2015-04-01 01:12:31 <tracker1> _mjones, I try to do the same, but it's hard getting deeper into system level stuff... I try to have some cursory familiarity with development options throughout a distributed application (from front end to the data store), but it's impossible to know it all.
2015-04-01 01:14:19 <_mjones> 99% won't get a lot of payback from diving into the very low-level. Usually only if you're building something at that level, or debugging a performance problem that manifests much higher in the stack.
2015-04-01 01:15:00 <_mjones> However, knowing the middle-level is a different story I'd say.
2015-04-01 01:15:32 <tracker1> _mjones, that's too true... and a benefit of having phenominal cosmic horsepower to throw at computing problems these days... I don't think I've had to sink lower than node or C# in 8 years.
2015-04-01 01:17:35 <_mjones> Yes, but. Horsepower doesn't solve problems of latency or jitter, doesn't change the properties of data gravity.
2015-04-01 01:17:56 <_mjones> And horsepower can still be expensive, as anyone who pays an EC2 bill well knows.
2015-04-01 01:18:25 <Diftraku> hoers
2015-04-01 01:18:54 <_mjones> The biggest payback is always to be found in the app, though. Either implementation or architecture.
2015-04-01 01:20:06 <_mjones> Some days, though, you just need to serve 20k hits per second.
2015-04-01 01:21:32 <Mo0O> thanks to AmatCoder for sharing his solution|alternative for flashplayer
2015-04-01 01:22:31 <Mo0O> I have made a gist which resume it: https://gist.github.com/WnP/21673281e68e2c0f9214
2015-04-01 01:24:00 <_mjones> Mo0O: it's working for you?
2015-04-01 01:24:09 <Mo0O> yes _mjones
2015-04-01 01:24:18 <Mo0O> tested also with grsec :)
2015-04-01 01:24:31 <_mjones> Kudos to AmatCoder. I would have guessed the solution would be uglier and more steps.
2015-04-01 01:24:38 <Mo0O> +1
2015-04-01 01:24:51 <Mo0O> me too
2015-04-01 01:50:55 <eve_> Is Alpine Linux recommended to use as a minimalist desktop?
2015-04-01 01:51:08 <owen1> i noticed a file called .ash_history that is being created. i have to delete it since otherwise my docker build fails - IOError: [Errno 13] Permission denied: './.ash_history'
2015-04-01 01:51:11 <eve_> This is assuming you're going for a bloatless install.
2015-04-01 01:51:18 <owen1> is there a way to tell alpine to stop creating this file?
2015-04-01 01:57:47 <Mo0O> eve_: I currently use alpine as a minimalist desktop -which is my all day one-
2015-04-01 01:57:50 <_mjones> owen1: uh, I think it's 'unset history'.
2015-04-01 01:58:17 <owen1> _mjones: i'll try it. thank you
2015-04-01 01:58:28 <_mjones> eve: I don't run desktop at the moment, but I think Alpine would make an excellent minimalist one. My no-GUI VMs use about 38mb RAM after boot on x86-64.
2015-04-01 01:58:51 <_mjones> owen1: best we should google it. All shells leave .history files by default.
2015-04-01 01:59:23 <owen1> also, what's the best place to put my aliases file? .bash_aliases? and do i have to RUN source /root/.bash_aliases ?
2015-04-01 01:59:33 <_mjones> eve: an xfce desktop will use more, but I'll let someone else tell you how much.
2015-04-01 01:59:48 <Mo0O> eve_: by the way you will have to do some tricky things to make running proprietary applications like flash or skype
2015-04-01 02:00:04 <_mjones> owen1: I usually do just .aliases so I can call them from any shell
2015-04-01 02:00:16 <eve_> Oh, I was just planning on running dwm with Chromium and St
2015-04-01 02:00:20 <_mjones> if it's docker, though, you could just do the aliases in the docker-file
2015-04-01 02:00:31 <_mjones> I'd do it in the docker-file so it documents itself.
2015-04-01 02:00:47 <_mjones> 
ACTION might add something about that to his documentation presentation.

2015-04-01 02:01:12 <Mo0O> eve_: so, you will love it ;)
2015-04-01 02:01:46 <Mo0O> eve_: at the moment there's no pkg for chromium
2015-04-01 02:02:10 <_mjones> I don't see a chromium aport.
2015-04-01 02:02:13 <owen1> _mjones: oh. i didn't think about it. i planned to have .bash_aliases in the folder that i share with -v
2015-04-01 02:02:34 <eve_> Firefox will have to do then :(, oh well I've been needing to try it again for a while.
2015-04-01 02:02:39 <owen1> _mjones: correct me if i am wrong, it might be better not to even use -v, so my docker containers will be self contained. just a guess.
2015-04-01 02:02:39 <_mjones> well I guess you need it for other shell invocations.
2015-04-01 02:03:07 <_mjones> Self contained is good but I can't answer the actual question without looking at docker docs.
2015-04-01 02:23:22 <Mo0O> is there an aport for scala?
2015-04-01 02:23:27 <Mo0O> looks not
2015-04-01 02:23:40 <Mo0O> correct me if I'm wrong :)
2015-04-01 02:27:33 <_mjones> no, you're correct.
2015-04-01 02:28:40 <Mo0O> ok
2015-04-01 02:29:28 <_mjones> While I'm curious about how easy it would be, the last five times I've found out it's been a big porting job. Not done with most of them.
2015-04-01 02:29:52 <_mjones> scala compiles to jvm bytecode, right?
2015-04-01 02:32:58 <Mo0O> _mjones: it looks like, I was asking about it because I want to learn it… so I'm not an expert :)
2015-04-01 02:35:45 <_mjones> I'm building openjdk now. Had a problem doing so earlier and assumed it was due to -DFORTIFY_SOURCE. Maybe not though, it pulls in all kinds of dependencies from within the package itself...
2015-04-01 02:37:19 <_mjones> No, I think I'm just seeing things. Regular dependency builds.
2015-04-01 02:38:01 <_mjones> If a build fails it deletes the dependencies installed, that's what I'm seeing.
2015-04-01 02:40:43 <_mjones> right, I have to install 90% of userland to compile openjdk. Makes sense. :P
2015-04-01 02:41:19 <_mjones> Mo0O: I was just thinking there are zero games on Alpine also.
2015-04-01 02:42:51 <Mo0O> _mjones: yes I guess so
2015-04-01 02:43:31 <Mo0O> but that's not realy a problem for me :D
2015-04-01 02:44:31 <Mo0O> 
ACTION doesn't play video games

2015-04-01 02:45:57 <BitL0G1c> vga_passthrough on kvm or xen to ubuntu guest is probably best for games (steam)
2015-04-01 02:47:53 <_mjones> I haven't worked with hardware passthroughs yet. They cause complications with live-migrations in general.
2015-04-01 02:48:08 <_mjones> BitL0G1c, does Steam work with musl?
2015-04-01 02:49:17 <BitL0G1c> I doubt it
2015-04-01 02:49:34 <BitL0G1c> i can't even get xpra working
2015-04-01 02:50:34 <BitL0G1c> vga_passthrough - arch forums have some instructions to do it with KVM
2015-04-01 02:52:01 <_mjones> I'm curious now as to what steam officially supports.
2015-04-01 02:52:56 <_mjones> "Currently, Steam for Linux is only supported on Ubuntu 12.04 LTS or 12.10 with the Unity, Gnome, or KDE desktop. Additional distributions will be examined for support as time permits."
2015-04-01 02:52:59 <BitL0G1c> ubuntu i think it targets for linux
2015-04-01 02:53:20 <_mjones> Could be a lot worse. Still, I use lxde or xfce when on ubuntu or debian.
2015-04-01 02:53:40 <BitL0G1c> or run windows with vga_passthrough https://bbs.archlinux.org/viewtopic.php?id=162768
2015-04-01 02:54:06 <_mjones> Although I'm due for another round of trying things out. I'll give i3 another shot.
2015-04-01 02:54:19 <_mjones> I pretty much don't run windows, and pretty much never have.
2015-04-01 02:54:48 <_mjones> I do have an 8.1 pro vm on storage, but it makes me angry every time I try to use it.
2015-04-01 02:54:56 <BitL0G1c> win only for games
2015-04-01 02:55:19 <_mjones> I never ran windows. Technically speaking, the first computer I bought myself was a sun 4/280.
2015-04-01 02:56:48 <BitL0G1c> BBC model B
2015-04-01 02:57:01 <_mjones> I was quite happy with an xbox360 for my occasional gaming until they started nonsense requiring live logins and things.
2015-04-01 02:57:16 <_mjones> I never saw a beeb computer; I was stateside.
2015-04-01 02:57:19 <_mjones> Nor acorn.
2015-04-01 02:57:41 <_mjones> I had 8-bit toys long before the sun, but my family owned those.
2015-04-01 02:58:13 <_mjones> In retrospect I should have gotten an amiga.
2015-04-01 02:59:23 <_mjones> Compiling openjdk is taking about 170 package installs as dependencies. _the_horror_.
2015-04-01 03:03:33 <BitL0G1c> temporarily installing the dependencies outside of the fakeroot saves a bit of time if there are build problems
2015-04-01 03:22:28 <cromo> Hey all, I'm trying to install a package and am getting an unsatisfiable constraints error, and I'm looking for help.
2015-04-01 03:23:29 <cromo> I'm running inside the official docker alpine:3.1 image, and attempting to run `apk add unzip`
2015-04-01 03:25:06 <cromo> The exact error I'm getting is: ERROR: unsatisfiable constraints:
2015-04-01 03:25:06 <cromo>   unzip (missing):
2015-04-01 03:25:06 <cromo>     required by: world[unzip]
2015-04-01 03:26:25 <cromo> The reason I'm trying to install unzip is because busybox's version of it is giving "unzip: short read" in an install script I'm running.
2015-04-01 03:26:26 <BitL0G1c> apk update
2015-04-01 03:27:15 <BitL0G1c> then apk add ....
2015-04-01 03:27:34 <cromo> Same error.
2015-04-01 03:28:10 <BitL0G1c> 'setup-apkrepos'
2015-04-01 03:36:02 <cromo> Tried a few with `apk update` inbetween the setup and add calls and I'm still getting the same error.
2015-04-01 03:39:08 <_mjones> cat /etc/apk/repositories
2015-04-01 03:41:14 <cromo> http://pastebin.com/xw5XzxDe
2015-04-01 03:42:49 <_mjones> looks fine.
2015-04-01 03:43:14 <_mjones> which package are you installing?
2015-04-01 03:50:44 <cromo> I'm installing unzip via `apk add unzip`
2015-04-01 04:01:20 <_mjones> oh I thought you were installing something else and unzip was a dependency
2015-04-01 04:01:28 <_mjones> does the network connectivity work?
2015-04-01 04:01:48 <_mjones> obviously those apk repositories are all online sites.
2015-04-01 04:02:16 <cromo> Network connectivity works; I've managed to install other things (build-base among them)
2015-04-01 04:17:25 <_mjones> this should be something simple.
2015-04-01 04:20:36 <cromo> I can't imagine what would cause the error other than a circular dependency for the package
2015-04-01 04:21:02 <cromo> I don't know what the difference is between unzip and world[unzip] though
2015-04-01 04:21:38 <cromo> For now, I'm trying to build the unzip program manually, and that seems to get me past needing to install it via apk
2015-04-01 04:21:59 <cromo> But it would be so much nicer if it worked
2015-04-01 04:24:48 <_mjones> no, world[unzip] is just the way it errors.
2015-04-01 04:26:00 <cromo> So if I'm reading it right, it's missing no dependencies but it thinks it is.
2015-04-01 04:31:50 <_mjones> it cant satisfy the dependency somehow.
2015-04-01 04:32:04 <_mjones> you have multiple repo mirrors...
2015-04-01 04:35:59 <cromo> It doesn't work even if there's only one; that was what I tried first
2015-04-01 04:36:45 <_mjones> multiple is good.
