <ncopa> can somebody help me with: http://bugs.alpinelinux.org/issues/5614
2016-05-26 20:22:45 <ncopa> just make sure the php module put the config in /etc/php5 instead of /etc/php
2016-05-26 20:29:26 <ncopa> #5593
2016-05-26 22:23:40 <nmeum> ncopa: any comments on the stuff I wrote above regarding the crond {pre,post}-upgrade script?
2016-05-26 22:24:01 <nmeum> also: is 9692e009385e05d9f60f259bc7d04713acb76727 still needed after the make downgrade or do you want to keep it around anyhow?
2016-05-26 23:27:33 <jirutka> ncopa?
2016-05-27 06:07:32 <fabled> ncopa, i'm thinking to move libfastjson to main and apply the rsyslog upgrade. is that ok, or you think they should go to community?
2016-05-27 06:10:49 <fabled> i guess i just update it. we can move it to community later if needed.
2016-05-27 06:15:33 <fabled> oh, it failed to build
2016-05-27 06:17:30 <fcolista__> package from main can pickup dependecies from community?
2016-05-27 06:17:35 <barthalion> nope
2016-05-27 06:17:37 <fcolista__> k
2016-05-27 06:17:38 <fcolista__> thx
2016-05-27 06:18:20 <fcolista__> py-mock dependencies are broken
2016-05-27 06:18:36 <fcolista__> that in turns cause certbot/letsencrypt to not work
2016-05-27 06:18:53 <fcolista__> i packagesd the missing dependencies, py-pbr and py-funcsigs
2016-05-27 06:19:00 <fcolista__> i'm going to move them to main
2016-05-27 08:09:41 <vkris> /usr/share/vim/vim74/tutor takes some 2mb~, dont't know, but can it be pkgd separtely ?
2016-05-27 08:11:45 <vkris> in vim-doc
2016-05-27 08:14:01 <ScrumpyJack> anyone know how to update a kernel from a package on the rpi? (modloop and squashfs are involved)
2016-05-27 09:00:52 <vkris> bump testing/h2o to v1.7.3 pls
2016-05-27 09:02:30 <ScrumpyJack> vkris: are you not near a alpine linux sdk dev environment? :)
2016-05-27 09:03:19 <vkris> not quite, anything important ?
2016-05-27 09:04:42 <ScrumpyJack> vkris said: "bump testing/h2o to v1.7.3 pls" <- this :)
2016-05-27 09:05:25 <vkris> coz of this https://github.com/h2o/h2o/pull/920
2016-05-27 09:08:45 <vkris> does not look like major change, http://blog.kazuhooku.com/search/label/H2O
2016-05-27 09:09:00 <ScrumpyJack> yes, sure, but you could bump it yourself if you're near your git repo with alpine-sdk installed no? ;)
2016-05-27 09:09:07 <ScrumpyJack> hense my question :)
2016-05-27 09:09:46 <vkris> sure !! ncopa ^^
2016-05-27 09:12:16 <ScrumpyJack> it's ok, i've done it
2016-05-27 09:15:38 <vkris> thanks
2016-05-27 09:20:19 <fabled> ScrumpyJack, update-kernel -a armhf -f rpi2 /path/to/boot
2016-05-27 09:27:02 <ncopa_> i am shutting down the build-3-4-armhf for a while
2016-05-27 09:27:07 <ncopa_> and build-3-3-armhf
2016-05-27 09:27:26 <ncopa_> will replace the usb disk with sata
2016-05-27 09:29:13 <ScrumpyJack> update-kernel gives me. ERROR: Failed to create lib/firmware/bnx2x/bnx2x-e2-6.0.34.0.fw: No space left on device
2016-05-27 09:29:32 <ScrumpyJack> it's trying to write to modloop
2016-05-27 09:30:00 <fabled> the error tells it all
2016-05-27 09:30:18 <fabled> it extracts stuff to /tmp when creating the update images
2016-05-27 09:30:23 <fabled> so your /tmp is not large enough
2016-05-27 09:30:25 <ScrumpyJack> ah ok
2016-05-27 09:30:52 <fabled> we'll probably get 3.4.0 or at least next _rc in few days time
2016-05-27 09:31:45 <ScrumpyJack> can i pass a different tmp folder? I'll check
2016-05-27 09:32:15 <ScrumpyJack> TMPDIR=/somewhere else
2016-05-27 11:39:50 <jirutka1> hi
2016-05-27 11:43:13 <jirutka1> ncopa: these patches are IMO ready to merge:
2016-05-27 11:43:13 <jirutka1> * https://github.com/alpinelinux/aports/pull/84.patch (by little-dude)
2016-05-27 11:43:13 <jirutka1> * http://tpaste.us/Gx6D (rebased https://github.com/alpinelinux/aports/pull/87 by AmatCoder)
2016-05-27 11:43:13 <jirutka1> * https://github.com/alpinelinux/aports/pull/89.patch (by AmatCoder)
2016-05-27 15:40:17 <ncopa> i drove to the server room where arm builder is today
2016-05-27 15:40:25 <ncopa> and replaced the usb disk with sata connected ssd
2016-05-27 15:40:34 <ncopa> should be significantly faster now
2016-05-27 16:21:13 <skarnet> "I replaced the arm builders with PCs, should be even faster!"
2016-05-28 06:53:42 <vkris> ncopa, any plans to support flashplugin in firefox ?
2016-05-28 10:15:01 <jirutka> vkris: Flash?! Do you really mean Adobe Flash, that piece of sh*t full of critical security vulnerabilities?
2016-05-28 11:27:12 <vkris> I am sure bbc only "gathers" news, and disperses whatever it gathers
2016-05-28 11:27:12 <vkris> or is it "home news" too..., why cannot they switch to alternative if browser supports it ?
2016-05-28 11:27:15 <vkris> this switch has to start at top level "good" organizations to set examples
2016-05-28 18:55:50 <ByteFlinger> Evening. Does anybody know how does one create a .apk package? I have tried using the alpine wiki (https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package) with one of the aports APKBUILD files but running "abuild -r" just builds and directly installs the compile binary
2016-05-28 18:56:10 <jirutka> hi
2016-05-28 18:56:20 <ByteFlinger> I'd like to create a .apk package like .deb but for alpine instead of installing the binaries on the host I am compiling
2016-05-28 18:56:23 <jirutka> abuild -r does not _install_ the package, it just builds .apk
2016-05-28 18:56:32 <jirutka> by default, it’s stored in ~/packages
2016-05-28 18:56:39 <ByteFlinger> Aha, let me take a look
2016-05-28 18:57:20 <ByteFlinger> Doh, yeah I found it there
2016-05-28 18:57:24 <ByteFlinger> Thank you
2016-05-28 18:58:13 <jirutka> you’re welcome
2016-05-28 18:58:25 <ByteFlinger> Is there any more official arm docker container other than "experimental" armhf one?
2016-05-28 18:58:56 <jirutka> I dunno, I don’t use Docker, I use just *real* containers ;)
2016-05-28 18:59:17 <ByteFlinger> Like?
2016-05-28 18:59:24 <jirutka> LXC and OpenVZ
2016-05-28 18:59:50 <ByteFlinger> Ah.
2016-05-28 19:00:03 <ByteFlinger> You mean "those" then :)
2016-05-28 19:01:18 <ByteFlinger> But I guess you are not deploying it on a raspberry pi
2016-05-28 19:01:59 <jirutka> nope, I’ve never tried it on Raspberry, but LXC should work even there
2016-05-28 19:03:20 <ByteFlinger> Yeah, docker is working fine, just need the arm stuff from alpine. The docker armhf is working OK so far, was just curious whether there was a more official one
2016-05-28 20:16:55 <barthalion> ByteFlinger: I don't think there is another one
2016-05-28 21:02:46 <ByteFlinger> barthalion: Yeah I kind of figured. I'll go with that one for now. Thanks
2016-05-29 10:59:33 <kaiyou_> Hey everyone, I am having some doubts about best practices and cant find any documentation
2016-05-29 11:00:28 <kaiyou_> I just submitted this http://patchwork.alpinelinux.org/patch/2055/ but am feeling like both system users could actually be attached to 'mail' instead of specific system groups
2016-05-29 11:00:40 <kaiyou_> Is there any documentation on the topic that I did not find?
2016-05-29 11:41:02 <ncopa> vkris: adobe flash works in ff
2016-05-29 12:15:36 <yth> What's the simplest way to build an edge modloop for rpi?
2016-05-29 13:11:11 <vkris> ncopa: http://irclogger.com/.alpine-linux/2016-05-28#1464418410
2016-05-29 13:11:57 <vkris> don't know if this is a ff's plugin_container issue or libflash ?
2016-05-29 13:14:43 <vkris> wonder if this effects tor in someway, though i don't use that application
2016-05-29 13:30:31 <ncopa> vkris: sorry. i dont understand what you are are trying to do, nor do i understadn what problem you have
2016-05-29 13:31:16 <ncopa> i understand that you have some kind of problem with flash in firefox, but not more that that
2016-05-29 13:33:03 <ncopa> i have no idea what kind of weird things you want to see with flash plugin
2016-05-29 13:45:19 <vkris> I don't know, just thought to let know the issue I noticed earlier with libflash, if its a bug I can report it upstream
2016-05-29 13:45:53 <vkris> anyway sorry bothering, would try on ff channel
2016-05-29 14:08:12 <vkris> why nothing weird, just some bbc news, any suggestions ;)
2016-05-29 15:21:41 <jirutka> ncopa, could I ask you for merging few patches? :)
2016-05-29 15:22:11 <oneinsect> folks...does our x86 distro work with u-boot
2016-05-29 15:22:55 <oneinsect> i am trying to get it working with u-boot however i am getting kernel panic...saying Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
2016-05-29 15:22:58 <oneinsect> any ideas?
2016-05-29 15:23:29 <oneinsect> i think its based on syslinux?
2016-05-29 15:59:24 <vkris> QEMU version 2.6.0 is out !
2016-05-29 15:59:42 <jirutka> what’s new in this version?
2016-05-29 16:00:02 <vkris> "Handling I/O Bursts"
2016-05-29 16:00:22 <vkris> http://www.phoronix.com/scan.php?page=news_item&px=QEMU-2.6-IO-Bursts
2016-05-29 16:15:04 <vkris> ah ! I am tied of 6 cross-domain clicks just to flag a pkg on pkgs.a.o
2016-05-29 16:15:23 <jirutka> ?