2015-04-01 04:38:42 <cromo> Well, I'm over that hump; now I'm getting "Error relocating /usr/local/lib/lua/5.3/socket/core.so: lua_pcallk: symbol not found"
2015-04-01 04:39:35 <cromo> This is when runtime loading a dynamic library
2015-04-01 04:42:00 <cromo> Running `ldd lua/5.3/socket/core.so` gives a whole slew of the same error, lots of "Error relocating ... symbol not found"
2015-04-01 04:42:02 <_mjones> was the binary compiled on alpine?
2015-04-01 04:42:10 <cromo> It was, I just built it.
2015-04-01 04:42:29 <_mjones>  /usr/local/lib is an unusual location, unless you set it
2015-04-01 04:42:34 <_mjones> er /usr/local.
2015-04-01 04:44:18 <_mjones> I only have one files in /usr/lib/lua/5.2/
2015-04-01 04:45:28 <cromo> So perhaps I need to set the base to something other than the default
2015-04-01 04:46:10 <_mjones> no
2015-04-01 04:46:42 <_mjones> can you tell me what you built? I want to look at the APKBUILD file.
2015-04-01 04:47:27 <cromo> I'm not building apk stuff, I'm building from the main lua release
2015-04-01 04:47:37 <cromo>  /usr/local/lib/lua/5.3 is the default install location when building from source
2015-04-01 04:48:32 <cromo> Really, what I'm trying to do is build and install lua, luarocks, and luasocket (to test dynamic loading)
2015-04-01 04:48:42 <_mjones> oh. language repos. I'm....not a fan.
2015-04-01 04:49:10 <_mjones> They inevitably conflict with os-dist repos.
2015-04-01 04:50:05 <__number5__> lua libs are very tricky to build since they have non-standard config
2015-04-01 04:50:26 <cromo> Well, the intent is to make a single purpose docker image for basic lua development; I don't need to worry about conflicting with official packages
2015-04-01 04:51:15 <__number5__> cromo: oh, docker image will be much simpler, I might able to find you some sample dockerfile
2015-04-01 04:51:17 <_mjones> cromo: that would seem the least-bad way to employ language repos. Still, you have something going on.
2015-04-01 04:51:25 <_mjones> Did you install the apk for the base lua5.3?
2015-04-01 04:52:57 <cromo> __number5__, if you can find an alpine based lua with luarocks that'll successfully require luasocket, I'd be grateful
2015-04-01 04:53:31 <cromo> _mjones, I did not. Searching the package list on pkgs.alpinelinux.org only shows a lua 5.1.5 package
2015-04-01 04:55:25 <_mjones> cromo: I checked 3.1.3 and it has full sets of packages for 5.1 and 5.2. 'Edge', or the beta/alpha development build that typically incorporates the 'testing' repo, also has basic packages for 5.3:
2015-04-01 04:55:37 <_mjones> lua5.3-5.3.0-r1 lua5.3-dev-5.3.0-r1 lua5.3-doc-5.3.0-r1 lua5.3-libs-5.3.0-r1 lua5.3-libs-5.3.0-r1
2015-04-01 04:55:40 <_mjones> It's fine though.
2015-04-01 04:56:03 <_mjones> I'm concerned about your dynamic linking issue.
2015-04-01 04:56:15 <cromo> I'm not picky about the version quite yet; I could try one of the other ones too.
2015-04-01 04:58:43 <cromo> Though pulling from official repos is somewhat preferable
2015-04-01 04:58:59 <_mjones> 'apk search lua | sort'
2015-04-01 04:59:12 <_mjones> If using 3.1, I don't suggest using any repo but main.
2015-04-01 05:03:08 <__number5__> _mjones: is the pkgs site not updated http://pkgs.alpinelinux.org/packages?package=lua&repo=all&arch=x86 just found one lua 5.1.3
2015-04-01 05:03:42 <_mjones> perhaps not.
2015-04-01 05:04:42 <_mjones> This is what you should see on 3.1.3:
2015-04-01 05:04:59 <_mjones> actually I'm not going to paste it, it's 164 packages.
2015-04-01 05:06:22 <cromo> I saw the list
2015-04-01 05:06:29 <_mjones> that list must be old. lua-5.1.5-r4 and lua5.1-5.1.5-r0; notice the change in naming convention over time.
2015-04-01 05:06:50 <_mjones> I mean pkgs.alpinelinux.org list.
2015-04-01 05:10:05 <__number5__> hmm, an outdated online pkgs db is no good for anyone
2015-04-01 05:10:06 <cromo> Sigh, I'll have to stop for now; perhaps I can get back to figuring this out tomorrow
2015-04-01 05:10:58 <__number5__> cromo: https://www.refheap.com/99120 this in a script I used in jenkins build to get lua/luarocks/luasockets running
2015-04-01 05:11:29 <__number5__> should works for alpine docker
2015-04-01 05:12:09 <__number5__> WORKSPACE can be set to your home directory or /usr/local
2015-04-01 05:12:24 <_mjones> __number5__: there was a recent problem with the server infrastructure that may be related.
2015-04-01 05:19:58 <cromo> Hrm. A cursory try of that script was unsuccessful.
2015-04-01 05:20:22 <cromo> I'm actually going to stop for tonight. I may be in here in the next few days trying to figure this out.
2015-04-01 05:20:31 <cromo> Thanks for all your help!
2015-04-01 05:20:46 <_mjones> ttyl.
2015-04-01 06:35:59 <owen1> does anyone uses ansible inside alpine? i get 'fatal: [localhost] => module ec2_facts not found in configured module paths.  Additionally, core modules are missing.'
2015-04-01 06:36:31 <owen1> ansible --version => 1.8.4  i used apk to get it.
2015-04-01 06:54:38 <owen1> it looks like ansible from apk comes with no plugins ):
2015-04-01 06:55:11 <owen1> i am trying to install it with pip and get this error: RuntimeError: autoconf error - http://paste.ubuntu.com/10716795/
2015-04-01 06:55:21 <owen1> any tips would be appreciated
2015-04-01 06:55:55 <_mjones> I just don't know -- haven't looked at it yet.
2015-04-01 06:56:17 <owen1> (:
2015-04-01 06:56:28 <owen1> do u ever sleep?
2015-04-01 06:56:31 <_mjones> there's a small article on wiki.alpinelinux.org but I doubt it has what you need.
2015-04-01 07:00:38 <_mjones> I need some sleep.
2015-04-01 07:06:05 <owen1> i am getting some help from #ansible
2015-04-01 07:06:16 <owen1> ansible is looking for a c compiler
2015-04-01 07:06:23 <owen1> can i install one on alpine?
2015-04-01 07:06:36 <owen1> like gcc
2015-04-01 07:13:15 <owen1> installed gcc
2015-04-01 07:13:34 <owen1> there is no python-devel
2015-04-01 07:13:37 <owen1> ):
2015-04-01 07:19:19 <owen1> installed make
2015-04-01 07:19:31 <owen1> pip install ansible => http://paste.ubuntu.com/10716859/
2015-04-01 07:19:40 <owen1> configure: error: C compiler cannot create executables
2015-04-01 07:26:34 <owen1> installed automake. same error
2015-04-01 07:34:11 <ncopa> owen1: install build-base and python-dev
2015-04-01 07:37:16 <owen1> ncopa: thanks. i'll try that
2015-04-01 07:38:58 <owen1> ncopa: http://paste.ubuntu.com/10716923/
2015-04-01 07:42:04 <clandmeter> owen1: do you have setuptools installed?
2015-04-01 07:42:49 <owen1> clandmeter: only py-setuptools
2015-04-01 07:43:04 <clandmeter> ok
2015-04-01 07:43:13 <clandmeter> i remember it was sometimes a missing dep
2015-04-01 07:44:00 <owen1> so far i have gcc, make, automake, build-base and python-dev
2015-04-01 07:44:25 <owen1> a few more and i'll have ubuntu!
2015-04-01 07:44:33 <clandmeter> lol
2015-04-01 07:45:27 <clandmeter> sign up for ubuntu one and let us know :)
2015-04-01 07:45:30 <owen1> UBUNTO MINT
2015-04-01 07:45:35 <owen1> ahha
2015-04-01 07:50:58 <ncopa> owen1: its an issue in pycrypto
2015-04-01 07:51:08 <ncopa> pycrypto + fortify
2015-04-01 07:51:39 <ncopa> fortify-headers
2015-04-01 07:51:40 <owen1> interesting
2015-04-01 07:52:35 <ncopa> i might have a clue...
2015-04-01 07:53:35 <ncopa> and it happens with the py-cypto package too
2015-04-01 07:53:53 <owen1> my alpine is 226MB ..
2015-04-01 07:55:07 <ncopa> how big is a corresponding ubuntu?
2015-04-01 07:56:56 <owen1> not sure. probably 600MB
2015-04-01 08:02:52 <owen1> ncopa: thanks for the help. i am going to sleep but available on irc any time. so feel free to spam me with good news (:
2015-04-01 08:03:09 <ncopa> ok
2015-04-01 08:03:15 <ncopa> i dont know what the issue is
2015-04-01 08:03:38 <ncopa> but i have notified the fortify-headers upstream maintainer
2015-04-01 08:04:00 <ncopa> it might be a bug in our toolchain
2015-04-01 08:07:25 <owen1> ncopa: thanks!
2015-04-01 08:12:58 <ncopa> any docker experts who want contribute with a dockerfile for redmine?
2015-04-01 08:13:33 <ncopa> the idea is to use it for testing if upgrades broke anything
2015-04-01 08:13:50 <ncopa> to test that the redmine apk we ship as .apk is actually lworking
2015-04-01 08:14:56 <ncopa> so it needs to 1. use latest edge repo, 2. use redmine apk and not ruby gems
2015-04-01 08:15:23 <ncopa> it could use sqlite db or mysql
2015-04-01 08:15:29 <ncopa> mariadb that is
2015-04-01 08:15:33 <ncopa> or maybe postgresql
2015-04-01 12:09:28 <ncopa> owen1: good news. fortify-headers-0.3-r1 should fix pycrypto build
2015-04-01 13:53:48 <clandmeter> dunno who requested path search option on pkgs db. its added now.
2015-04-01 16:37:37 <owen1> ncopa: is there any action i can take to solve it or just wait for fortify-headers-0.3-r1?
2015-04-01 16:37:55 <owen1> what is fortify-headers-0.3-r1 ? what apk package is it part of?
2015-04-01 16:38:10 <owen1> fortify-headers-0.3-r1
2015-04-01 16:38:23 <owen1> oh. apk search fortify => fortify-headers-0.3-r1
2015-04-01 16:41:31 <owen1> wow. it worked. apk add python-dev build-base && pip install ansible
2015-04-01 16:41:42 <owen1> ncopa: ^
2015-04-01 16:51:29 <stateless> owen1: nice, fortify-headers are basically an overlay over the system headers to mitigate against some set of buffer overflow attacks
2015-04-01 16:51:53 <owen1> it might be an ansible question but i got error when trying to use ansible with boto:
2015-04-01 16:52:04 <owen1> failed: [localhost -> 127.0.0.1] => {"failed": true}
2015-04-01 16:52:06 <owen1> msg: No handler was ready to authenticate. 1 handlers were checked. ['HmacAuthV4Handler'] Check your credentials
2015-04-01 16:52:18 <owen1> ansible-1.8.4-r0, py-boto-2.35.2-r0
2015-04-01 16:53:03 <owen1> btw, i installed gcc and make but maybe they are not needed
2015-04-01 16:53:13 <owen1> i'll keep them until i solve the issue
2015-04-01 17:22:24 <_mjones> owen1: boto error, I believe.