2016-05-29 16:15:41 <barthalion> wut
2016-05-29 16:16:14 <vkris> tired*
2016-05-29 16:16:59 <vkris> pkgs.a.o uses google captcha, that requires users to click 4+ times to get through it
2016-05-29 16:17:04 <vkris> and that too on images
2016-05-29 16:17:54 <vkris> I see some amount of dubiousness in this
2016-05-29 16:18:42 <barthalion> you do that once for rather long time
2016-05-29 16:18:45 <vkris> today its is "donut" tomorrow who knows ?
2016-05-29 16:18:50 <barthalion> wat
2016-05-29 16:18:59 <barthalion> sorry for us not wanting spam
2016-05-29 16:19:53 <vkris> there are 1/2 a dozen simpler number method, even yandex uses a nice obfuscated number based captcha
2016-05-29 16:20:20 <vkris> I don't think 1/2 a dozen clicks in needed to pass a captcha
2016-05-29 16:20:44 <leo-unglaub> hey friends
2016-05-29 16:21:28 <jirutka> hi
2016-05-29 16:21:45 <barthalion> vkris: and that simpler methods are too trivial
2016-05-29 16:21:54 <skazz> I only get asked to do recaptcha stuff if I am using somthing like Tor, only other reasons are things like a previously spam IP address, shared NAT network where some device is spamming (like teathering), or something on your own network that has spammed google at some point, etc.
2016-05-29 16:21:54 <barthalion> you verify yourself once, then you need one click to get through it
2016-05-29 16:22:20 <barthalion> and usually you don't have to do that once unless your IP has been used for spam
2016-05-29 16:22:26 <skazz> ^
2016-05-29 16:23:15 <barthalion> show us superior and simpler method that is so widely used and causes no headache to anyone including you
2016-05-29 16:23:16 <vkris> why use google script for that ?
2016-05-29 16:23:36 <barthalion> I already said why
2016-05-29 16:23:44 <vkris> are they paying or is it being used for gather traffic ?
2016-05-29 16:23:55 <barthalion> 
ACTION sighs

2016-05-29 16:23:58 <barthalion> this is pointless
2016-05-29 16:29:01 <barthalion> https://paste.xinu.at/jKy0C/ doesn't seem difficult, does it?
2016-05-29 16:31:45 <leo-unglaub> 
ACTION is also not a big fan of that captha

2016-05-29 16:31:52 <leo-unglaub> but i hate all captchas
2016-05-29 16:33:51 <vkris> well, I am not flgging till its simple
2016-05-29 16:37:34 <leo-unglaub> barthalion: when i do it i have to do way more than you
2016-05-29 16:37:47 <leo-unglaub> i get 9 images and have to select all images that show "flowers"
2016-05-29 16:38:02 <barthalion> sure
2016-05-29 16:38:04 <barthalion> *once*
2016-05-29 16:38:05 <leo-unglaub> or "sky" and that is fucking stupid because sometimes you are not sure if it counts, ...
2016-05-29 16:38:12 <leo-unglaub> no, every fucking time i do it
2016-05-29 16:38:20 <skazz> You on a static IP address leo?
2016-05-29 16:38:24 <barthalion> sorry for your ISP then
2016-05-29 16:38:35 <leo-unglaub> in austria you dont get a static ip
2016-05-29 16:38:42 <leo-unglaub> unless you pay more
2016-05-29 16:38:42 <barthalion> hum
2016-05-29 16:38:44 <skazz> Or a near-static IP which only renews if your router looses power?
2016-05-29 16:39:10 <barthalion> still, you can't eat cake and have cake
2016-05-29 16:39:18 <leo-unglaub> the number of countries where you get a static ip by default is very low
2016-05-29 16:39:30 <leo-unglaub> most countries rotate there ipv4 adresses because they dont have enought
2016-05-29 16:39:32 <leo-unglaub> like germany
2016-05-29 16:39:38 <leo-unglaub> they rotate it all the time
2016-05-29 16:39:53 <leo-unglaub> thats why that captcha is horrible ...
2016-05-29 16:40:11 <leo-unglaub> i never flag stuff .... i just ping ncopa here in the irc because its easier that this captcha
2016-05-29 16:41:35 <vkris> leo-unglaub: it flowers, then ferns than hashis ...
2016-05-29 16:42:12 <vkris> and that gets clicked and impressions in your brain
2016-05-29 16:42:20 <vkris> impressioned*
2016-05-29 16:42:49 <vkris> get the bigger picture now !!
2016-05-29 16:43:08 <leo-unglaub> ???
2016-05-29 16:44:04 <vkris> then its different kinds of store-fronts, then foods that sell
2016-05-29 16:44:10 <vkris> still not getting it ?
2016-05-29 16:44:45 <vkris> and you click on those images
2016-05-29 16:44:54 <leo-unglaub> its some kind of a drug joke, but i dont get the connection to the captchas
2016-05-29 16:46:05 <vkris> its mild form of ads
2016-05-29 16:46:14 <leo-unglaub> ah, now it get it
2016-05-29 16:46:21 <vkris> gee !!!!
2016-05-29 16:46:42 <barthalion> mild form of ads? you are really overreacting
2016-05-29 16:47:39 <vkris> barthalion: get some nice books on human succeptibility, coginitative and impressionative behaviourism
2016-05-29 16:48:02 <vkris> its basically psychology books
2016-05-29 16:48:21 <vkris> don't just still to your laptop
2016-05-29 16:48:38 <vkris> and codes
2016-05-29 16:49:17 <barthalion> I'm fine with my books, thank you
2016-05-29 16:49:34 <vkris> ;)
2016-05-29 16:51:18 <vkris> it doesn't have to be cross-domain and 6clicks
2016-05-29 16:51:43 <vkris> hire developers if they cannot or do not have one
2016-05-29 16:51:53 <vkris> plain and simple
2016-05-29 16:51:58 <barthalion> err, no, it doesn't work like that
2016-05-29 16:52:01 <barthalion> you are complaining
2016-05-29 16:52:10 <barthalion> you are supposed to suggest alternative
2016-05-29 16:52:57 <barthalion> it's not 6 clicks, except for you, but that's not our fault anyhow
2016-05-29 16:53:52 <barthalion> and sunday, 7PM CET is not the time any of website-resposinble devs are hanging around here
2016-05-29 16:55:01 <vkris> barthalion: I know, I am not asking to make changes immediately
2016-05-29 16:55:39 <barthalion> you are not proposing anything except trying to convince everyone that captcha is trying to wash your brain
2016-05-29 16:55:39 <vkris> I would let clandmeter know at work hours of my opinion, (logs are avilable)
2016-05-29 16:57:41 <vkris> I don't care its ads or not, today it made me click 5 different "Store fronts, with name"
2016-05-29 16:58:28 <vkris> also try getting some info on PPC
2016-05-29 16:58:32 <jirutka> I agree with vkris that captchtas are horrible
2016-05-29 16:59:51 <jirutka> I have a suggestions – centralize user accounts for Redmine, wiki, … and allow log-in to pkgs.a.o, then at least people with account will be saved from captchta
2016-05-29 17:00:21 <vkris> and if someone wants my suggestion on captchas can pay me
2016-05-29 17:00:34 <vkris> even its temporary hiring
2016-05-29 17:00:46 <vkris> I am free for last 6yrs
2016-05-29 17:01:17 <jirutka> I also rerely flag stuff, I rather upgrade the package and open a pull request :)
2016-05-29 17:01:51 <jirutka> afk
2016-05-30 09:11:49 <leo-unglaub> hey :(
2016-05-30 09:11:53 <leo-unglaub> wrong smilie
2016-05-30 09:11:55 <leo-unglaub> hey :)
2016-05-30 09:13:06 <clandmeter> hallo
2016-05-30 09:52:09 <jirutka> hi
2016-05-30 11:42:47 <jirutka> ncopa: please look at PRs on GitHub, there are 14 of them ready to be merged https://github.com/alpinelinux/aports/pulls?q=is%3Aopen+is%3Apr+label%3Aready
2016-05-30 11:42:47 <jirutka> you should probably take an attention to https://github.com/alpinelinux/aports/pull/91 (moving qemu-openrc to main), https://github.com/alpinelinux/aports/pull/96 (ruby2.1) and https://github.com/alpinelinux/aports/pull/97 (ruby2.2)
2016-05-30 15:53:39 <leo-unglaub> hey :)
2016-05-30 15:58:31 <leo-unglaub> we really have to teach people here on how to use a mailinglist. i got 4 responses now to my question but none of those to the mailinglist, all to my personal address
2016-05-30 16:06:38 <ncopa> heh
2016-05-30 16:07:28 <leo-unglaub> but in generall people seam to agree with the idea ... but they would prefer ~/bin instead of ~/.bin if i get it correctly
2016-05-30 16:12:09 <leo-unglaub> brb reboot
2016-05-30 16:14:57 <leo-unglaub> ncopa: maybe you can explain this to me: when i place a script into /etc/profiles.d and login to TTY1 it gets sourced and all aliases in there are available but when i do a startxfce4 all aliases are gone again
2016-05-30 16:15:02 <leo-unglaub> do you know why this happens?
2016-05-30 16:15:57 <jirutka> leo: maybe you should add address of the mailing list to Reply-To?
2016-05-30 16:15:59 <skarnet> shells in your xterms aren't login shells, I gyess
2016-05-30 16:16:32 <jirutka> why is there your mail address if you want us to response to the mailing list?
2016-05-30 16:16:35 <skarnet> jirutka: no, but the list should set a Mail-Followup-To header if it isn't the case already
2016-05-30 16:17:07 <skarnet> http://marc.merlins.org/netrants/reply-to-harmful.html
2016-05-30 16:17:19 <jirutka> this is not the point
2016-05-30 16:17:29 <jirutka> the point is that it’s not people fault
2016-05-30 16:17:48 <leo-unglaub> jirutka: you never set the reply-to header for mailinglists
2016-05-30 16:17:48 <skarnet> it's exactly the point. It *is* people's fault if they can't hit their "reply-to-all" button on their mailer.