2015-04-01 17:33:41 <owen1> _mjones: ok
2015-04-01 18:09:55 <mrahmed> Hi every1. how do i install alpinelinux & set it for dual boot with win7
2015-04-01 18:28:06 <owen1> _mjones: works now. .boto config credential file was missing
2015-04-01 18:45:17 <owen1> _mjones: i have RUN unset history in my Dockerfile but i the .ash_history is still being created ): how do i even start my google search for this issue? ash is not super popular (:
2015-04-01 18:51:15 <_mjones> one second.
2015-04-01 18:52:22 <_mjones> well, this will definitely work.
2015-04-01 18:52:35 <_mjones> RUN ln -s /dev/null .ash_history
2015-04-01 18:52:52 <_mjones> that will create a symlink -- hopefully no perm problems -- and never write anything.
2015-04-01 18:52:58 <owen1> ok. i'll add it to my Dockerfile
2015-04-01 18:54:09 <owen1> my images is 266.9 MB - RUN apk add --update iojs py-pip build-base python-dev py-boto && rm -rf /var/cache/apk/**/    RUN pip install awscli ansible
2015-04-01 19:04:54 <owen1> i have ENV PATH /roon/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
2015-04-01 19:05:17 <owen1> and when i sh into my alpine container i see echo $PATH => /roon/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
2015-04-01 19:05:42 <owen1> but which myscript.sh shows nothing and i can't run it. (it's in /root/bin).  any clues
2015-04-01 19:05:44 <owen1> +
2015-04-01 19:05:46 <owen1> ?
2015-04-01 19:16:59 <xoritor> _mjones, did you finish the port of windows 10 yet?
2015-04-01 19:17:19 <xoritor> sorry i just couldnt help spewing something today
2015-04-01 19:31:30 <nsz> owen1: /roon/bin != /root/bin
2015-04-01 19:36:03 <owen1> nsz: ??
2015-04-01 19:36:23 <owen1> oh
2015-04-01 20:46:48 <_mjones> xoritor: I'll be happy enough to finish one of the difficult aports ones on which I've been working.
2015-04-02 02:44:14 <xaocon> anyone know how I can get colemak working on my console? loadkmap doesn't seem to work with the colemak file from the website. the directions on the site says to use loadkeys but i think that's in console-tools which doesn't seem to be in the repo.
2015-04-02 02:58:05 <Mo0O> someone have solve this java issue: the trustAnchors parameter must be non-empty
2015-04-02 02:58:07 <Mo0O> ?
2015-04-02 03:16:31 <xaocon> Mo0O: a quick google shows some answers. i don't know much about alpine but i'd probably try installing the ca-certificates apk if you haven't already.
2015-04-02 03:17:24 <Mo0O> xaocon: allready done ;)
2015-04-02 03:17:41 <Mo0O> and also already googled
2015-04-02 03:18:38 <Mo0O> `update-ca-certificates` return: WARNING: ca-certificates.crt does not contain exactly one certificate or CRL: skipping
2015-04-02 03:18:56 <xaocon> i'm out of ideas. i don't work with java and i'm new with alpine. good luck.
2015-04-02 03:19:14 <Mo0O> thanks xaocon
2015-04-02 03:19:43 <__number5__> xaocon: install-keymap linux_console/colemak.iso15.kmap this don't work?
2015-04-02 03:21:34 <xaocon> install-keymap isn't in my path. i'll check it out and see if i can get it going with that.
2015-04-02 03:26:51 <__number5__> xaocon: do you have a folder like /usr/share/kbd/keymaps/?
2015-04-02 03:28:58 <xaocon> __number5__: no, this is a liveboot of 3.1.3 and doesn't seem to have that included.
2015-04-02 03:30:32 <__number5__> what about /lib/kbd/keymaps?
2015-04-02 03:30:46 <xaocon> installing kbd and kbd-misc doesn't seem to change that either
2015-04-02 03:31:09 <__number5__> try just loadkeys linux_console/colemak.iso15.kmap
2015-04-02 03:32:16 <xaocon> yeah... i have loadkeys now. i guess it's in the kbd package.
2015-04-02 03:32:38 <xaocon> __number5__: thanks that should do it
2015-04-02 03:32:53 <__number5__> np
2015-04-02 04:07:53 <Mo0O> I've fix my problem by copying `/etc/ssl/certs/java/cacerts` form archlinux to `/usr/lib/jvm/java-1.7-openjdk/jre/lib/security/` in alpine
2015-04-02 04:09:07 <Mo0O> don't you think it could be nice to add this file to the openjdk7-jre-base aport?
2015-04-02 04:09:45 <Mo0O> correct me if I'm wrong :)
2015-04-02 04:11:23 <Mo0O> I know that's a binary file, at that sucks… but java already sucks :-°
2015-04-02 07:12:53 <iwearsable> Is it possible to get graphical applications to run from a chroot with alpine grsec as host?
2015-04-02 07:13:30 <iwearsable> I get 'grsec: denied connect() to abstract AF_UNIX socket outside of chroot'. I thought that 'sysctl -w kernel.grsecurity.chroot_deny_chroot=0
2015-04-02 07:13:55 <iwearsable> Sorry, I thought that the above sysctl command might help, but it does not
2015-04-02 07:24:11 <ScrumpyJack> morning
2015-04-02 07:24:14 <iwearsable> I figured it out run 'sudo sysctl -w kernel.grsecurity.chroot_deny_unix=0
2015-04-02 07:24:22 <iwearsable> Howdy
2015-04-02 07:25:29 <ScrumpyJack> nice find. can you put that in the FAQ in the wiki? :)
2015-04-02 07:28:23 <ScrumpyJack> or here http://wiki.alpinelinux.org/wiki/Chroot
2015-04-02 07:29:38 <iwearsable> Sure thing, I'll add it to the section on graphical applications in the article above. Still working out some other grsec issues that happen when I try to start something complex like say Midori
2015-04-02 07:30:49 <ScrumpyJack> there is a section in there about grsec preperation
2015-04-02 07:46:58 <iwearsable> For now I put it here but if I find some other things to add I may put it with grsec preparation http://wiki.alpinelinux.org/wiki/Chroot#Run_graphical_applications_from_the_chroot
2015-04-02 09:35:35 <cultureulterior1>  Mo0O https://registry.hub.docker.com/u/denvazh/java/dockerfile/
2015-04-02 13:11:38 <coredumb> all of you docker folks, what are you using as process supervision ?
2015-04-02 14:05:23 <kahiru1> hey
2015-04-02 14:09:01 <BitL0G1c> coredumb - there is supervisor in the repos
2015-04-02 14:10:35 <BitL0G1c> http://supervisord.org
2015-04-02 15:33:19 <kahiru1> i was wondering, if i run alpine on rpi, it boots to ram? could i then take the sdcard out and put there another one as storage?
2015-04-02 15:34:31 <clandmeter> kahiru1: you could, but you have to umount modloop
2015-04-02 15:34:58 <clandmeter> you wont be able to modprobe anymore
2015-04-02 15:35:37 <clandmeter> you cannot boot from the storage sd?
2015-04-02 15:36:59 <kahiru1> i have one large sdcard (originally bout that for my tablet but somehow it doesnt work) and smaller one with alpine apkovl and apk cache
2015-04-02 15:37:15 <kahiru1> i was just curious if i could hotswap them or have to reformat the larger one
2015-04-02 15:37:41 <clandmeter> if nothing is mounted from the sd, you can just pull it out
2015-04-02 15:38:03 <clandmeter> but like you said, you will also loose apkovl and cache
2015-04-02 15:40:04 <clandmeter> I think i created a primary partition of ~ 200 mb for boot/apk/cache/ovl
2015-04-02 15:40:35 <clandmeter> the rest used for system like var
2015-04-02 15:41:36 <kahiru1> makes sense
2015-04-02 15:43:10 <clandmeter> or use lxc
2015-04-02 15:43:48 <clandmeter> and have the containers on sd
2015-04-02 15:52:56 <kahiru1> nah, im not really a container person
2015-04-02 18:01:35 <coredumb> BitL0G1c: ok
2015-04-02 18:56:19 <_mjones> I'm going to mount the soapbox for a moment and remind everyone that blocking icmp, and ping (echo request/echo reply) in particular, is pointless and people who do it should be deeply ashamed.
2015-04-02 18:57:44 <_mjones> I have a situation right now where I can't test a server with ping. Earlier in the same crossconnect setup I was assured I could test the other end's router with ping, but it wasn't working and I strongly suspect some misguided icmp blocking was happening.
2015-04-02 18:58:56 <_mjones> So I say again, blocking icmp is pointless, you're mostly hurting yourself (especially if you end up breaking pmtud) but everyone else is going to have to put in extra work because of it.
2015-04-02 18:59:17 <_mjones> It doesn't do anything for security, but if you feel better you can rate-limit all icmp and be done with it.
2015-04-02 18:59:31 <_mjones> I definitely won't respect you if you're blocking icmp.
2015-04-02 19:00:49 <_mjones> I despise cargo-culting in infosec.
2015-04-02 19:17:03 <coredumb> _mjones: network team here find it interesting to block icmp on routers
2015-04-02 19:17:19 <coredumb> how fun it is to traceroute issues -_-
2015-04-02 19:22:20 <_mjones> that's precisely the nonsense to which I refer. Utterly pointless cargo-cult security masquerading as 'best practises'.
2015-04-02 19:23:25 <_mjones> Then separately are yesteryear's best practises, to which most security and network people cling desperately. I was fulltime network ops for years, long ago, but I wouldn't implement a net like that today.
2015-04-02 19:24:31 <_mjones> the big zmap survey of the internet found, iirc, 36 million devices which didn't ping but which responsed on at least one port. I cannot express the disdain.
2015-04-02 19:27:21 <coredumb> _mjones: oh and default gw must be changed from .254 to .1
2015-04-02 19:27:25 <coredumb> cause you know
2015-04-02 19:27:30 <coredumb> more secure indeed
2015-04-02 19:27:32 <coredumb> ...
2015-04-02 19:27:34 <coredumb> -_-
2015-04-02 19:28:57 <coredumb> "oh and ... yes we couldn't do it in all subisdiaries... so half on .254 half on .1... have fun configuring your servers mate"
2015-04-02 19:29:48 <_mjones> consistency is important in i.t. ops but I see a lot of cargo-culting about that too. They're totally forgetting _why_ they're doing these things.
2015-04-02 19:30:33 <coredumb> forgetting ?
2015-04-02 19:30:38 <_mjones> I prefer .1 myself and reserve the first ten of a /24 or larger, but that's purely aesthetics. It's easy enough to figure out from dhcp or arp replies of v6 ra.
2015-04-02 19:30:39 <coredumb> do they even knew ?
2015-04-02 19:30:45 <coredumb> i don't think so
2015-04-02 19:31:24 <_mjones> Much of it boils down to principle of least privilege but it's virtually always at the expense of simplicity.
2015-04-02 19:32:16 <coredumb> _mjones: that's the point no matter what you gw is, there's so much ways to find a gw it doesn't mean jack setting it a specific IP for the sake of _security_
2015-04-02 19:32:20 <_mjones> I don't run a lot of vlans; even years ago I'd run up to a /22 for client office lans. Network traffic is almost all tcp so collision domains don't need to be /24.
2015-04-02 19:32:42 <_mjones> I ignored the security part of your remark because I can't even fathom the stupidity.
2015-04-02 19:33:37 <_mjones> Anyone who told you that either didn't tell you about some factor that might make it more secure, like maybe there's an ACL everywhere for all IPs below 10, or they don't know what they're talking about and won't admit it.
2015-04-02 19:33:38 <coredumb> that's the point
2015-04-02 19:33:39 <coredumb> :)
2015-04-02 19:34:08 <_mjones> someone is repeating something from someone, and at least one of them doesn't know that they don't know.
2015-04-02 19:34:17 <_mjones> But then that's almost the definition of cargo-cult, neh?
2015-04-02 19:34:31 <coredumb> they don't know jack but pretend to because you know - <put the name you like> certified blahblahblha -
2015-04-02 19:34:51 <coredumb> that the definition of dumbass bulshit
2015-04-02 19:35:07 <_mjones> I have to work not to mentally down-rate people who have certs.