2016-05-30 16:18:12 <jirutka> I’ve replied from my mobile phone…
2016-05-30 16:18:24 <skarnet> Your mail app should still have such a button
2016-05-30 16:18:29 <jirutka> I don’t even know if there’s Reply To All on my mobile mail client
2016-05-30 16:18:36 <skarnet> if there isn't, get a better mail client
2016-05-30 16:19:11 <jirutka> ah, there is Reply To All, I’ve just overlooked it when writing the response from bus
2016-05-30 16:19:23 <skarnet> QED
2016-05-30 16:19:27 <leo-unglaub> jirutka: dont worry, its not the end of the world ;)
2016-05-30 16:20:08 <leo-unglaub> skarnet: hmm, maybe you are right but when the aliases are already loaded, whouldn't my terminal in an x session extend from the existing context?
2016-05-30 16:21:11 <skarnet> I don't think so, because shells in your session aren't considered children of your starting shell, or something like that.
2016-05-30 16:21:54 <skarnet> My memory is fuzzy, and may be wrong because I haven't done X in a long while, but IIRC when you startx there's a lot of env cleaning happening
2016-05-30 16:22:30 <skarnet> ISTR the point being that startx and some form of xdm should behave the same
2016-05-30 16:22:49 <skarnet> and when you log in via xdm, you don't get your login shell environment
2016-05-30 16:23:19 <skarnet> this may all be bullshit, but this would explain it
2016-05-30 16:23:41 <skarnet> X is a weird beast anyway
2016-05-30 16:23:49 <leo-unglaub> hmmm, yours makes sence to me so i think it has to be true
2016-05-30 16:23:57 <leo-unglaub> but yeah, X is always confusing as hell
2016-05-30 16:34:59 <leo-unglaub> ncopa: would you be okay with adding "who" to busybox?
2016-05-30 16:37:55 <ncopa> i thought we had it
2016-05-30 16:38:07 <ncopa> is it a config option in busybox?
2016-05-30 16:41:37 <leo-unglaub> ncopa: currently the file contains: # CONFIG_WHO is not set
2016-05-30 16:41:45 <leo-unglaub> i think thats the correct option
2016-05-30 16:42:22 <ncopa> does git blame say anything on why it was set to that?
2016-05-30 16:43:04 <leo-unglaub> i think its the default if no one has set it yet
2016-05-30 16:43:40 <ncopa> commit e99310f967946358b74c86fa77143dbda6f630bf
2016-05-30 16:43:40 <ncopa> Author: Natanael Copa <ncopa@alpinelinux.org>
2016-05-30 16:43:40 <ncopa> Date:   Wed Oct 28 15:48:13 2015 +0000
2016-05-30 16:43:40 <ncopa>     main/busybox: fix comatibility with kmod
2016-05-30 16:43:51 <ncopa> but it did more than that
2016-05-30 16:43:55 <ncopa> i think who is broken
2016-05-30 16:44:05 <ncopa> i dont think it will work
2016-05-30 16:44:42 <ncopa> would need to build your own utmpx implementation for it i think
2016-05-30 16:45:47 <ncopa> http://wiki.musl-libc.org/wiki/FAQ#Q:_why_is_the_utmp.2Fwtmp_functionality_only_implemented_as_stubs_.3F
2016-05-30 16:46:30 <leo-unglaub> ah, i see
2016-05-30 16:46:50 <leo-unglaub> hmm, he has a point. its easy to see who does what whti that command
2016-05-30 16:48:43 <ncopa> thats the reason who was disabled
2016-05-30 16:48:55 <leo-unglaub> i see and i agree with it
2016-05-30 16:49:17 <leo-unglaub> better to keep it out then
2016-05-30 16:49:19 <ncopa> :)
2016-05-30 16:49:25 <leo-unglaub> security first
2016-05-30 16:50:09 <ncopa> i think that was my reaction too when i realized it does not work
2016-05-30 16:51:41 <leo-unglaub> yeah, understandable
2016-05-30 16:52:07 <leo-unglaub> one more question about the busyboxconfig. is there a special reason you did disable CONFIG_XZ in d0b4f922?
2016-05-30 16:52:23 <leo-unglaub> or is it just a by-product during the upgrade?
2016-05-30 16:52:38 <leo-unglaub> because i think it would make sence to enable it
2016-05-30 17:24:07 <leo-unglaub> ncopa: the wpa_supplicant fixed are horrible
2016-05-30 17:24:27 <leo-unglaub> code execution via dbus ... that tells you everything you need to know about the developers of it ...
2016-05-30 17:24:55 <jirutka> uh, is there any alternative to wpa_supplicant?
2016-05-30 17:26:43 <leo-unglaub> nope
2016-05-30 17:27:18 <jirutka> hm, so maybe it’s a time to fork it… as eudev etc.
2016-05-30 17:27:37 <leo-unglaub> rewrite would be better ;)
2016-05-30 17:27:42 <jirutka> or rewrite it from scratch… in Rust :P
2016-05-30 17:28:34 <jirutka> btw do you know http://os.phil-opp.com/ ?
2016-05-30 17:29:15 <jirutka> it’s really awesome blog series of writing kernel in Rust, starting with basics
2016-05-30 17:29:27 <leo-unglaub> yeah, i know it
2016-05-30 17:29:30 <leo-unglaub> great idea
2016-05-30 17:36:19 <leo-unglaub> jirutka: his blog post about page tables is very awesome
2016-05-30 17:36:25 <leo-unglaub> it helped me a lot to understand it
2016-05-30 17:49:55 <Adran> mmmm
2016-05-30 18:12:16 <jirutka> is there some simple way how to determine which dependencies should be added to -dev subpackage? /cc skarnet
2016-05-30 18:42:39 <skarnet> simple as in automatable? I'm afraid no. For simple cases, a few heuristics work, but nothing replaces actual human thought. Yet.
2016-05-30 18:42:53 <jirutka> what heuristics?
2016-05-30 18:44:38 <skarnet> if foo depends on bar, then foo-dev will probably depend on bar-dev, for instance.
2016-05-30 18:44:58 <skarnet> Sorry, I'm too tired to focus on that right now.
2016-05-30 18:45:12 <jirutka> thanks for a hint
2016-05-30 18:45:23 <jirutka> no problem, this does not hurry :)
2016-05-30 19:02:09 <jirutka> it seems that https://github.com/openSUSE/snapper is unusable without dbus :(
2016-05-30 19:49:56 <oneinsect> quick question friends... setenv bootargs /vmlinuz modules=loop,squashfs,sd-mod,usb-storage initrd=04000000 initrd=/initramfs-edison modloop=/modloop-edison console=ttyMFD2 g_multi.ethernet_config=rndis systemd.unit=multi-user.target hardware_id=00 g_multi.iSerialNumber=b33f335811680145fdc309c8f189130c g_multi.dev_addr=02:00:86:89:13:0c platform_mrfld_audio.audio_codec=dummy init=/init
2016-05-30 19:50:23 <oneinsect> is there a way to specify modloop file by mmc 0:6
2016-05-30 19:50:47 <oneinsect> i am trying to run alpine on intel edison
2016-05-30 19:51:06 <oneinsect> it boots but dont seem to find modloop file
2016-05-30 19:51:29 <oneinsect> hence prompt doesnt come...
2016-05-30 21:24:10 <Adran> it seems the alpine mirrors are all over the place. :\
2016-05-30 21:24:21 <Adran> dl-1 doesn't work, dl-2 works, dl-6 has super old software
2016-05-30 21:24:25 <Adran> :|
2016-05-30 21:25:09 <Adran> (24/28) Upgrading linux-firmware (20151021-r0 -> 20160516-r0)
2016-05-30 21:25:12 <Adran> thats on edge
2016-05-30 21:25:40 <jirutka> could you please report it on bugs.alpinelinux.org?
2016-05-30 21:26:45 <jirutka> ncopa once recommended me http://dl-cdn.alpinelinux.org/alpine/ as most up-to-date and fast mirror
2016-05-30 21:28:19 <jirutka> https://repository.fit.cvut.cz/mirrors/alpine/ should be also fine and it supports HTTPS and IPv6 ;) (it’s my mirror on university)
2016-05-30 21:35:39 <leo-unglaub> oh boy, we have some real experts on that mailinglist *g*
2016-05-30 21:37:27 <jirutka> yeah… exploit using “hidden” directory, lol!
2016-05-30 21:37:43 <Adran> wut?
2016-05-30 21:37:43 <leo-unglaub> yeah, the evil hidden folder *g*
2016-05-30 21:38:00 <leo-unglaub> maybe he was watching to much CSI *g*
2016-05-30 21:38:05 <jirutka> heh
2016-05-30 21:42:51 <jirutka> leo: but seriously, it’d be great to support also OpenBSD kernel :)
2016-05-30 21:46:30 <leo-unglaub> sure it would
2016-05-30 21:46:43 <leo-unglaub> but that has nothing todo with the ml question
2016-05-30 21:46:55 <jirutka> yeah
2016-05-30 21:47:02 <leo-unglaub> that would simply make life easier because i have scripts on every machine i run ... and i run a lot
2016-05-30 21:57:33 <jirutka> omg, so many ideas and so little time :(
2016-05-30 21:57:52 <leo-unglaub> hehe, welcome to my world :)
2016-05-30 21:58:12 <skarnet> you're both guests in mine!
2016-05-30 21:58:40 <jirutka> this is really depressing…
2016-05-30 21:59:17 <leo-unglaub> hehehe
2016-05-30 21:59:41 <skarnet> much better than coming to an old discipline where everything has already been done and you can never invent anything new.
2016-05-30 22:00:05 <skarnet> s/discipline/domain/
2016-05-30 22:02:07 <Adran> that .bin thing should be entertaining
2016-05-30 22:02:12 <Adran> hide a binary in there, nobody ever checks
2016-05-30 22:02:21 <jirutka> I’d like to focus only on few projects, but instead I get a new idea once a week and it depresses me that I can’t realize it
2016-05-30 22:03:03 <Adran> live in a box
2016-05-30 22:03:07 <Adran> thats what I d o
2016-05-30 22:03:18 <jirutka> I’m not a cat to live in a box…
2016-05-30 22:04:02 <skarnet> living in a box, or in a bubble, is the only way to get things done and not be dragged down by the world
2016-05-30 22:04:18 <jirutka> but how to do it?
2016-05-30 22:04:34 <coredumb> first buy a box
2016-05-30 22:04:38 <jirutka> XD
2016-05-30 22:04:49 <skarnet> or blow a giant soap bubble!