2015-04-02 19:35:28 <coredumb> i don't mind ppl having certs
2015-04-02 19:35:40 <coredumb> i certainly do those that brag about them :)
2015-04-02 19:35:52 <_mjones> I used to have a team member that liked certs, in large part because he found the process of getting one made him disciplined to learn the tech. He was good. Not the best, but good.
2015-04-02 19:36:27 <coredumb> well
2015-04-02 19:37:05 <coredumb> i use to say that "you need cert when you're not good enough to prove yourself"
2015-04-02 19:38:42 <_mjones> Blunt, but generally a good summary.
2015-04-02 19:41:12 <xen_roger> _mjones: sounds funny  ... :)
2015-04-02 19:41:31 <_mjones> xen_roger: which?
2015-04-02 19:42:31 <xen_roger> _mjones: all what you wrote ... icmp, cert ...  just funny - nothing more!
2015-04-02 19:46:55 <_mjones> I want to evangelise a specific item and to discourage on the strongest terms the opposite bit of stupidity which is _somehow_ regarded as a good idea. I should create a blog, and one of my favourite writing topics is anti-patterns t o avoi, but in the meantime if anyone has suggestions on convincing the world not to block ICMP, I'm all ears.
2015-04-02 19:47:45 <_mjones> This no echo response thing was funny for a minute in 1995, but clearly some have taken the joke far too far.
2015-04-02 19:49:03 <_mjones> I need to write a guest editorial someplace with a lot of traffic, where the readership is largely empowered to make these sorts of decisions.
2015-04-02 19:50:09 <xen_roger> it just sounds you do not unterstand why some people block icmp. sure, there are a lot of them how think "no icmp - so I am invisible" - but! why should you not filter icmp, like you do with tcp and udp and all the other protocols? blocking in general is mostly not really clever - filtering does make sense!
2015-04-02 19:50:33 <xen_roger> s/how/who/g
2015-04-02 19:51:01 <coredumb> block != filter
2015-04-02 19:55:02 <xen_roger> filtering is a more specific term of block!
2015-04-02 19:58:00 <_mjones> xen_roger: filtering is pretty much blocking, I agree.
2015-04-02 19:58:26 <_mjones> xen_roger: what benefit do you think you're getting by blocking icmp? Remember though that I said a default icmp rate-limit is fine.
2015-04-02 19:58:59 <_mjones> global default icmp rate-limit. To maybe 1% or .001% of bandwidth
2015-04-02 19:59:35 <_mjones> If I know why people do it, it will help me convince them to stop doing it.
2015-04-02 20:00:01 <xen_roger1> _mjones: prevent tunneling e.g.
2015-04-02 20:01:14 <xen_roger1> or just to not answer with "port unreachable" if there is no service to reach on specific port...
2015-04-02 20:01:32 <xen_roger1> not knowing why they do that, means not they have no reason
2015-04-02 20:01:57 <_mjones> Everything has a reason. These are just absolutely awful reasons.
2015-04-02 20:02:54 <_mjones> I can always find any service that my src address can reach, for obvious reasons. blocking icmp port unreachable doesn't help security at all.
2015-04-02 20:03:10 <_mjones> it just makes things hard to debug, and adds uncertainty to legitimate users.
2015-04-02 20:04:02 <_mjones> icmp administratively blocked  should be used, in many cases, instead of unreahables. silent drops should be avoided because they cause performance problems and uncertainty for no benefit.
2015-04-02 20:07:14 <xen_roger1> so ping is your debuging tool ? mabey tool nr. 1?
2015-04-02 20:10:54 <xen_roger1> it also depends, performance problems are maybe welcome - simple and I mean simple example is to slow down a port scan. not in case of any kind of security - not to get me wrong. in a cooperate network there should also be other options to verify the host I would reach is up if the service I want to use does not answer or acting not the way it should. so no ping is no debug or diagnostic tool!
2015-04-02 20:12:22 <_mjones> slow down port scan? these things are automated. I don't need to port scan before sending an exploit if I don't want.
2015-04-02 20:12:41 <_mjones> ping is a great debug tool, but I use them all.
2015-04-02 20:12:56 <_mjones> I like to run lldp or cdp on hosts, too. Very handy.
2015-04-02 20:22:53 <xen_roger1> sorry, but you statements are to general. host mostly do not run lldp or cdp... - most of them. ping isn´t a diagnostic tool - it just says host is up or not and if up the delay of the pks. nothing more. how could it help you finding the problem? not ping does not mean host is down - if ping is working it does not mean services running at this host are also fine. so ping does not help you to troubleshoot failure.
2015-04-02 20:25:47 <xen_roger1> long story short - I do not agree with you and non of your arguments seem to be valid.
2015-04-02 20:28:25 <_mjones> xen_roger: I'm setting up a new intercage cross-connect right now, and not being able to ping far router could mean that something is wrong at layer 3, or it could mean that far end security policy blocks icmp. Do I keep looking for problem on my end or send mail to far end and wait?
2015-04-02 20:28:36 <_mjones> ping is absolutely a diagnostic tool.
2015-04-02 20:29:07 <_mjones> Link is up on layer2, but I'm not getting traffic back.
2015-04-02 20:29:55 <_mjones> I honestly can't believe that people think blocking icmp echo reply is beneficial, and mor eimportantly, that the benefits outweigh the costs.
2015-04-02 20:30:17 <owen1> i have RUN ln -s /dev/null /root/.ash_history   but when i sh into my container ls -al /root  doesn't show me the symlink. any ideas?
2015-04-02 20:32:47 <_mjones> You must already have a .ash_hitory at that point. Use ln -af to force it.
2015-04-02 20:32:55 <_mjones> I just checked and that works.
2015-04-02 20:33:41 <xen_roger1> owen1: chmod -w /root/.ash_history
2015-04-02 20:36:10 <owen1> xen_roger1: i tried adding RUN chmod -w /root/.ash_history after RUN ln -vsf /dev/null /root/.ash_history    but it didn't help
2015-04-02 20:39:37 <xen_roger1> owen1: RUN means Dockerfile?
2015-04-02 20:40:37 <xen_roger1> owen1: what about not creating symlink and just chmod the "real" file
2015-04-02 20:41:11 <owen1> xen_roger1: let me try!
2015-04-02 20:41:20 <_mjones> ln -af works
2015-04-02 20:41:43 <_mjones> RUN ln -fs /dev/null /root/.ash_history
2015-04-02 20:41:55 <_mjones> Not the cleanest answer, but will work for now.
2015-04-02 20:44:13 <rpu3uO8PEVZZ> hmmm, i just checked and apparently firehol disables icmp by default.
2015-04-02 20:45:49 <_mjones> rpu3uO8PEVZ: good info. Changing defaults would be very helpful. I need to convince the maintainers to do so, and also convince their userbase.
2015-04-02 20:47:39 <_mjones> rpu3uO8PEVZZ: please tell me it doesn't block icmp fragmentation needed.
2015-04-02 20:49:32 <_mjones> http://www.netheaven.com/pmtu.html
2015-04-02 20:49:41 <_mjones> http://www.znep.com/~marcs/mtu/
2015-04-02 20:50:13 <_mjones> From the latter: There is one solution, and several workarounds, for this problem. They include:      Fix your filters! The real problem here is filtering ICMP messages without understanding the consequences. Many packet filters will allow you to setup filters to only allow certain types of ICMP messages through. If you reconfigure them to let ICMP can't fragment (type 3, code 4) messages through, the problem should disappear. 
2015-04-02 20:50:58 <_mjones> It's hell to get the other end to accept that they need to not block icmp, even while they're complaining that they can't reach your servers. Trust me.
2015-04-02 20:58:16 <rpu3uO8PEVZZ> _mjones: i think it does not discriminate, all icmp is blocked. but haven't checked the code
2015-04-02 20:59:23 <owen1> xen_roger1: chmod: /root/.ash_history: No such file or directory
2015-04-02 20:59:34 <owen1> _mjones: RUN ln -fs /dev/null /root/.ash_history    still create the file
2015-04-02 20:59:53 <owen1> it doesn't add a symlink
2015-04-02 21:00:27 <owen1> http://paste.ubuntu.com/10726438/
2015-04-02 21:01:05 <owen1> sorry  - http://paste.ubuntu.com/10726441/
2015-04-02 21:02:33 <_mjones> rpu3uO8PEVZZ: hugely bad. A user doing this is shooting themselves in the foot, and causing problems for others when she or he demands things be 'fixed'.
2015-04-02 21:14:47 <_mjones> rpu3uO8PEVZZ: I see these things a lot. Another one is blocking tcp/53 because you think only zone transfers use tcp.
2015-04-02 21:36:05 <bOSKE> will forum ever work again?
2015-04-02 21:45:30 <clandmeter> bOSKE: yes
2015-04-02 21:45:41 <clandmeter> i will restore it soon
2015-04-02 21:46:54 <bOSKE> clandmeter, what was the problem?
2015-04-02 21:47:51 <clandmeter> there is no problem
2015-04-02 21:48:15 <clandmeter> we have migrated to infra box to 3.1 and lxc
2015-04-02 21:48:31 <clandmeter> s/to/the
2015-04-02 21:49:09 <clandmeter> and its the last container to restore
2015-04-02 21:49:20 <clandmeter> i will also do some maintenance to it.
2015-04-02 21:51:47 <clandmeter> we were looking into alternatives, but it seems there are not many good ones.
2015-04-02 21:52:29 <bOSKE> k,just wondering see forum is not working for a while..
2015-04-02 21:58:19 <darkfader> what's the last word on rapsberry pi 2?
2015-04-02 21:59:06 <clandmeter> send one to fabled and he will take care of it :)
2015-04-02 22:01:26 <clandmeter> i think he mentioned as long as he doesnt have the device, he cannot add support.
2015-04-02 22:01:48 <clandmeter> but he was interested in getting one. so could be its already on the way.
2015-04-02 22:01:57 <clandmeter> not sure what current stock is.
2015-04-02 22:12:46 <systmkor> hey all I'm writing some initial boostrapping scripts for linux boxes that I'm building
2015-04-02 22:12:52 <systmkor> any suggestions on a proper method of doing this
2015-04-02 22:13:26 <systmkor> meaning one example a friend of my suggested is basically an init script that fetchs the current bootstrap script
2015-04-02 22:13:28 <systmkor> and then executes it
2015-04-02 22:15:20 <systmkor> basically going from a base Alpine install to a set of packages, configurations, services enabled, custom code compiled and added, etc.
2015-04-02 22:15:40 <systmkor> all of that automated
2015-04-02 22:16:08 <systmkor> this is mainly for development on ARM dev boards at moment
2015-04-02 22:16:56 <clandmeter> systmkor: yes we have a user who adds notes to his wiki profile how to do that.
2015-04-02 22:17:11 <clandmeter> i think its Jean-Scotch
2015-04-02 22:20:21 <clandmeter> systmkor: http://wiki.alpinelinux.org/wiki/User_talk:Jch
2015-04-02 22:26:26 <knite> what's the right way to request a package version bump? ioJS released 1.6.3 and edge has 1.6.2.
2015-04-02 22:28:34 <systmkor> clandmeter: thanks, looks like related material but not exactly what I need at the moment
2015-04-02 22:28:53 <systmkor> clandmeter: will continue to go through the wiki article
2015-04-02 22:29:09 <owen1> ansible -i hosts donate-1 -m ping => http://paste.ubuntu.com/10726925/
2015-04-02 22:29:22 <owen1> (issue when trying to run ansible script)
2015-04-02 22:32:17 <_mjones> knite: yes, that happened a day or two ago.
2015-04-02 22:32:19 <owen1> how to install ssh?
2015-04-02 22:32:25 <owen1> acf-openssh-0.9.0-r0?
2015-04-02 22:32:34 <_mjones> 'apk add openssh' or 'apk add dropbear'
2015-04-02 22:32:42 <_mjones> no, acf is a management framework.
2015-04-02 22:33:01 <knite> _mjones: right, so is there a mechanism by which I can request that the package be bumped?
2015-04-02 22:33:03 <_mjones> dropbear is an alternative lightweight ssh daemon, but the one you're more likely familiar with is openssh.
2015-04-02 22:33:21 <owen1> _mjones: thanks
2015-04-02 22:33:26 <_mjones> knite: there's a bugtracker but asking on #alpine-devel is the way to go.
2015-04-02 22:33:33 <_mjones> knite: consider it requested.
2015-04-02 22:33:55 <_mjones> ok, maybe not 'the way to go', but let's say efective.,
2015-04-02 22:35:05 <knite> _mjones: excellent. :-) second request: mainline nginx in addition to stable.
2015-04-02 22:35:43 <_mjones> knite: I'll compile 1.6.3 now, the version bump is probably trivial.
2015-04-02 22:37:31 <knite> _mjones: iojs has been following a "move fast and break things" philosophy. changelog doesn't have any red flags, but I've been trying to keep up with releases so long as they're becoming more stable.
2015-04-02 22:38:39 <_mjones> knite: I threw the nginx question out to #alpine-devel.
2015-04-02 22:39:21 <_mjones> knite: technically iojs has 6 or 7 failing tests ('make test' tests) but there's no reason to fuss about 1.6.3 as long as there's no regression.
2015-04-02 22:39:45 <knite> _mjones: hm, will stand pat, then!
2015-04-02 22:40:22 <_mjones> I mean I'm compiling on my system now to make sure 1.6.3 works as least as well, then intend to submit the version bump.
2015-04-02 22:47:42 <agd> hello.
2015-04-02 22:47:47 <agd> anybody around?
2015-04-02 22:48:10 <_mjones> yes.
2015-04-02 22:48:22 <agd> _mjones; do you love alpine?
2015-04-02 22:48:50 <_mjones> i like it. am i on candid camera?
2015-04-02 22:49:07 <agd> _mjones; more than arch and openbsd combined? (the lightweightness of arch and the security of openbsd)?
2015-04-02 22:49:41 <agd> _mjones; i haven't updated or changed my distro since 2.5 years now
2015-04-02 22:49:45 <_mjones> I'm not going to make that judgement. I will point out that arch has systemd and udev by default.
2015-04-02 22:49:58 <agd> _mjones; and quite blatantly, i need a change
2015-04-02 22:50:12 <agd> im looking for a distro that is miserably secure and stupendously lightweight
2015-04-02 22:50:13 <_mjones> alpine uses openrc, and while udev is installable, doesn't have udev by default. Not that udev is heavy, but still.
2015-04-02 22:50:44 <agd> _mjones; how much memory does alpine use upon first-time startup?
2015-04-02 22:50:57 <_mjones> desktop? I believe the grsec/pax kernel doesn't play nicely with X.org.
2015-04-02 22:51:20 <agd> does alpine not have selinux/apparmor or anything else?
2015-04-02 22:51:40 <_mjones> mjones: x86-64 running dropbear (openssh is equally available) base install is 38MB or less of RAM.
2015-04-02 22:52:05 <_mjones> alpine has grsec, pax, fortify_source, and so forth.
2015-04-02 22:52:18 <agd> _mjones; alpine has grsec/pax installed by default?!
2015-04-02 22:52:23 <_mjones> yes.
2015-04-02 22:52:47 <agd> very nice
2015-04-02 22:52:56 <agd> _mjones; im assuming you use xfce?
2015-04-02 22:54:01 <_mjones> I don't currently run alpine on the desktop, but I run xfce and lxde on unix machines running X, yes.
2015-04-02 22:54:28 <_mjones> I just rebooted a VM to check memory consumptin with openssh.
2015-04-02 22:54:59 <BitL0G1c> xfce is very quick on alpine - much quicker than on debian
2015-04-02 22:55:25 <_mjones> I like alpine a lot but the next Linux install I do on a desktop might be arch. one of the factors is that I think systemd might be worthwhile on a mobile desktop.
2015-04-02 22:55:49 <agd> BitL0G1c; you have xfce on alpine?
2015-04-02 22:56:02 <_mjones> traditionally I run ubuntu or debian; I only started with Alpine towards the end of last year.
2015-04-02 22:56:08 <systmkor> agd: Alpine has XFCE
2015-04-02 22:56:21 <agd> BitL0G1c; how much RAM does alpine consume with no applications running?
2015-04-02 22:56:26 <BitL0G1c> I've had it working in a vm with http://wiki.alpinelinux.org/wiki/XFCE_Setup
2015-04-02 22:57:01 <systmkor> _mjones: thoughts on porting systemd to work with Alpine?
2015-04-02 22:57:34 <iwearsable> I'm running alpine as a desktop with openbox. Works well. Ram usage on boot to graphical environment is less than 200mb I think
2015-04-02 22:58:20 <BitL0G1c> as a reverse proxy / vpn it uses less than 128meg on kvm
2015-04-02 22:58:47 <_mjones> agd: looks like 24mb ram without sshd, 30mb ram with opensshd started and no connections.
2015-04-02 22:59:28 <iwearsable> Biggest complaint would be lack of some of my favorite software in the repo and problems building stuff myself due to musl. I'm not very knowedgable but I worked around most of that using an archlinux chroot. Even figured out how to get programs from the chroot to play with pax and grsec (using paxctl etc)
2015-04-02 22:59:31 <agd> _mjones; and how much with xfce?
2015-04-02 22:59:58 <iwearsable> I don't think that alpine is the best choice for a hassle-free desktop but at least I'm learning something...
2015-04-02 23:00:22 <agd> i need an absolutely hassle-free desktop
2015-04-02 23:00:35 <systmkor> agd: probably go check out ElementaryOS
2015-04-02 23:00:41 <iwearsable> It's not hassle free unless you have modest needs
2015-04-02 23:00:44 <agd> systmkor; too heavy
2015-04-02 23:00:51 <_mjones> systmkor: systemd doesn't work with musl; dalias says Poettering considers glibc to be a dependency and has no intention or desire of changing that.
2015-04-02 23:01:05 <systmkor> agd: so you want a light weight and hassle free?
2015-04-02 23:01:11 <agd> systmkor; ofcourse
2015-04-02 23:01:18 <systmkor> that's going to be a nearly non-existant find
2015-04-02 23:01:23 <agd> systmkor; hassle-free beyond imagination
2015-04-02 23:01:28 <_mjones> iwearsable: we're working on changing both things, but besides thinking systemd might pay for itself on a laptop, those would be my concerns.
2015-04-02 23:01:29 <systmkor> as of current state of Linux distros
2015-04-02 23:01:33 <agd> systmkor; is alpine slightly hassling?
2015-04-02 23:01:50 <systmkor> _mjones: one guy has patched systemd to work with MUSL
2015-04-02 23:02:11 <systmkor> i'm planning on getting in contact with him to see if he would be interested in maintaining that with some help
2015-04-02 23:02:21 <dalias> uhg, why?
2015-04-02 23:02:22 <_mjones> agd: and you're running Arch now?!
2015-04-02 23:02:36 <dalias> do you actually want to use systemd??
2015-04-02 23:02:38 <BitL0G1c> agd - Antix is quite nice & light - based on Debian Testing
2015-04-02 23:02:51 <systmkor> dalias: I have my various implementation issues with systemd
2015-04-02 23:03:05 <systmkor> but ignoring it doesn't make my issues of its implementation choices go away
2015-04-02 23:03:13 <agd> _mjones; ofcourse
2015-04-02 23:03:15 <systmkor> however porting it and actively working on it so it starts to have better
2015-04-02 23:03:22 <systmkor> implementation does
2015-04-02 23:03:22 <dalias> it seems you could get all the benefits and none of the disadvantages just by writing a utility to use systemd service files with a non-crap supervision system
2015-04-02 23:03:40 <_mjones> dalias: I intend to try systemd so I can at least crticise it from experience as well as theory. I'm not entirely convinced about it, despite having major reservations.
2015-04-02 23:04:02 <systmkor> dalias: well systemd is more than a supervisions system in the runit/supervisord sense
2015-04-02 23:04:15 <dalias> systmkor, yes
2015-04-02 23:04:30 <systmkor> dalias: there are a lot of good efforts/design choices made by systemd
2015-04-02 23:04:30 <dalias> but my approach to a problem like this is much like my approach to musl/libc
2015-04-02 23:04:43 <systmkor> that said I think they made a variety of bad implementation choices
2015-04-02 23:04:51 <dalias> start off asking: "what are the interface surfaces?" "which ones actually matter?"
2015-04-02 23:05:00 <systmkor> dalias: yup
2015-04-02 23:05:18 <dalias> and while systemd is huge and architecturally monolithic and horribly bad design...
2015-04-02 23:05:19 <systmkor> however for something like systemd I would rather push to upstream
2015-04-02 23:05:23 <systmkor> than start from scratch
2015-04-02 23:05:34 <dalias> the interface surfaces that actually matter can be distilled out into something very small
2015-04-02 23:05:38 <systmkor> dalias: it's monolithic and not
2015-04-02 23:06:01 <systmkor> dalias: what would you change about systemd
2015-04-02 23:06:08 <systmkor> but with the same goals in mind
2015-04-02 23:06:27 <dalias> that's a loaded question
2015-04-02 23:06:31 <systmkor> meaning not having a hodgepodge of a variety of crappy DIY scripts kinda sorta maintaining the system
2015-04-02 23:06:33 <dalias> i wouldn't start with systemd and i wouldn't have the same goals
2015-04-02 23:06:47 <systmkor> dalias: wasn't meant to be a loaded questoin
2015-04-02 23:06:52 <dalias> re: hodgepodge of crappy DIY scripts...
2015-04-02 23:07:02 <dalias> https://xkcd.com/927/
2015-04-02 23:07:11 <dalias> that's systemd for you
2015-04-02 23:07:16 <dalias> #15 :)
2015-04-02 23:07:20 <systmkor> kinda sorta
2015-04-02 23:07:35 <dalias> the only way systemd "solves" the problem is through forced adoption and imposing policy
2015-04-02 23:07:36 <systmkor> except all the other major distros are ditching their existing methods for systemd
2015-04-02 23:07:39 <systmkor> so it would be more like
2015-04-02 23:07:41 <dalias> it actively tries to kill all the existing ones
2015-04-02 23:07:42 <systmkor> new standard
2015-04-02 23:07:46 <systmkor> now 10 standards
2015-04-02 23:08:02 <dalias> and replace them with something new that's just as bad or worse
2015-04-02 23:08:14 <systmkor> well so unpack that
2015-04-02 23:08:22 <_mjones> oh, this outfit isn't sending me any icmp. So my tcptraceroute wasn't showing anything useful and I didn't notice when they fixed it because I wasn't testing with telnet. Just incredible.
2015-04-02 23:08:23 <systmkor> because I've been looking into where it succeeds and fails
2015-04-02 23:08:40 <systmkor> dalias: just started though, so limited knowledge
2015-04-02 23:09:18 <_mjones> sysV init is awfully clever actually. way better than traditional BSD, and I'm a big fan of bsd.
2015-04-02 23:09:28 <dalias> anyway if you want to ask me "what i would do different" i think first we'd need to be on the same page about goals
2015-04-02 23:09:37 <systmkor> so yah explain your goals
2015-04-02 23:09:41 <_mjones> systemd is definitely the other extreme.
2015-04-02 23:09:57 <_mjones> dalias, are you familiar with illumos/solaris smf?
2015-04-02 23:10:05 <dalias> _mjones, no
2015-04-02 23:11:03 <systmkor> dalias: so explain goals and design choices
2015-04-02 23:11:08 <systmkor> please :D
2015-04-02 23:11:35 <iwearsable> journald pisses me off, I mean why do I have to install another logging daemon and hook it up instead of just being able to tell journald to save as text? Binary logs cmon now most users don't need that
2015-04-02 23:11:51 <_mjones> not needing special systems to upgrade itself live would be nice.