2016-05-30 22:05:50 <skarnet> seriously though, learn to disconnect. It's hard but it's vital.
2016-05-30 22:06:09 <skarnet> Even IRC is a distraction for me sometimes, and I'm in only 4 channels.
2016-05-30 22:06:24 <skarnet> Unfollow all the projects that aren't top-priority for you.
2016-05-30 22:06:30 <skarnet> Stop watching the news.
2016-05-30 22:07:25 <skarnet> Don't let negativity or unimportant things (to you) take away fron your energy or your time.
2016-05-30 22:09:01 <jirutka> for example, today I’ve been installing Alpine on a virtual machine (for the first time, I normally use it in LXC containers); I realize that we don’t have package for etckeeper… and that etckeeper needs perl that is quite huge, so I’d be best to write just some very simple tool to persist file permissions… then I started thinging about using Btrfs for rootfs and instead of tracking just /etc, track entire rootfs… so some
2016-05-30 22:09:56 <Adran> lol
2016-05-30 22:09:59 <skarnet> ^ that's called shaving the yak
2016-05-30 22:10:19 <skarnet> You can't shave a yak in a day, or even in a month.
2016-05-30 22:10:33 <skarnet> You have to decide on a small patch of skin you're going to shave, and focus on getting that done.
2016-05-30 22:10:37 <jirutka> than I found https://github.com/sysnux/btrfs-snapshots-diff/blob/master/btrfs-snapshots-diff.py, it doesn’t look so complicated, so I’ve started thinking about writing such utility
2016-05-30 22:10:56 <Adran> The only time i went down that path was making packaging for bro for alpine
2016-05-30 22:10:59 <Adran> each package needed packaging
2016-05-30 22:11:04 <Adran> 4 packages later, its ready to be submitted
2016-05-30 22:11:22 <Adran> as soon as I figure out how to make sure this thing works reliably on other platforms
2016-05-30 22:13:01 <jirutka> skarnet: and that’s just today! and that project seems to be quite small in comparison with other ideas of mine :(
2016-05-30 22:13:40 <coredumb> jirutka: yeah btrfs sucks
2016-05-30 22:13:47 <coredumb> see I saved you time
2016-05-30 22:13:48 <coredumb> \o/
2016-05-30 22:13:57 <jirutka> I disagree, btrfs is great
2016-05-30 22:14:25 <skarnet> I save time by not discussing the virtues and flaws of various filesystems
2016-05-30 22:14:25 <coredumb> until it gets corrupted for whatever reason like the height of the moon or such
2016-05-30 22:14:40 <coredumb> :)
2016-05-30 22:14:53 <coredumb> see skarnet we all save time ^^
2016-05-30 22:15:53 <jirutka> well, at least you can rebalance Btrfs, gradually add drives to RAID etc.… ZFS is extremly rigid
2016-05-30 22:16:16 <skarnet> you just lost 2 minutes explaining something about btrfs that nobody here is caring about
2016-05-30 22:16:22 <jirutka> okay okay
2016-05-30 22:16:24 <coredumb> I wouldn't say "extremely"
2016-05-30 22:16:46 <jirutka> I would say rigid as a concrete :)
2016-05-30 22:17:08 <barthalion> leo-unglaub: we also got an expert about introducing system-wide changes for something that can be done under 15 seconds
2016-05-30 22:17:21 <coredumb> but yeah not that flexible when it comes to vdevs counts
2016-05-30 22:17:38 <barthalion> leo-unglaub: also weren't you just complaining about people sending replies directly to you instead to the ml? )))
2016-05-30 22:30:37 <leo-unglaub> i was complaining about people not understanding how mailinglists work
2016-05-30 22:32:00 <Adran> did I break things
2016-05-31 04:05:21 <Adran> this is a new one http://i.imgur.com/UYrkCV2.png and this was going to be a training thing too. :D
2016-05-31 06:36:22 <tmh1999> Hi everyone, I have a question and I hope some guru here could help me out :). I am trying to build apk-tools from source and it seems alpine provides lua5.1 only. I am wondering if what is the role of lua in apt-tools ? I am looking at the src/Makefile file and trying to change LUA_VERSION to 5.1. What possible problem could I have?  2nd question : When I changed to lua5.1, src/url.c complains about missing fetch.h. I cannot find fetch.h on the
2016-05-31 06:36:23 <tmh1999> Internet. Any help is welcome, thanks!
2016-05-31 06:38:15 <tmh1999> Oops. 2nd question might be answered by installing libfetch.
2016-05-31 06:39:20 <tmh1999> Humm.. Seems compiled okay using lua5.1.
2016-05-31 08:12:10 <ncopa> need to do release 3.4.0 today
2016-05-31 08:12:22 <ncopa> what do we need to include in the release notes?
2016-05-31 08:12:27 <ncopa> rename of crond -> cron
2016-05-31 08:12:44 <ncopa> ping is no longer suid root, so you need to be in netdev group for ping
2016-05-31 08:13:00 <ncopa> 4.4 kernel
2016-05-31 08:15:10 <ncopa> new flavor of iso: alpine-virt
2016-05-31 08:23:49 <ScrumpyJack> morning
2016-05-31 08:24:25 <ScrumpyJack> what's in alpine-virt?
2016-05-31 08:24:40 <ncopa> same as alpine.iso but with linux-virtgrsec kernel
2016-05-31 08:24:51 <ncopa> kernel for vm
2016-05-31 08:24:58 <ncopa> so its much smaller
2016-05-31 08:25:51 <ScrumpyJack> found it http://git.alpinelinux.org/cgit/alpine-iso/commit/?id=aba4bd4a15280a91ed16e65680a399760ca66045
2016-05-31 09:31:16 <leo-unglaub> hey >(
2016-05-31 09:31:27 <leo-unglaub> :)
2016-05-31 09:43:00 <ncopa> hi
2016-05-31 09:43:45 <ncopa> i need feedback on http://lists.alpinelinux.org/alpine-devel/5311.html
2016-05-31 09:44:15 <ncopa> we allow ping for group id 999 -> 59999  by default?
2016-05-31 09:45:25 <ncopa> that is, all normal users
2016-05-31 09:45:45 <ncopa> or do we want users to be members of netdev to be able to ping?
2016-05-31 09:51:30 <leo-unglaub> hmmm
2016-05-31 09:51:54 <leo-unglaub> i dont see any danger in people using ping
2016-05-31 09:52:13 <leo-unglaub> the change to do some demage on the network is very slim and you have to do it on purpose
2016-05-31 09:53:15 <leo-unglaub> on an ubuntu i would say the target userbase does not need ping by default so it cann be disabled for them but i would asume alpine users to be more technical and therefor it should work by default
2016-05-31 09:55:05 <leo-unglaub> the downside ot the > 999 usergroup is that if you dont know it its harder to understand. a usergroup is the more default way to go on *nix devices
2016-05-31 09:55:43 <leo-unglaub> ncopa: i vote for the netdev group, but with the option that the installer adds the user automatically to that group
2016-05-31 10:00:14 <ncopa> it is not easy to implement add users automatically to that group
2016-05-31 10:00:18 <ncopa> i think...
2016-05-31 10:01:39 <leo-unglaub> i would do it in setup-alpine
2016-05-31 10:01:48 <leo-unglaub> and simply ask if a user should be able to ping
2016-05-31 10:01:56 <leo-unglaub> if yet we run the usermod command
2016-05-31 10:04:55 <leo-unglaub> ncopa: forgett about what i wrote. users should add themself to the group if they need ping. plain, clean and simple
2016-05-31 10:05:01 <leo-unglaub> netdev group allows to ping
2016-05-31 10:05:17 <leo-unglaub> if you or a deamon needs ping, use the group, its the same with audio, video, ...
2016-05-31 10:06:05 <ncopa> i am thinking call gid '999' for 'ping'
2016-05-31 10:06:21 <ncopa> if daemon needs execute /bin/ping they'd need to be in ping group
2016-05-31 10:06:29 <ncopa> users would be default be able to ping
2016-05-31 10:06:47 <leo-unglaub> i see
2016-05-31 10:07:01 <ncopa> i am faraid that if they dont by default can ping, users will use sudo ping or roll back to suidroot ping binary
2016-05-31 10:07:20 <leo-unglaub> hmm, thats true to
2016-05-31 10:07:40 <leo-unglaub> the problem with the >999 solution is that it is not very transparent
2016-05-31 10:07:59 <leo-unglaub> to debug why someone can ping and someone cannot would be harder
2016-05-31 10:08:07 <leo-unglaub> but i get your point about the sudo ping
2016-05-31 10:08:10 <leo-unglaub> and no one wants that
2016-05-31 10:08:24 <ncopa> same with require that you are in a special magic group
2016-05-31 10:08:49 <leo-unglaub> well, the group is more common, there are printer, audio, video, ... groups
2016-05-31 10:08:50 <ncopa> ok
2016-05-31 10:08:54 <leo-unglaub> the principle is more common
2016-05-31 10:09:06 <ncopa> yes and i will add a 'ping' group, with gid 999
2016-05-31 10:09:21 <leo-unglaub> okay, fine with me :)
2016-05-31 10:13:42 <skarnet> I don't understand the value of that change. Are there security issues with the ping code? Were we unhappy with busybox-setuid? The benefits have to balance the hassle of adding users to netdev group, in order for the change to be worth it.
2016-05-31 10:14:15 <ncopa> suid root is always risky
2016-05-31 10:14:21 <skarnet> no shit, Sherlock
2016-05-31 10:14:30 <ncopa> recent kernels has a special socket for ping
2016-05-31 10:14:53 <ncopa> but annoyingly you need set some magic sysctl values to be allowed to use it
2016-05-31 10:15:11 <skarnet> so my question stands
2016-05-31 10:15:21 <skarnet> you have looked at the busybox ping code
2016-05-31 10:15:24 <skarnet> you know what it does
2016-05-31 10:15:53 <ncopa> yes, and i am unhappy with running it as suid root
2016-05-31 10:16:06 <skarnet> do you think there's a potential security issue in the trusted code, with the probability of it worth the hassle of making the change?
2016-05-31 10:16:16 <skarnet> ok.
2016-05-31 10:16:20 <ncopa> yes
2016-05-31 10:17:23 <skarnet> A better solution to me would be "simplify the ping code so we can trust it". That would avoid the breakages that *will* happen when you release the modidied thing.