2015-04-02 23:12:51 <_mjones> I need to sit down and read the design history again. the argument is that all these absorbed systems are necessary, but how true is that?
2015-04-02 23:13:06 <dalias> systmkor, well i need to know where you're coming from
2015-04-02 23:13:17 <systmkor> i'm coming from a pretty empty shell place
2015-04-02 23:13:22 <dalias> to me systemd tries to solve a bunch of problems that were solved correctly decades ago but not adopted
2015-04-02 23:13:30 <systmkor> meaning I honestly have rather limited knowledge on the boot sequence of UNIX and Linux
2015-04-02 23:13:36 <dalias> so i think it has zero useful goals _except_ forced-adoption
2015-04-02 23:13:50 <systmkor> dalias: okay but I need more than generic statement
2015-04-02 23:13:55 <systmkor> examples, hows, and whys
2015-04-02 23:14:16 <dalias> first explain to me what is convincing about systemd to you
2015-04-02 23:14:20 <dalias> why it sounds useful to you
2015-04-02 23:14:24 <dalias> why it's something you would want
2015-04-02 23:14:49 <dalias> then i can answer (1) whether that's actually valid, and (2) if there's a different way i would achieve the same thing
2015-04-02 23:15:03 <systmkor> the one thing that is something I would want is the fact is that you describe your services (as far as I have seen)
2015-04-02 23:15:07 <systmkor> as apposed to coding them
2015-04-02 23:15:10 <systmkor> that is to say init scripts
2015-04-02 23:15:21 <systmkor> which while in some case yes you want more control
2015-04-02 23:15:38 <systmkor> for custom, edge case, fine tuning, or dev reasons
2015-04-02 23:15:59 <systmkor> however for 99.9% of OS use that means general server and laptop, desktop uses
2015-04-02 23:16:00 <_mjones> so idempotency, or desired-state configuration. nixos does that, incidentally.
2015-04-02 23:16:04 <systmkor> thats the exapt opposite of what you want
2015-04-02 23:16:15 <dalias> _mjones, i don't think that's equivalent to what systmkor said
2015-04-02 23:16:34 <systmkor> dalias: it's along the lines but not exactly the same
2015-04-02 23:17:47 <systmkor> _mjones: nixos does that but in the packages and configurations I don't think it does it beyond that (i.e. no one has mentioned it doing anything beyond it)
2015-04-02 23:18:02 <systmkor> dalias: but yes effectively a declaritve language to describe your system
2015-04-02 23:18:05 <_mjones> What's wrong with coding services? Like I said, sysV init is quite clever.
2015-04-02 23:18:18 <systmkor> rather than a Turing complete language
2015-04-02 23:18:39 <systmkor> _mjones: there is nothing wrong with coding services for developers of a distro/OS/dameon/service
2015-04-02 23:18:45 <systmkor> beyond that
2015-04-02 23:18:56 <dalias> so having a declarative process-supervision system?
2015-04-02 23:18:58 <systmkor> you are generally gaining nothing and giving your self a turing complete language
2015-04-02 23:19:00 <systmkor> to hang yourself
2015-04-02 23:19:31 <systmkor> dalias: what I would want ignoring systemd
2015-04-02 23:19:56 <_mjones> systmkor: I write or rewrite the occssional init script, but it's rather rare. And it's usually required when the original author didn't implement all of the features, or did so improperly.
2015-04-02 23:20:26 <systmkor> _mjones: so that would be the same with a declaritive situation
2015-04-02 23:20:29 <_mjones> systmkor: if you don't want the possibility of writing bad code, maybe one should be using macos classic or something.
2015-04-02 23:20:32 <systmkor> minus a noose to hang yourself
2015-04-02 23:20:55 <_mjones> This strikes me as some weird salesmanship.
2015-04-02 23:20:57 <systmkor> _mjones: I want the removal of overly complex languages were not needed
2015-04-02 23:21:23 <_mjones> I'm familiar with the sales technique of taking the worst thing about your product and trying to make it a feature, but still.
2015-04-02 23:21:25 <_mjones> We need /bin/sh anyway.
2015-04-02 23:21:26 <systmkor> init sequence for sysadmin I don't think needs a turing complete language
2015-04-02 23:21:46 <systmkor> not really
2015-04-02 23:21:48 <_mjones> We don't need ruby interpreter and a ruby dsl anyway.
2015-04-02 23:22:03 <systmkor> a base shell isn't needed for an OS
2015-04-02 23:22:10 <systmkor> how Linux is designed it's pretty much needed
2015-04-02 23:22:14 <_mjones> Telling me I need fewer features for safety isn't a good idea.
2015-04-02 23:22:32 <systmkor> that's not really what I'm saying
2015-04-02 23:23:04 <systmkor> dalias: so basically a declarative system personally with interchangable parts that communicate in a consistent way
2015-04-02 23:24:30 <systmkor> _mjones: more complex automata to solve the same problem
2015-04-02 23:24:37 <dalias> systmkor, *nod*
2015-04-02 23:24:42 <systmkor> isn't more end features
2015-04-02 23:24:50 <systmkor> it's just more complexity for the same end goal
2015-04-02 23:24:52 <systmkor> _mjones: https://www.youtube.com/watch?v=3kEfedtQVOY
2015-04-02 23:24:59 <systmkor> will explain were I'm coming from
2015-04-02 23:25:14 <systmkor> dalias: so for example if I could wave a magic wand and change systemd
2015-04-02 23:25:17 <dalias> if none of the existing process supervision systems (runit, s6, etc.) meet your needs there (they're not as declarative as i would like either)
2015-04-02 23:25:44 <systmkor> dalias: well those are a peice of the pie meaning they handle the process supervision
2015-04-02 23:25:59 <dalias> the problem could still be solved with just a new supervision system, without demanding that it be put in init and integrated with logging, device hotplug, and all sorts of other unrelated things
2015-04-02 23:26:14 <systmkor> _mjones: if you are in more of reading mood check out http://www.cs.dartmouth.edu/~sergey/langsec/
2015-04-02 23:26:23 <_mjones> get to the part where systemd needs to liste on on :631 (cups) for a good reason.
2015-04-02 23:26:26 <dalias> however i think if you want declarative it's trivially handled by making your s6 or runit run commands "make" invocations :-)
2015-04-02 23:27:20 <_mjones> heh, that's clever.
2015-04-02 23:27:39 <dalias> make is the canonical universally-available declarative language :)
2015-04-02 23:28:37 <dalias> in any case, even if you do want to implement a new declarative supervision system
2015-04-02 23:28:42 <dalias> there's no reason it has to be pid 1/init
2015-04-02 23:28:59 <systmkor> yah again design issues of systemd not saying that was a good choice
2015-04-02 23:29:05 <dalias> there's no reason it has to have a public dbus-based interface surface the your desktop environment interfaces with and depends on
2015-04-02 23:29:08 <dalias> etc.
2015-04-02 23:29:26 <systmkor> okay so ignore how systemd is implemented and that it has any code for a moment
2015-04-02 23:29:43 <dalias> so far all of this is design/public-interface criticisms
2015-04-02 23:29:47 <dalias> not implementation criticisms
2015-04-02 23:29:47 <systmkor> if all of the existing init systems, supervisions systems, logging, hot plug etc.
2015-04-02 23:29:51 <systmkor> had a common user interface
2015-04-02 23:29:57 <_mjones> systmkor: I don't see a lot of sysV init-related CVEs, and if I did I wouldn't blame a turing-complete language for them.
2015-04-02 23:29:59 <systmkor> systemctl bootctl journaltcl and such
2015-04-02 23:30:35 <systmkor> were you could put in your choice of init system, superivisor system, syslog daemon etc.
2015-04-02 23:30:48 <dalias> no that's not a choice
2015-04-02 23:30:52 <systmkor> but had a common API in a sense and a common declartive language
2015-04-02 23:30:59 <dalias> and the problem is that you created unneeded and useless interface surfaces
2015-04-02 23:31:06 <systmkor> yes I know that's why I said ignore it's current design and buiild etc.
2015-04-02 23:31:09 <dalias> there's no need for systemctl or bootctl to exist AT AL
2015-04-02 23:31:17 <dalias> these are not interfaces applications should ever be aware of
2015-04-02 23:31:19 <systmkor> i would argue there is
2015-04-02 23:31:22 <systmkor> you already have them
2015-04-02 23:31:25 <systmkor> rc-service
2015-04-02 23:31:27 <systmkor> service
2015-04-02 23:31:34 <dalias> it's not an interface surface apps can use
2015-04-02 23:31:43 <systmkor> '/etc/init.d/<service>'
2015-04-02 23:31:45 <dalias> it's an administrative detail that you only care about as the administrator who chose the system to use
2015-04-02 23:31:50 <dalias> that's not a public interface
2015-04-02 23:31:53 <dalias> my systems don't have that
2015-04-02 23:32:10 <dalias> it's an internal detail of sysvinit
2015-04-02 23:32:12 <systmkor> *sigh* agreed
2015-04-02 23:32:16 <_mjones> technically it's now "service foo {start|stop|status}
2015-04-02 23:32:18 <systmkor> okay so again pause for a second
2015-04-02 23:32:20 <dalias> and there's no reason to try to make it public
2015-04-02 23:32:37 <systmkor> so I won't use the term systemd
2015-04-02 23:32:40 <dalias> doing so _constrains_ your choices and achieves nothing useful
2015-04-02 23:32:44 <systmkor> sys admins do need
2015-04-02 23:32:54 <systmkor> a common API, and user interface to adminster there systems
2015-04-02 23:32:56 <dalias> sys admins can use the system they chose to use
2015-04-02 23:33:05 <dalias> there is no need for it to have a "common api"
2015-04-02 23:33:19 <systmkor> dalias: i would disagree at that because
2015-04-02 23:33:29 <systmkor> that makes it less exchangable between iplementations
2015-04-02 23:33:32 <_mjones> it's useful to have a common api. But that API is sysV init, or 'service'.
2015-04-02 23:33:35 <systmkor> because what you are  doing with an API
2015-04-02 23:33:37 <dalias> your perspective is wrong
2015-04-02 23:33:43 <dalias> these are not simply implementations of a common api
2015-04-02 23:33:50 <systmkor> is making your mechanics orthogonal to your dynamics
2015-04-02 23:33:52 <systmkor> that is to say
2015-04-02 23:34:00 <dalias> they're different system designs that are invisible and irrelevant to anything high-level going on on the syste,
2015-04-02 23:34:04 <systmkor> the way you describe something is done seperate from how it's actually being done
2015-04-02 23:34:10 <dalias> which only matter from a standpoint of how you want to do lowlevel sysadmin stuff
2015-04-02 23:34:26 <dalias> and then you choose the system that fits with your preferences for how to administer the system
2015-04-02 23:34:48 <systmkor> yes so for specific systems this is and will always be the case
2015-04-02 23:34:52 <systmkor> with the nature of an open source OS
2015-04-02 23:35:00 <dalias> there's no reason they all need to look the same
2015-04-02 23:35:05 <systmkor> dalias: however my perspective is neither wrong nor irrelevant
2015-04-02 23:35:15 <systmkor> it's clear despite systemd's strong arming that
2015-04-02 23:35:21 <systmkor> a good chunk of sys admins
2015-04-02 23:35:26 <systmkor> do want a common UI and API
2015-04-02 23:35:31 <systmkor> for managing their system on linux
2015-04-02 23:35:32 <dalias> if you're designing some control panel for untrained admins to pretend they can admin unix systems, you can write thin wrappers to make them all look similar to the control panel
2015-04-02 23:35:40 <_mjones> Your packaging system needs to know your init, unless you design it otherwise. 'Low level' and 'high level' can be irrelevant when someone asks you if the software is compatible.
2015-04-02 23:35:42 <systmkor> not necessarily a common implementation
2015-04-02 23:35:48 <systmkor> a common UI and API
2015-04-02 23:36:05 <dalias> common UI and API is unreasonable
2015-04-02 23:36:11 <systmkor> no it isn't
2015-04-02 23:36:12 <dalias> this is a goal i'm absolutely against
2015-04-02 23:36:14 <systmkor> that's like saying
2015-04-02 23:36:19 <systmkor> a common linux kernel
2015-04-02 23:36:22 <systmkor> system functions
2015-04-02 23:36:24 <dalias> no it's no
2015-04-02 23:36:24 <systmkor> is unreasonable
2015-04-02 23:36:29 <dalias> it's a completely different domain
2015-04-02 23:36:55 <systmkor> which has isomorphic problems
2015-04-02 23:37:04 <dalias> you need common APIs for things that ordinary applications need to do in a portable way
2015-04-02 23:37:08 <systmkor> you have something that the mechanics and dynamics of the system
2015-04-02 23:37:18 <systmkor> one huge perk and standpoint of engineering
2015-04-02 23:37:28 <systmkor> is abstraction
2015-04-02 23:37:35 <systmkor> seperating the dynamics and mechanics of a system
2015-04-02 23:37:39 <stateless> systmkor: wtf are you talking about?
2015-04-02 23:38:01 <systmkor> stateless: dynamics and mechanics as in reference to physics
2015-04-02 23:38:17 <systmkor> meaning dynamics describes the system on how an netwonian object moves
2015-04-02 23:38:24 <stateless> systmkor: are you high?
2015-04-02 23:38:24 <systmkor> mechanics is why it moves that way
2015-04-02 23:38:34 <systmkor> stateless: no
2015-04-02 23:38:50 <systmkor> dalias: stateless so let's take another common engineering problem way simpler
2015-04-02 23:38:55 <systmkor> driving a car
2015-04-02 23:39:26 <systmkor> do you think that having a accelerator, break pedals, and a stearing wheel
2015-04-02 23:39:28 <systmkor> is useless?
2015-04-02 23:39:37 <dalias> not the car analogy...
2015-04-02 23:39:44 <systmkor> dalias: fine hot water tap
2015-04-02 23:39:48 <systmkor> wtf you want your call
2015-04-02 23:39:49 <dalias> it's almost always unjustified and bogus
2015-04-02 23:39:56 <dalias> analogies are not a valid form of argument
2015-04-02 23:39:56 <systmkor> no it isn't
2015-04-02 23:40:07 <systmkor> I'm not using it as an argument
2015-04-02 23:40:08 <dalias> they're a way of helping people explain something that's otherwise confusing to them
2015-04-02 23:40:11 <dalias> thats about it
2015-04-02 23:40:12 <dalias> then don't
2015-04-02 23:40:17 <systmkor> I'm using it to show that they have the same sets of problems
2015-04-02 23:40:27 <dalias> "to show..." = "as an argument"
2015-04-02 23:40:34 <dalias> so don't claim you're not
2015-04-02 23:40:41 <dalias> and analogies are not valid arguments
2015-04-02 23:40:51 <systmkor> well depends but sure w/e
2015-04-02 23:40:55 <systmkor> if your claim is that
2015-04-02 23:40:59 <systmkor> OSes by definition
2015-04-02 23:41:08 <systmkor> can't have a common UI/API of admistration
2015-04-02 23:41:11 <systmkor> from how they function
2015-04-02 23:41:22 <dalias> my claim is that it's harmful to impose commonality where you don't get a large benefit for it
2015-04-02 23:41:25 <systmkor> you are categorically wrong
2015-04-02 23:41:29 <dalias> that's called monoculture
2015-04-02 23:41:34 <systmkor> dalias: I didn't say impose
2015-04-02 23:41:43 <_mjones> monoculture considered harmful.
2015-04-02 23:42:04 <_mjones> That's an argument that the distributions shouldn't have adopted systemd so uniformly.
2015-04-02 23:42:11 <systmkor> dalias: did I say there should only be one UI and API ever
2015-04-02 23:42:14 <systmkor> no I didn't
2015-04-02 23:42:30 <systmkor> I said that admins do want want a (i.e. at least one) common UI/API
2015-04-02 23:42:34 <dalias> for a language/api that's meant for writing applications, it needs to be well-defined and different implementations need to exhibit the same outward characteristics where the behaviors are defined
2015-04-02 23:42:51 <dalias> but there is no way to treat every possible interface surface like this
2015-04-02 23:42:59 <dalias> if you do, you end up imposing exactly one system
2015-04-02 23:43:06 <dalias> and making diversity of implementations impossible
2015-04-02 23:43:23 <dalias> there is utterly no need for a standard API for system administration
2015-04-02 23:43:27 <systmkor> having a common API doesn't imply only one implementation
2015-04-02 23:43:36 <dalias> and most people actually running sites don't even use/want the ones we have
2015-04-02 23:43:55 <dalias> they write flat scripts to start (and possibly run supervision to restart) the services they need in the required order
2015-04-02 23:44:13 <systmkor> okay since I currently feel like you are ignoring what I'm saying, I'll take a moment to gather my thoughts, you explain what should be the means for system admin
2015-04-02 23:44:18 <systmkor> and just saying we need diversity
2015-04-02 23:44:22 <systmkor> isn't really an answer
2015-04-02 23:44:50 <dalias> i'm not going to keep wasting time on this discussion
2015-04-02 23:44:52 <_mjones> it might be when you're arguing we need the 15th standard.
2015-04-02 23:44:56 <dalias> you've drunk the systemd koolaid
2015-04-02 23:45:03 <dalias> and i have more useful things to do
2015-04-02 23:45:20 <dalias> sorry
2015-04-02 23:45:23 <systmkor> ...
2015-04-02 23:45:41 <systmkor> on every point you said systemd is failing I agree
2015-04-02 23:45:54 <systmkor> however every time i bring up a generic hypothetical solution
2015-04-02 23:45:59 <systmkor> you then conflate that with systemd
2015-04-02 23:46:13 <systmkor> and get mad at me for saying systemd is awesome which isn't what I said
2015-04-02 23:46:41 <systmkor> if you think that administring an OS should never have any common API
2015-04-02 23:46:49 <systmkor> that doesn't mean one and only one
2015-04-02 23:46:56 <systmkor> that is a different claim
2015-04-02 23:47:24 <systmkor> then you are also saying that every OS should be a unique snowflake with a variety of bootstrapping and interfacing to
2015-04-02 23:47:28 <systmkor> in the end get the same effect
2015-04-02 23:47:33 <dalias> just one final remark. if the goal was really to have an API for programmatically doing sysadmin stuff, then you could just write a tool to do that which knows the way to achieve admin tasks on all 14 existing systems, rather than making a 15th system and providing the programmatic API only for the 15th system and forcing everybody to adopt the 15th system
2015-04-02 23:48:01 <systmkor> dalias: when did I say systemd should be imposed
2015-04-02 23:48:18 <systmkor> if I did, I didn't mean it but I don't recall saying that
2015-04-02 23:48:42 <dalias> you didn't, but that's the whole approach of systemd
2015-04-02 23:48:51 <dalias> without forced adoption systemd is just the 15th standard
2015-04-02 23:49:11 <dalias> the adoption of the APIs is used as a tool for forced adoption of the implementation
2015-04-02 23:49:18 <systmkor> okay fine here then can we talk about things that do similar things like systemd without talking about systemd
2015-04-02 23:49:28 <systmkor> since that clearly is a trigger issue for you were I don't think
2015-04-02 23:49:35 <systmkor> I can segment what I'm trying to say or example
2015-04-02 23:49:41 <systmkor> vs. what the actual developers
2015-04-02 23:49:42 <dalias> *sigh*
2015-04-02 23:49:45 <systmkor> of systemd are doing
2015-04-02 23:50:51 <systmkor> so ignoring systemd
2015-04-02 23:51:06 <systmkor> what is fundamentally about providing a common UI/API for a OS
2015-04-02 23:51:21 <systmkor> wrong
2015-04-02 23:51:33 <stateless> systmkor: just drop this discussion
2015-04-02 23:51:54 <dalias> systmkor, it precludes OS's that are radically different
2015-04-02 23:52:26 <dalias> for instance if your API has mounting/unmounting filesystems, it imposes an assumption that filesystems are mounted and that there's a global mount namespace
2015-04-02 23:52:27 <systmkor> dalias: while that seems somewhat true I don't think that's entirely true
2015-04-02 23:52:37 <dalias> which are stupid legacy designs that we could otherwise do away with
2015-04-02 23:52:52 <systmkor> APIs aren't necessarily static
2015-04-02 23:53:06 <dalias> if they're not stable then they're worse than useless
2015-04-02 23:53:22 <systmkor> so your saying that all API for an OS are useless
2015-04-02 23:53:26 <dalias> whoever controls the API keeps changing it, rendering independent implementations impossible
2015-04-02 23:53:30 <dalias> no
2015-04-02 23:53:30 <systmkor> or thtat they needlessly impose
2015-04-02 23:53:52 <systmkor> oh, okay then what are you saying more specifically
2015-04-02 23:53:58 <dalias> i'm saying that it's a _hard_ problem choosing what aspects of the OS API are things that should look identical regardless of what system you're on, and which should be system-specific
2015-04-02 23:54:14 <systmkor> agreed that it is a hard problem
2015-04-02 23:54:14 <dalias> and there are smart people who have been thinking about these problems for decades and mostly gotten it right
2015-04-02 23:54:39 <dalias> and idiots like poettering show up with no understanding of that distinction and want to throw it all away :)
2015-04-02 23:55:12 <systmkor> dalias: with the API for an OS I think the whole point of the trade off is with abstraction of the system
2015-04-02 23:55:16 <systmkor> with loss of some control
2015-04-02 23:55:21 <systmkor> automation is gained
2015-04-02 23:56:07 <systmkor> so for example if a sys-admin has a means of controlling 95% of file systems
2015-04-02 23:56:09 <dalias> i think you underestimate the extent to which this kind of automation is _useless_ for real sites
2015-04-02 23:56:10 <_mjones> ubuntu already had upstart and they're surprisingly quick to give that up.
2015-04-02 23:56:14 <systmkor> which usually 100% of the ones they use
2015-04-02 23:56:17 <systmkor> why shouldn't they use that
2015-04-02 23:56:21 <dalias> nobody says "i want apache running"
2015-04-02 23:57:12 <systmkor> dalias: expand on how it's useless for real sites
2015-04-02 23:57:14 <dalias> instead they're starting it as its own user, with its own settings/paths controlled by the proper user, on a high port, with some sort of iptables rules or outside-world-facing proxy making it available on port 80/443
2015-04-02 23:57:34 <dalias> and this is not something that belongs in the root-owned system config
2015-04-02 23:57:43 <dalias> it belongs local to the user account it goes with
2015-04-02 23:57:58 <systmkor> okay one sec I'm half following the example
2015-04-02 23:58:14 <_mjones> this 28c3 presentation is about the halting problem and languages. I'm not particularly sympathetic to lang fixers, but regardless, init isn't the first (or 99th) place I would fix this.
2015-04-02 23:58:28 <dalias> so just having a generic "apache service" defined for some global service admin system is not useful
2015-04-02 23:58:53 <systmkor> dalias: pause one sec
2015-04-02 23:59:48 <systmkor> what I meant by a generic "apache service" is not Apache service equals an immutable setting of what port it is running on by what user
2015-04-03 00:00:03 <systmkor> *generic description is not
2015-04-03 00:01:07 <systmkor> _mjones: not sure what your sympathy towards 'lang fixers' has to do with having a problem solved by two languages 1 which is turing complete and potentially another that isn't which by defintion reduces your attack surface
2015-04-03 00:01:13 <dalias> my point is that the config is in the wrong place when you do it the API way you want
2015-04-03 00:01:52 <systmkor> dalias: how so?