2016-05-31 10:18:10 <skarnet> Because you *know* people rely on ping in their automation, and until they add their automation user to netdev, they'll freak out. :)
2016-05-31 10:18:45 <leo-unglaub> adding the special sockets for ping to the kernel was a huge discussion if i remember correctly
2016-05-31 10:19:10 <leo-unglaub> i dont care how its done, what socket is used as long as the code is clean.
2016-05-31 10:19:38 <leo-unglaub> if you want to get it more secure you have to go the OpenBSD route and pledge the the binary
2016-05-31 10:19:54 <leo-unglaub> if ping does somethign else than it has a pledge for, kill it off
2016-05-31 10:20:10 <skarnet> don't get me wrong: ultimately, the less suid root, the better. But I find it a problem that we can't trust something as simple as ping.
2016-05-31 10:20:43 <leo-unglaub> skarnet: i agree with your premise, but in this case i think ncopa's solution is okay as well
2016-05-31 10:21:38 <skarnet> sure it's okay. Just brace yourselves for the influx of user complaints when their stuff breaks and they haven't read the release notes that say "put your ping users into netdev group".
2016-05-31 10:22:15 <leo-unglaub> i am not sure, but i think this issues is comming from the docker comminuty ....
2016-05-31 10:23:46 <skarnet> oh, don't worry, you will get a lot of docker-using devops types with milk still dribbling out of their nose that will come crying to the forums. :P
2016-05-31 10:24:46 <leo-unglaub> i think they freaked out because they just saw a suid binary and freaked out without thinking about it for a secound
2016-05-31 10:25:01 <skarnet> 10 zorkmids say something like npm or whatever automated package manager for $LANGUAGE_OF_THE_MONTH internally uses ping, and hilarity will ensue
2016-05-31 10:25:28 <ncopa> i think it was dalias who first requested it
2016-05-31 10:25:51 <ncopa> and so did a docker engineer
2016-05-31 10:26:23 <skarnet> Obviously dalias would love it. I'm more impressed by the request coming from a docker engineer :)
2016-05-31 10:26:48 <leo-unglaub> well, i can understand both sides
2016-05-31 10:27:02 <ncopa> me too
2016-05-31 10:27:14 <leo-unglaub> the question is, what is more secure ... a special kernel socket or a small simple ping suid binary
2016-05-31 10:27:27 <skarnet> again, theoretically it's a good thing. I'm more concerned about such a change happening between major releases
2016-05-31 10:27:48 <ncopa> i am relatively convinced a special kernel socket is more secure
2016-05-31 10:27:51 <skarnet> it it were 4.0 you wouldn't hear from me at all
2016-05-31 10:28:20 <ncopa> maybe we should call it 4.0 :)
2016-05-31 10:28:44 <leo-unglaub> 4.0?
2016-05-31 10:28:54 <leo-unglaub> nooooo, call it 410.1 like systemd!!!!!
2016-05-31 10:29:03 <ncopa> lol
2016-05-31 10:29:09 <skarnet> if there are a few changes like that that can break user workflow, then yes, a major version bump is warranted
2016-05-31 10:29:11 <ncopa> oh no.. he said the s* word...
2016-05-31 10:29:27 <skarnet> quick, erase the logs
2016-05-31 10:31:00 <leo-unglaub> hehe, have you read the latest systemd thing? where they are now trying to force other projects to depend on libsystemd for no reason? they break stuff now on purpose and then write tickets to depend on libsystemd so it will work again?
2016-05-31 10:31:24 <ncopa> i'm not surprised at all
2016-05-31 10:31:31 <ncopa> thats how they roll
2016-05-31 10:52:44 <leo-unglaub> hmm, something is still off
2016-05-31 10:52:45 <leo-unglaub> ping: permission denied (are you root?)
2016-05-31 10:53:47 <leo-unglaub> ah, it maybe needs a reboot
2016-05-31 10:54:36 <leo-unglaub> brb, reboot
2016-05-31 10:57:22 <leo-unglaub1> re
2016-05-31 10:57:30 <leo-unglaub1> yeah, after a reboot it works
2016-05-31 11:08:20 <ncopa> i suppose restarting sysctl would work too
2016-05-31 11:10:35 <ncopa> can anyone help me test upgrade?
2016-05-31 11:10:40 <ncopa> from v3.3 to v3.4
2016-05-31 11:11:17 <leo-unglaub1> ncopa: sure, i have to fire up virtualbox for it
2016-05-31 11:11:26 <leo-unglaub1> any special things i should watch out for?
2016-05-31 11:11:32 <ncopa> yes
2016-05-31 11:11:38 <ncopa> cron -> crond rename
2016-05-31 11:11:54 <ncopa> before upgrade,  check that cron runs
2016-05-31 11:12:02 <ncopa> init.d script is "cron"
2016-05-31 11:12:10 <ncopa> after upgrade it should be "crond"
2016-05-31 11:12:21 <ncopa> and you should be able to stop/restart crond
2016-05-31 11:12:53 <ncopa> and also test if you can ping with normal user
2016-05-31 11:12:59 <ncopa> create a normal user in v3.3
2016-05-31 11:13:02 <ncopa> test that ping works
2016-05-31 11:13:08 <ncopa> do upgrade, and test ping
2016-05-31 11:13:25 <leo-unglaub1> sure thing, let me reboot into windows where i have virtualbox running
2016-05-31 11:19:41 <leo-unglaub> 3.3.3 x64 okay?
2016-05-31 11:19:48 <ncopa> perfect
2016-05-31 11:20:36 <leo-unglaub> some special disk setup?
2016-05-31 11:21:02 <ncopa> no
2016-05-31 11:24:36 <leo-unglaub> i hate virtual machines, but for debugging the snapshots of virtuabox are awesome :)
2016-05-31 11:26:10 <ncopa> +1
2016-05-31 11:26:34 <jirutka> leo: what do you prefer, containers or bear metal? :)
2016-05-31 11:27:22 <leo-unglaub> containers suck even more
2016-05-31 11:27:27 <leo-unglaub> i prefer real hardware
2016-05-31 11:27:32 <leo-unglaub> but i guess i am oldschool
2016-05-31 11:27:40 <jirutka> so you like bears, okay :)
2016-05-31 11:29:42 <leo-unglaub> ncopa: that was my first alpine install where i did use the defaults and my first alpine install where everything worked *g*
2016-05-31 11:29:54 <skarnet> jirutka: someone saw http://fredrikdesigns.com/wp-content/uploads/2015/02/growlybear-blue.png :)
2016-05-31 11:30:07 <jirutka> skarnet: yeah :)
2016-05-31 11:30:42 <ScrumpyJack> ncopa: i think i might have bluetooth working for the RPi3. will i be too late for 3.4?
2016-05-31 11:32:49 <ncopa> ScrumpyJack: yes. but file it for 3.4.1
2016-05-31 11:33:09 <ScrumpyJack> ok
2016-05-31 11:34:34 <leo-unglaub> ncopa: /usr/sbin/crond -c /etc/ctontabs is running after the upgrade
2016-05-31 11:34:52 <leo-unglaub> ncopa: my normal user can ping
2016-05-31 11:35:45 <leo-unglaub> but the group 999 ping does not exists because the group and group.apk-new file are not merged
2016-05-31 11:53:10 <jirutka> group 999?
2016-05-31 12:02:29 <jirutka> vpsfree.cz, a community VPS hosting on OpenVZ, oficially announced support for Alpine Linux https://blog.vpsfree.cz/pridali-jsme-alpine-linux-3-3-a-slackware-14-1/ :)
2016-05-31 12:02:29 <jirutka> Translation of relevant parts:
2016-05-31 12:02:29 <jirutka> Alpine Linux is a minimalistic distribution focused on security. It’s based on musl libc, Busybox, uses OpenRC as an init system. Unfortunately, you can’t use PaX and grsecurity on our infrastructure.
2016-05-31 12:02:30 <jirutka> With the size 5.5 MB, Alpine is the smallest template that we ever had. For comparison, Debian 8 has 121 MB and CentOS 7 needs 244 MB. This template created Jakub Jirutka, thanks!
2016-05-31 12:02:31 <jirutka> :)
2016-05-31 12:09:39 <skarnet> in Czech Republic, templates create you?
2016-05-31 12:09:56 <jirutka1> XD
2016-05-31 12:10:04 <jirutka1> sorry, my English kinda sucks :(
2016-05-31 12:10:19 <skarnet> "was created by" :)
2016-05-31 12:10:38 <jirutka1> thanks
2016-05-31 12:24:28 <leo-unglaub> ncopa: something more you want tested? if not i shutdown the virtualbox and reboot into my work alpine instal
2016-05-31 12:49:28 <leo-unglaub> ncopa: you should do a blog post about alpine users now beeing able to use ping again and you could use the title: "Make Alpine ping again"!
2016-05-31 12:50:45 <jirutka> XD
2016-05-31 12:53:48 <ncopa> jirutka: awesome!
2016-05-31 12:54:56 <ncopa> leo-unglaub: i think that was it basically
2016-05-31 12:55:20 <skarnet> leo-unglaub: I'd rather avoid references to that quite sad political climate, even in a joking fashion
2016-05-31 12:55:37 <jirutka> ^ agree
2016-05-31 13:10:01 <leo-unglaub> hmm, i think we have a boot dependency init problem
2016-05-31 13:11:30 <leo-unglaub> \/etc/init.d/dmcrypt needs to be run after mdadm-raid
2016-05-31 13:12:03 <leo-unglaub> because if you want to decrypt a device mapper that comes from an mdadm it will fail otherwize
2016-05-31 13:12:51 <barthalion> skarnet: yeah, actually "was created by" sounds fine if you translate it back word by word to Czech or Polish :P
2016-05-31 13:13:13 <barthalion> ah, but you are complaining about template created
2016-05-31 13:14:01 <ncopa> leo-unglaub: that is a config issue. you can in theory also use a dmcrypted disk as pv for lvm
2016-05-31 13:14:45 <leo-unglaub> hmmm, thats true
2016-05-31 13:14:59 <ncopa> leo-unglaub: i think you need to do: echo 'rc_need="mdadm-raid"' >> /etc/conf.d/dmcrypt
2016-05-31 13:15:50 <barthalion> ncopa: wrt python3, I'd like to push it forward when 3.4.0 is out
2016-05-31 13:16:03 <barthalion> question is if it wouldn't be better to handle python libraries the ruby way
2016-05-31 13:16:15 <barthalion> just leave some widely used or crucial for other packages
2016-05-31 13:16:59 <ncopa> agree, i want fix python3
2016-05-31 13:17:17 <leo-unglaub> i tested the upgrade again and i noticed that the installer still writes "rw,relatime,data=ordered" into fstab
2016-05-31 13:17:19 <ncopa> using pip would make things simple
2016-05-31 13:17:27 <leo-unglaub> rw,relatime,data=ordered is not the same as "defaults"
2016-05-31 13:18:05 <ncopa> leo-unglaub: i think the setup script will mount use whatever it got mounted with
2016-05-31 13:18:22 <ncopa> so if you manually mount the partition with some extra options, it will get into fstab
2016-05-31 13:18:36 <leo-unglaub> ah, perfect
2016-05-31 13:18:40 <ncopa> i think it gets the values from /proc/mounts
2016-05-31 13:18:44 <ncopa> iirc
2016-05-31 13:18:46 <barthalion> ncopa: so I'll send a message tomorrow to ml
2016-05-31 13:18:53 <ncopa> good
2016-05-31 13:18:58 <ncopa> i'll do 3.4.0 now
2016-05-31 13:21:08 <ncopa> rnalrd: do you think that your setup-interfaces patch for wifi is enough to close #1008?
2016-05-31 13:22:39 <rnalrd> ncopa, i think yes
2016-05-31 13:23:01 <rnalrd> it cycles between PSK and open networks, and lets you pick one
2016-05-31 13:23:12 <ncopa> thanks!
2016-05-31 13:30:45 <barthalion> ncopa: one remaining question… I'd actually like to drop non-crucial python2 libraries either, not sure how would that affect release notes for 3.5.0
2016-05-31 13:31:09 <barthalion> "sup no packages for you anymore, enjoy apk del && pip install cycle"
2016-05-31 13:31:25 <barthalion> doesn't sound too good
2016-05-31 13:31:53 <ncopa> non-crucial python libs
2016-05-31 13:32:00 <ncopa> how do we find those?
2016-05-31 13:32:17 <barthalion> anything that isn't widely used as dependency for all the rest
2016-05-31 13:32:29 <ncopa> hm
2016-05-31 13:32:31 <ncopa> might work
2016-05-31 13:32:32 <barthalion> I think we have lots of weird libraries in testing I never heard about
2016-05-31 13:32:39 <ncopa> yes
2016-05-31 13:32:47 <barthalion> things like python-requests would stay packages for both py2 and py3
2016-05-31 13:32:48 <ncopa> or we could convert those to python3
2016-05-31 13:32:58 <barthalion> s/packages/packaged/
2016-05-31 13:33:06 <ncopa> lets talk about it later this week
2016-05-31 13:33:10 <ncopa> i want 3.4 out now
2016-05-31 13:33:18 <barthalion> sure thing
2016-05-31 13:33:28 <ncopa> #3524
2016-05-31 13:33:39 <ncopa> i am cleaning up bugs.a.o
2016-05-31 13:33:53 <ncopa> oh
2016-05-31 13:33:55 <ncopa> one thing more
2016-05-31 13:33:58 <ncopa> php7
2016-05-31 13:34:01 <ncopa> never made it to main?
2016-05-31 13:34:31 <barthalion> guess not
2016-05-31 13:34:47 <barthalion> let's handle this in 3.5 too
2016-05-31 13:35:17 <ncopa> oh no... i didnt do #3951
2016-05-31 13:35:24 <barthalion> heh
2016-05-31 13:35:31 <ncopa> i think its 80% done
2016-05-31 13:35:31 <barthalion> critical blocker indeed
2016-05-31 13:35:44 <ncopa> i have it in my git stash history somwhere deep
2016-05-31 13:37:49 <barthalion> I don't even try to find anything in mine
2016-05-31 13:38:02 <leo-unglaub> hahaha, git stash ... the worst command ever invented ... what i have in that stash ... i dont even want to know *g*
2016-05-31 13:44:39 <clandmeter> should rename it to recycle.bin
2016-05-31 13:56:08 <ncopa> can anyone help me clean up release notes? http://tpaste.us/Apzm
2016-05-31 13:56:21 <ncopa> i added a stupid note on ping issue and cron rename
2016-05-31 13:56:28 <ncopa> it needs to be written properly
2016-05-31 13:56:36 <ncopa> ok
2016-05-31 13:56:43 <ncopa> php was renamed to php5
2016-05-31 13:57:16 <ncopa> new major version of samba
2016-05-31 13:57:25 <ncopa> samba dc support was removed
2016-05-31 13:57:44 <ncopa> support for rpi2?
2016-05-31 13:58:08 <skazz> rpi3* ?
2016-05-31 13:58:16 <ncopa> rpi3
2016-05-31 13:58:31 <leo-unglaub> ncopa: i am on it
2016-05-31 13:58:34 <leo-unglaub> give me a sec
2016-05-31 13:58:35 <skazz> Is it arm64 or still just the pi2 armhf?
2016-05-31 13:58:49 <ncopa> still 32bit
2016-05-31 13:58:57 <ncopa> might want mention that
2016-05-31 13:58:58 <skazz> fair enough
2016-05-31 13:59:05 <ncopa> ScrumpyJack: i pushed bluez fix
2016-05-31 14:00:20 <leo-unglaub> ncopa: https://paste.debian.net/plainh/f7452415
2016-05-31 14:00:42 <ncopa> LOL
2016-05-31 14:02:37 <leo-unglaub> for every developer that uses geany, this looks like a nice plugin: http://plugins.geany.org/git-changebar.html
2016-05-31 14:03:05 <jirutka> leo: Rust is finally available?
2016-05-31 14:03:12 <skazz> ^
2016-05-31 14:03:26 <skazz> It still says it is Early Access on steam
2016-05-31 14:03:50 <ncopa> i think he is just messing with us
2016-05-31 14:04:15 <leo-unglaub> i made a joke, dont get overexited *g*
2016-05-31 14:04:27 <ncopa> we should also move some of the issues from "new features" to "important changes" or similar
2016-05-31 14:05:07 <jirutka> ncopa: what kind of changes you’d like to mention in release notes?
2016-05-31 14:05:24 <ncopa> things that people need to know before upgrading
2016-05-31 14:06:06 <ncopa> i have used to include mentions of major releases of things like postgresql because it requires dump/restore
2016-05-31 14:06:11 <jirutka> ruby 2.2 → 2.3 ?
2016-05-31 14:06:16 <ncopa> yes
2016-05-31 14:06:28 <ncopa> that should be mentioned
2016-05-31 14:06:36 <leo-unglaub> hmmm, i think most of the software in this release is going to upgrade without users even notice
2016-05-31 14:06:47 <ncopa> thinkgs like ping change
2016-05-31 14:06:54 <ncopa> rename of crond service
2016-05-31 14:06:57 <ncopa> php5 rename
2016-05-31 14:07:01 <ncopa> ruby 2.3
2016-05-31 14:07:13 <ncopa> and some usefule information
2016-05-31 14:07:17 <ncopa> like what kernel version it is
2016-05-31 14:07:44 <ncopa> the desktop app list was only added to get a bit higher score on distrowatch :)
2016-05-31 14:08:08 <jirutka> maybe it’s worth to mention that ruby 2.2 is available in ruby2.2 package in testing
2016-05-31 14:08:17 <ncopa> we should probably move it to community
2016-05-31 14:08:21 <ncopa> once we have tested it
2016-05-31 14:08:29 <ncopa> testing is not available for stable branches
2016-05-31 14:08:49 <ncopa> i am not sure we want support ruby2.1 btw
2016-05-31 14:09:01 <barthalion> I think it actually is, but it's a snapshot from release date?
2016-05-31 14:09:03 <skazz> ncopa: Why not have a snapshot for testing as of release? 3.3.3 release firefox 42 was working, but now it is not and I cannot test it =/
2016-05-31 14:09:09 <barthalion> I guess it should be removed completely
2016-05-31 14:09:13 <jirutka> should I send PR for moving ruby2.2 to community?
2016-05-31 14:09:34 <barthalion> we either support it or not, no reason to give shotgun to user
2016-05-31 14:10:03 <ncopa> barthalion: not following...
2016-05-31 14:10:09 <ncopa> are you talking about testing repo?
2016-05-31 14:10:12 <barthalion> yes
2016-05-31 14:10:25 <ncopa> we dont build it on 3.4 builder
2016-05-31 14:10:35 <ncopa> so there are no testing packages under v3.4
2016-05-31 14:10:49 <barthalion> okay, outdated info then
2016-05-31 14:11:00 <barthalion> I remember it used to be there long time ago
2016-05-31 14:11:02 <ncopa> might be we shipped a snapshot once
2016-05-31 14:11:10 <skazz> Would be nice to have testing packages as of a release which use the stable required ones when they are avaliable... So we don't 'have' to use edge for everything just to use one package which worked at that point in time
2016-05-31 14:11:12 <barthalion> moving on then
2016-05-31 14:11:32 <skazz> I'll sush now tho =x
2016-05-31 14:11:39 <barthalion> skazz: we had a repository called "backports" for 2.7, but I think ncopa wasn't very fond of it
2016-05-31 14:11:53 <ncopa> because it added maintenance work
2016-05-31 14:12:11 <ncopa> we have hard time maintaining the branches we already have
2016-05-31 14:12:17 <barthalion> I never sat on stable release so I guess I'll sush as well
2016-05-31 14:12:29 <barthalion> 
ACTION sushes

2016-05-31 14:13:02 <ncopa> :)
2016-05-31 14:13:10 <skazz> Would it not just be a copy of edge/testing to v3.4/testing as of the release? Does not need maintaining, just make it clear on a wiki that the release/testing is static and edge/testing is what you use the rest of the time if you have issues with them
2016-05-31 14:13:23 <barthalion> it's a no go
2016-05-31 14:13:28 <barthalion> people enable it, then cry it's broken
2016-05-31 14:13:35 <skazz> ah, ok =P
2016-05-31 14:14:05 <skazz> I forget about the people factor sometimes
2016-05-31 14:14:32 <barthalion> it's not very difficult to backport a package manually, using docker or lxc
2016-05-31 14:14:52 <skazz> You cannot backport testing packges however, your stuck with edge or compiling it all yourself
2016-05-31 14:15:05 <barthalion> this is what I mean by backport
2016-05-31 14:15:09 <barthalion> compiling a package
2016-05-31 14:15:39 <ncopa> might be an idea to try make that simpler for users
2016-05-31 14:15:40 <skazz> Oh, I just don't like compiling and having tools on router setups or raspberry pis
2016-05-31 14:15:42 <ncopa> like aur
2016-05-31 14:15:54 <ncopa> anyway
2016-05-31 14:16:03 <ncopa> i'm tagging 3.4.0 now
2016-05-31 14:16:10 <barthalion> well, aports are like aur already
2016-05-31 14:16:15 <barthalion> especially community repo
2016-05-31 14:16:17 <ncopa> please don't push anything now
2016-05-31 14:16:27 <barthalion> aur isn't binary repository, just place for PKGBUILDs
2016-05-31 14:16:28 <skazz> 
ACTION pushes barthalion

2016-05-31 14:16:31 <ncopa> please dont push or barthalion will sush
2016-05-31 14:16:42 <barthalion> :P
2016-05-31 14:17:08 <scadu> #alpine-hiphop
2016-05-31 14:17:23 <ncopa> when i'm done i want see the release notes ready ;)
2016-05-31 14:17:45 <jirutka> ncopa: ruby2.2 to the community repo https://github.com/alpinelinux/aports/pull/104.patch
2016-05-31 14:18:55 <jirutka> skazz: you can simply create your own repository and use e.g. Travis as a build server ;) https://github.com/jirutka/aports#how-to-setup-your-own-repository
2016-05-31 14:19:14 <jirutka> skazz: however I don’t know if Travis supports ARM, but there are also other public CI servers
2016-05-31 14:23:43 <ScrumpyJack> hmm, why does my rpi zero only show half the RAM
2016-05-31 14:23:54 <ScrumpyJack> grep MemTotal /proc/meminfo 251676 kB
2016-05-31 14:25:38 <ScrumpyJack> gpu_mem=16
2016-05-31 14:27:10 <stwa> ScrumpyJack: thought that it's a known issue that gpu_mem is not taken into account
2016-05-31 14:27:27 <barthalion> jirutka: that's fancy
2016-05-31 14:27:43 <ScrumpyJack> stwa: yikes!
2016-05-31 14:28:21 <jirutka> barthalion: fancy?
2016-05-31 14:28:55 <ScrumpyJack> update-kernel requires more tmpfs than that.
2016-05-31 14:28:59 <skazz> jirutka: Thanks for the gude, setting up my own CI for my projects is something I need to do in the next month so Ill use that for aports too
2016-05-31 14:29:49 <stwa> ScrumpyJack: forgot where I read about it - but trying to remember...
2016-05-31 14:31:35 <stwa> ScrumpyJack: http://forum.alpinelinux.org/forum/kernel-and-hardware/raspberry-pi-and-ram-split ?
2016-05-31 14:31:44 <ScrumpyJack> stwa: i just changed all the pi sections in config.txt to gpu_mem=16 and it's now bricked :(
2016-05-31 14:33:29 <ncopa> sounds bad
2016-05-31 14:34:34 <ScrumpyJack> only annoying, in that the RPi0 in question is at home, and i'm accessing it from work (trying to figure out what's missing for USB gadget)
2016-05-31 14:35:30 <leo-unglaub> now thats what i call minimal: http://core.suckless.org/sinit
2016-05-31 14:35:30 <stwa> i had the same problem about 3 weeks ago, sorry, should have warned you :/
2016-05-31 14:38:31 <ncopa> is the release notes ready yet?
2016-05-31 14:39:27 <jirutka> ncopa: what about moving ruby2.2 to community?
2016-05-31 14:39:38 <ncopa> to late sorry
2016-05-31 14:39:50 <leo-unglaub> ncopa: i am waiting on your desicion on jirutka question
2016-05-31 14:40:04 <ncopa> it would keep the arm builder busy for an hour or so
2016-05-31 14:40:07 <ncopa> cannot wait for that
2016-05-31 14:40:11 <jirutka> okay :(
2016-05-31 14:40:18 <jirutka> so 3.4.1?
2016-05-31 14:40:28 <ncopa> yes i think so
2016-05-31 14:40:38 <ncopa> leo-unglaub: what q?
2016-05-31 14:41:02 <leo-unglaub> the ruby one
2016-05-31 14:41:22 <leo-unglaub> but you answered it while i was typing *g*
2016-05-31 14:41:28 <ScrumpyJack> same behaviour on RPi3, something isn't right
2016-05-31 14:41:39 <jirutka> when will be ISO image ready?
2016-05-31 14:41:49 <ncopa> it already is
2016-05-31 14:41:52 <ncopa> i just signed them
2016-05-31 14:42:11 <ncopa> http://rsync.alpinelinux.org/alpine/v3.4/releases/x86_64/
2016-05-31 14:42:22 <fcolista> weee
2016-05-31 14:42:25 <jirutka> aha, great :)
2016-05-31 14:42:25 <fcolista> \o/
2016-05-31 14:42:28 <fcolista> welcome baby
2016-05-31 14:42:40 <ncopa> :D
2016-05-31 14:42:49 <fcolista> 
ACTION is always exicted when a new alpine version is shipped

2016-05-31 14:43:09 <ncopa> 
ACTION is always stressed when a new alpine version is shipped

2016-05-31 14:43:18 <clandmeter> 
ACTION hands fcolista some tissues

2016-05-31 14:44:23 <fcolista> :)
2016-05-31 14:46:40 <skarnet> 
ACTION is always wary when a new .* version is shipped

2016-05-31 14:47:28 <leo-unglaub> ncopa: https://paste.debian.net/plainh/fff31b05
2016-05-31 14:47:35 <leo-unglaub> is that okay for you?
2016-05-31 14:49:51 <ncopa> i dont know if we need to include why changes was made
2016-05-31 14:51:46 <stwa> i thought cron got renamed to crond?
2016-05-31 14:51:52 <ncopa> yes
2016-05-31 14:51:53 <ncopa> that too
2016-05-31 14:52:13 <leo-unglaub> what?
2016-05-31 14:52:19 <leo-unglaub> no, the other way around
2016-05-31 14:52:25 <leo-unglaub> or am i stupid nowß
2016-05-31 14:52:26 <leo-unglaub> ?
2016-05-31 14:53:27 <leo-unglaub> ah, damn ... you guys are right *g*
2016-05-31 14:54:05 <clandmeter> is the bin renamed or the init script?
2016-05-31 14:54:57 <ncopa> init script
2016-05-31 14:55:03 <ncopa> - The service `cron` got renamed to `crond`.
2016-05-31 14:55:07 <clandmeter> what about ruby pkgs?
2016-05-31 14:55:47 <clandmeter> or were they already removed previously?
2016-05-31 14:55:55 <ncopa> no
2016-05-31 14:55:58 <ncopa> we need to mention those
2016-05-31 14:56:31 <jirutka> yeah, the ruby packages genocide, that should be mentioned :P
2016-05-31 14:56:46 <fcolista> jirutka, rotfl :D
2016-05-31 14:56:59 <ncopa> we should write a wiki page on the busybox ping issue
2016-05-31 14:57:47 <clandmeter> for ping i need to be a member of a certain group?
2016-05-31 14:58:10 <ncopa> http://tpaste.us/AWkB
2016-05-31 14:58:40 <ncopa> based on leo-unglaub's
2016-05-31 14:58:49 <jirutka> why it starts with 999 instead 1000 ?
2016-05-31 14:59:08 <jirutka> typo detected: grou wiht
2016-05-31 14:59:10 <ncopa> we introduce a group named 'ping' with gid 999
2016-05-31 14:59:19 <jirutka> uh, why 999? o.O
2016-05-31 14:59:49 <ncopa> because linux requires a range
2016-05-31 15:00:17 <ncopa> so the special group for deamons needs to be directly before user ids (1000) or directly after
2016-05-31 15:00:57 <clandmeter> a group to use ping, it sounds unreal.
2016-05-31 15:01:11 <ncopa> its a group range
2016-05-31 15:01:15 <ncopa> welcome to linux
2016-05-31 15:01:18 <clandmeter> :)
2016-05-31 15:01:22 <jirutka> what’s different between alpine and alpine-virt ISO?
2016-05-31 15:01:29 <ncopa> kernel
2016-05-31 15:01:43 <ncopa> kernel is smaller in alpine-virt as it only has the drivers for vm
2016-05-31 15:01:52 <jirutka> like virtio drivers?
2016-05-31 15:01:55 <ncopa> yes
2016-05-31 15:02:09 <ncopa> and some of the hw drivers that are known to be emulated
2016-05-31 15:02:39 <jirutka> okay, maybe I’ll try to use distribution kernel instead of my custom made for once :)
2016-05-31 16:12:35 <ncopa> should we include the commit statistics? with email addresses
2016-05-31 16:12:45 <ncopa> i think maybe we should strip the email addr
2016-05-31 16:15:08 <ncopa> the dns issue?
2016-05-31 16:15:14 <ncopa> its new in 3.4 right?
2016-05-31 16:15:24 <ncopa> the search support in resolv.conf
2016-05-31 16:27:10 <barthalion> ncopa: the fact that it's fixed? yes, it's new
2016-05-31 16:34:31 <ncopa> http://wwwtest.alpinelinux.org/posts/Alpine-3.4.0-released.html
2016-05-31 16:34:36 <ncopa> anything else should be mentioned?
2016-05-31 16:36:17 <ncopa> should we mention rpi3 support?
2016-05-31 16:36:27 <ncopa> ScrumpyJack: does rpi3 work?
2016-05-31 16:36:43 <jirutka> yay, I made it into TOP10! \o/
2016-05-31 16:36:53 <ncopa> good job :)
2016-05-31 16:38:02 <jirutka> I’d like to be in TOP5 for v3.5 :P
2016-05-31 16:39:43 <letoram> did uefi make it in there?
2016-05-31 16:39:55 <ncopa> unfortunally not :-(
2016-05-31 16:40:15 <skazz> ncopa: pi3 works with 3.3.3
2016-05-31 16:40:40 <skazz> I can try an 3.4 image if you like
2016-05-31 16:41:08 <ncopa> would be nice
2016-05-31 16:41:10 <ncopa> 3.4.0 is out
2016-05-31 16:42:05 <jirutka> I’ve just tried alpine-virt on KVM VM
2016-05-31 16:42:14 <jirutka> and it works
2016-05-31 16:47:40 <skazz> Just upgrading the pi, not done the upgrade before as I just do reinstalls so far, replaced repos with v3.4, apk update, apk upgrade, reboot... all seems fine, running the firefox html5 application as before
2016-05-31 16:48:04 <skazz> shows as Alpine 3.4, Kernel 4.1.20-0-rpi2 on armv7
2016-05-31 16:48:14 <skazz> upgrading kernel seperate or is that right?
2016-05-31 16:48:54 <skazz> Ah, ill do the kernel upg now
2016-05-31 16:49:10 <ncopa> depends on if you run from ram or from disk
2016-05-31 16:51:05 <skazz> Run from ram
2016-05-31 16:51:45 <ncopa> then you need to do it separately
2016-05-31 16:51:48 <skazz> Did apk add linux-rpi2, failed to create /lib/firmware ... that folder does not exist and probably would not have enough space in tmpfs
2016-05-31 16:51:56 <ncopa> correct
2016-05-31 16:52:09 <ncopa> you need to replace the /boot dir with the new tarball
2016-05-31 16:52:18 <skazz> My trick was to and i will do it with firmware, remove the /lib/modules symlink and symlink that and /lib/firmware to the 2nd partition that is ext4
2016-05-31 16:52:37 <skazz> It worked with the last upgrade
2016-05-31 16:52:51 <ncopa> http://dl-cdn.alpinelinux.org/alpine/v3.4/releases/armhf/
2016-05-31 16:52:59 <ncopa> i think you need replace content with new tarball
2016-05-31 16:53:02 <Adran> ncopa: why not update the alpine setup stuff to just use dl-cdn?
2016-05-31 16:53:24 <ncopa> not had the time
2016-05-31 16:53:30 <fabled> skazz, just unpack the new tarball to mmcblk0p1
2016-05-31 16:53:52 <ncopa> skazz: maybe make backup of boot/ and apks/ first
2016-05-31 16:54:22 <barthalion> I can imagine fancy coreos-like upgrade mechanism with another partition
2016-05-31 16:54:34 <ncopa> would be nice yes
2016-05-31 16:54:45 <ncopa> ijust push this now? http://wwwtest.alpinelinux.org/posts/Alpine-3.4.0-released.html
2016-05-31 16:54:45 <skazz> You can actually just create a loopback file with dd, format it ext4 and mount it
2016-05-31 16:54:47 <barthalion> not sure if it's possible with rpi?
2016-05-31 16:55:00 <barthalion> does it always boot from first partition?
2016-05-31 16:55:35 <skazz> Yes, but you can just create seperate folders/files and add it too the boot ini file
2016-05-31 16:55:46 <skazz> But it does not do the fallback I think
2016-05-31 16:55:48 <fabled> yes, the rpi bootloader uses partition #1's config.txt
2016-05-31 16:56:58 <ncopa> postgresql 9.5
2016-05-31 17:21:03 <jirutka> ncopa: link to ISO image is still pointing to v3.3 https://alpinelinux.org
2016-05-31 17:30:04 <ScrumpyJack> ncopa: is it too late for release note changes?
2016-05-31 17:30:13 <ncopa> kind of
2016-05-31 17:30:15 <ScrumpyJack> ok
2016-05-31 17:30:20 <ncopa> but i can fix the web page
2016-05-31 17:30:28 <ncopa> email was sent out
2016-05-31 17:31:03 <ScrumpyJack> just though we could put them through native British-english speaker
2016-05-31 17:31:28 <ncopa> ScrumpyJack: are you native British?
2016-05-31 17:31:42 <ncopa> i had someone native american to look at it
2016-05-31 17:31:53 <ncopa> and someone native British too
2016-05-31 17:32:18 <jirutka> ncopa: you’re not native american?
2016-05-31 17:32:33 <ncopa> no. native swedish
2016-05-31 17:32:38 <ncopa> or more like half native
2016-05-31 17:32:52 <jirutka> where do you currently live?
2016-05-31 17:32:57 <ncopa> norway
2016-05-31 17:33:00 <Adran> "native american"
2016-05-31 17:33:15 <ncopa> que bueno!
2016-05-31 17:33:21 <ncopa> :)
2016-05-31 17:38:15 <ncopa> jirutka: a.o download url is still 3.3 because i havent updated latest-stable symlink
2016-05-31 17:38:19 <ncopa> will fix it now
2016-05-31 17:38:32 <jirutka> okay
2016-05-31 17:39:06 <ScrumpyJack> ok, i'll leave it then
2016-05-31 17:39:17 <ncopa> ScrumpyJack: I'm open to suggestions!
2016-05-31 17:39:29 <ncopa> if you have an opinion let me know
2016-05-31 17:47:33 <ScrumpyJack> ncopa: what's the name of the ping permitted group?
2016-05-31 17:47:37 <ScrumpyJack> netsomething?
2016-05-31 17:47:40 <ScrumpyJack> netdev?
2016-05-31 17:48:12 <ncopa> 'ping'
2016-05-31 17:48:18 <ncopa> or gid 999
2016-05-31 17:48:31 <ncopa> all normal login users should also be able to ping
2016-05-31 17:48:42 <ScrumpyJack> what's with the 999 to 5999 thing?
2016-05-31 17:48:45 <ncopa> you might need re-run sysctl.conf
2016-05-31 17:48:54 <ncopa> that is the group range
2016-05-31 17:49:14 <ncopa> you need to be in one of the groups in the gid rand 999-59999
2016-05-31 17:49:51 <ScrumpyJack> so there's a ping group that gets created and it takes a gid between 999 and 59999
2016-05-31 17:50:03 <ncopa> no
2016-05-31 17:50:08 <ncopa> you set a group range
2016-05-31 17:50:36 <ScrumpyJack> range of users?
2016-05-31 17:50:40 <ncopa> net.ipv4.ping_group_range=999 59999
2016-05-31 17:51:06 <ncopa> there is a /proc/sys/net/ipv4/ping_group_range
2016-05-31 17:51:28 <ScrumpyJack> ok
2016-05-31 17:51:37 <ncopa> you can set it to 0 - 40000000000 or something
2016-05-31 17:51:52 <ncopa> and everyone gets the privilege to create ICMP ping sockets
2016-05-31 17:53:50 <ncopa> https://sturmflut.github.io/linux/ubuntu/2015/01/17/unprivileged-icmp-sockets-on-linux/
2016-05-31 17:54:45 <jirutka> where can I find some documentation for /etc/network/interfaces ?
2016-05-31 17:57:54 <ScrumpyJack> ncopa: http://tpaste.us/AeXb
2016-05-31 17:58:11 <ncopa> diff?
2016-05-31 17:58:20 <ScrumpyJack> failed :(
2016-05-31 17:58:39 <ncopa> ScrumpyJack:  http://git.alpinelinux.org/cgit/alpine-mksite/
2016-05-31 17:59:04 <ncopa> http://git.alpinelinux.org/cgit/alpine-mksite/tree/posts/Alpine-3.4.0-released.md
2016-05-31 18:00:19 <ScrumpyJack> patch gave me tpaste.us/2k89 which seems wrong
2016-05-31 18:00:32 <ncopa> yeah
2016-05-31 18:00:38 <ncopa> http://tpaste.us/AO9V
2016-05-31 18:00:59 <ncopa> hav been removed.
2016-05-31 18:00:59 <ScrumpyJack> what went wrong with my patch?
2016-05-31 18:02:06 <ncopa> ScrumpyJack: I like your suggestions
2016-05-31 18:02:52 <ncopa> like this? http://tpaste.us/3Ja1
2016-05-31 18:03:06 <ScrumpyJack> tpaste.us/3wMO
2016-05-31 18:03:45 <ncopa> looks llike you just show one of mine commits
2016-05-31 18:04:28 <ScrumpyJack> get me try again, don't apply yet :)
2016-05-31 18:04:34 <ncopa> ok :)
2016-05-31 18:06:46 <ncopa> i have only one more pullet point in my release check list
2016-05-31 18:06:51 <ncopa> celebrate!
2016-05-31 18:08:13 <jirutka> \o/
2016-05-31 18:08:26 <ncopa> ScrumpyJack: it would be nice if you had a look at the release image descriptions too: http://git.alpinelinux.org/cgit/alpine-mksite/tree/_scripts/generate_releases.lua
2016-05-31 18:10:32 <ncopa> have a nice evening!
2016-05-31 18:10:33 <ncopa> \o/
2016-05-31 18:13:11 <ScrumpyJack> kk.
2016-05-31 18:13:52 <ScrumpyJack> i'll send patches in the morning
2016-05-31 18:20:21 <barthalion> what's the difference between nginx and nginx-lua again?
2016-05-31 18:20:27 <barthalion> except the latter is terribly outdated
2016-05-31 18:20:51 <jirutka> IMO nginx-lua is compiled with lua module?
2016-05-31 18:21:10 <barthalion> well, no, nginx also has lua support applied
2016-05-31 18:22:22 <barthalion> what's the current state of the art method of package replacement?
2016-05-31 18:22:50 <barthalion> ncopa: ↑ provides? replaces?
2016-05-31 18:32:44 <jirutka> barthalion: ncopa is probably celebrating ;)
2016-05-31 18:33:17 <barthalion> I figure
2016-05-31 19:43:48 <coredumb> talking about nginx there's already a security fix out :D
2016-05-31 19:45:08 <Adran> for what  vuln?
2016-05-31 19:46:45 <coredumb> apparently segfault
2016-05-31 19:47:15 <coredumb>     *) Security: a segmentation fault might occur in a worker process while
2016-05-31 19:47:17 <coredumb>        writing a specially crafted request body to a temporary file
2016-05-31 19:47:19 <coredumb>        (CVE-2016-4450); the bug had appeared in 1.3.9.
2016-05-31 19:51:45 <coredumb> hey so now that 3.4 is cut can we send py3- packages ?
2016-05-31 19:51:47 <coredumb> :)
2016-05-31 21:42:06 <barthalion> coredumb: I pushed a fix for this CVE to edge and 3.4
2016-05-31 21:42:37 <barthalion> wrt py3… hold your horses, I'll send a monologue tomorrow to alpine-devel