2015-04-03 00:02:05 <_mjones> systmkor: attack surface between init and services isn't in my top 100 concerns about input validation and robustness.
2015-04-03 00:02:27 <_mjones> I think dalias is overselling the 'nobody uses a default services' argument, also.
2015-04-03 00:02:58 <systmkor> dalias: meaning what do you mean by config is in the wrong place
2015-04-03 00:03:36 <systmkor> _mjones: not certain about what your top 100 concerns has to do with the argument
2015-04-03 00:04:03 <dalias> _mjones, "nobody" is a stretch, but i find it's often true in practice
2015-04-03 00:05:32 <_mjones> systmkor: your argument is that having a non-turing DSL for init to interface with services decreases opportunity for error, and I'm saying that I'm vastly more concerned with input validation and the rest in every other situation. This isn't an attack surface or buggy interface about which I'm concerned.
2015-04-03 00:06:17 <dalias> the real problem with legacy init systems is racy use of pid files
2015-04-03 00:06:33 <dalias> and systemd doesn't solve it
2015-04-03 00:06:41 <systmkor> dalias: again what do you mean by wrong place of config file?
2015-04-03 00:06:54 <_mjones> dalias: I do believe everything should have sensible defaults. Not just to save work, but to encourage best practises and standardisation when non-defaults aren't necessary.
2015-04-03 00:08:19 <systmkor> to add, defaults I thought are not there for the person who is interested in customization but the person who isn't interested or even knows about customization
2015-04-03 00:11:12 <systmkor> _mjones: I agree that it isn't on the top 100 day to day concerns of a sysadmin but if someone handed you two options and with them you got equivelent end results but one innately had a reduced attack surface why wouldn't you choose the one with less attack surface
2015-04-03 00:11:45 <dalias> this isn't an attack surface
2015-04-03 00:11:56 <dalias> but your hypothetical question is unrelated to systemd
2015-04-03 00:12:08 <dalias> systemd has a massively expanded attack surface compared to all other init systems
2015-04-03 00:12:19 <dalias> as a result of exposing a dbus interface
2015-04-03 00:12:20 <systmkor> dalias: i didn't mean in relation to systemd
2015-04-03 00:12:33 <systmkor> I was talking about a comparison of a shell script and a non-turing complete language
2015-04-03 00:12:36 <_mjones> because one of them is by Lennart Poettering? x_x
2015-04-03 00:12:37 <systmkor> that's it
2015-04-03 00:12:53 <dalias> systmkor, as long as neither has external inputs, the attack surface of both is ZERO
2015-04-03 00:13:18 <dalias> zero surface => zero attack surface
2015-04-03 00:13:19 <_mjones> let's say he means robustness surface and not just attack.
2015-04-03 00:13:36 <systmkor> _mjones: yes slip of the tongue that's what I mean
2015-04-03 00:13:59 <_mjones> it would be easy to make init call a hardcoded interpreter, or make it selectable
2015-04-03 00:14:05 <systmkor> dalias: also if you have systems were you have active users on that system such as Universities then it is apart of your surface of attack
2015-04-03 00:14:25 <_mjones> service -> nginx, interpreter tcl, execpath /usr/local/etc/...
2015-04-03 00:14:49 <_mjones> only if your scripts are writable, really.
2015-04-03 00:15:09 <systmkor> _mjones: or that the scripts don't read
2015-04-03 00:15:12 <systmkor> anything from the system
2015-04-03 00:15:23 <systmkor> so for example if your init script is not writable
2015-04-03 00:15:35 <systmkor> but it reads from /var/lib/loldongs.conf
2015-04-03 00:15:44 <systmkor> and that file is writable
2015-04-03 00:16:07 <dalias> if it reads from paths writable by any user other than root (or the owner of the service if you have separate owners) then that's a bug in itself
2015-04-03 00:16:07 <systmkor> then yes it becomes back into scope as an attack surface
2015-04-03 00:16:13 <dalias> and the same issue would apply to declarative services
2015-04-03 00:16:27 <dalias> this does not happen though in any existing system i'm aware of
2015-04-03 00:16:33 <systmkor> dalias: yes however with the declartive service
2015-04-03 00:16:41 <systmkor> you wouldn't have a turing complete language
2015-04-03 00:16:56 <systmkor> depending on what that language did and how it called things
2015-04-03 00:17:03 <systmkor> I'm meaning within respect of that language initself
2015-04-03 00:17:39 <_mjones> at the cost of flexibility. But clearly this isn't the main purpose of systemd.
2015-04-03 00:17:49 <systmkor> ...
2015-04-03 00:17:50 <_mjones> upstart and launchd did it already.
2015-04-03 00:17:59 <dalias> it doesn't really make any difference
2015-04-03 00:18:05 <systmkor> I wasn't talking about systemd's declartive langauge
2015-04-03 00:18:14 <dalias> the program interpreting the declarative language is written in a turing-complete language
2015-04-03 00:18:21 <systmkor> I was talking about a theoretical declartive language as apposed to an init script
2015-04-03 00:18:33 <dalias> and even if the language is declarative, it's describing how to invoke programs written in imperative languages
2015-04-03 00:18:48 <dalias> declarative language is NICE but it's not a security measure
2015-04-03 00:19:11 <systmkor> it is a security measure but it doesn't magically solve everything
2015-04-03 00:19:13 <_mjones> most declaratives have callouts for procedurals.
2015-04-03 00:20:06 <systmkor> okay so pause on the langsec because if you think you have an argument I'm not the best person to talk about more detailed answer about
2015-04-03 00:20:08 <systmkor> chat with
2015-04-03 00:20:46 <_mjones> np.
2015-04-03 00:20:59 <systmkor> @maradydd
2015-04-03 00:21:15 <_mjones> Ugh, iojs1.6.2 only failed 7 tests but iojs1.6.3 fails 21. sigh.
2015-04-03 00:21:35 <systmkor> and @sergeybratus
2015-04-03 00:21:37 <_mjones> I did not expect this.
2015-04-03 00:22:19 <systmkor> dalias: so wait were were you talking about API in relation to location of a service's configuration file?
2015-04-03 00:22:41 <dalias> this conversation is really a time waste...
2015-04-03 00:23:09 <systmkor> did I miss that explenation?
2015-04-03 00:24:52 <dalias> no it's just a really simple matter that's frustrating to waste time explaining in detail
2015-04-03 00:25:57 <systmkor> m'kay
2015-04-03 00:26:33 <systmkor> well didn't mean it to turn out as such
2015-04-03 00:26:50 <systmkor> if that's what you are experiencing
2015-04-03 00:27:20 <systmkor> for that I apologize
2015-04-03 00:28:30 <dalias> for me it's been a series of demands for explanations of one thing after another
2015-04-03 00:28:45 <systmkor> as in general or my communication with you?
2015-04-03 00:28:53 <dalias> this conversation
2015-04-03 00:28:55 <dalias> i've got a meeting in a few minutes tho
2015-04-03 00:29:00 <dalias> so we'll have to talk later
2015-04-03 00:29:13 <systmkor> oh wow, I'm really sorry about that. Wasn't trying to make demands.
2015-04-03 00:29:55 <systmkor> Thanks. I do appreciate the time.
2015-04-03 00:30:02 <systmkor> dalias: good luck with the meeting
2015-04-03 00:30:07 <dalias> thanks
2015-04-03 01:22:55 <_mjones> I'm watching Poettering's presentation on systemd at fosdem15. systemd is apparently now an entire userland; there might be an init system there too.
2015-04-03 01:23:43 <_mjones> dns resolvers, ntp, networking. the logging over http sounds neat, but should be a totally separate program.
2015-04-03 01:27:30 <dalias> :)
2015-04-03 01:27:53 <_mjones> he's getting to the part about implementing their own selinux now, except not lame.
2015-04-03 01:32:54 <BitL0G1c> systemd have forked the kernel
2015-04-03 01:33:32 <_mjones> kdbus?
2015-04-03 01:34:25 <BitL0G1c> oops - maybe it was an april fool from distrowatch
2015-04-03 01:36:13 <BitL0G1c> is there any reason no elasticsearch in alpine ?
2015-04-03 01:36:59 <_mjones> someone was working on it the other day.
2015-04-03 01:38:02 <_mjones> I suspect there was technical difficulty.
2015-04-03 01:39:27 <BitL0G1c> oh
2015-04-03 01:40:20 <BitL0G1c> i sent a patch in for nginx-naxsi 1.7.11
2015-04-03 01:47:46 <systmkor> _mjones: kdbus, because dbus is clean and secure so let's put it in the kernel :P lol *barf*
2015-04-03 01:48:24 <dalias> :)
2015-04-03 01:50:08 <systmkor> is there any good portable IPC libraries?
2015-04-03 01:55:42 <systmkor> i'm just going to take the silence as a no :P
2015-04-03 01:58:25 <systmkor> dalias: thoughts on a separate library or a set of libraries that provide extra features that glibc does but keeping it out of a libc as it should?
2015-04-03 03:31:43 <Mo0O> hi there, do you know how can I fix this `mount` error: mounting /dev/loop0 on /mnt/cdrom failed: Invalid argument
2015-04-03 03:31:46 <Mo0O> ?
2015-04-03 03:35:10 <Mo0O> every mount commande return this on alpine
2015-04-03 03:36:57 <dalias> odd
2015-04-03 03:37:05 <dalias> care to show the whole command?
2015-04-03 03:37:22 <dalias> doesn't happen for me
2015-04-03 03:39:25 <Mo0O> mount -o loop,exec /home/scl/Downloads/systemrescuecd-x86-4.5.2.iso /mnt/cdrom
2015-04-03 03:39:29 <Mo0O> dalias: ^
2015-04-03 03:39:57 <Mo0O> strange, not?
2015-04-03 03:40:08 <Mo0O> it's a pretty common command
2015-04-03 03:40:55 <Mo0O> no related erro in dmesg
2015-04-03 03:42:18 <Mo0O> error*
2015-04-03 03:43:21 <dalias> does it only happen for loop mounts?
2015-04-03 03:53:07 <Mo0O> dalias: trying `mount /home/scl/Downloads/systemrescuecd-x86-4.5.2.iso /mnt/cdrom` result the same error
2015-04-03 03:54:10 <Mo0O> also with `mount /dev/sdb1 /mnt/usb`
2015-04-03 03:57:34 <dalias> ok the last is not a loop mount
2015-04-03 03:57:42 <dalias> but the .iso one is
2015-04-03 03:57:49 <dalias> try strace mount ...
2015-04-03 03:57:57 <dalias> and see what syscall returns the error
2015-04-03 04:00:16 <Mo0O> dalias: http://sprunge.us/WeVC
2015-04-03 04:13:12 <dalias> is the filesystem on the cd ext3?
2015-04-03 04:14:03 <Mo0O> dalias: on the iso?
2015-04-03 04:14:07 <dalias> yeah
2015-04-03 04:14:28 <dalias> (that was a rhetorical question. of course the answer is no; it's iso9660 or udf)
2015-04-03 04:14:41 <Mo0O> ^^
2015-04-03 04:14:45 <dalias> so either add the right -t option for the type of filesystem it is
2015-04-03 04:15:03 <dalias> or modprobe the right fs driver first so it's loaded and gets auto-tried
2015-04-03 04:15:15 <dalias> either solution should fix your problem :)
2015-04-03 04:16:40 <Mo0O> great! thanks a lot dalias, you rox!
2015-04-03 04:16:57 <Mo0O> I haven't ever need this option in the past
2015-04-03 04:17:11 <dalias> perhaps other dists auto-load the fs modules by default
2015-04-03 04:17:29 <dalias> or gnu mount (vs busybox mount) may auto-try well-known ones even if they're not loaded already
2015-04-03 04:18:15 <systmkor> dalias: have you looked at toybox http://landley.net/toybox/
2015-04-03 04:18:43 <systmkor> I got it to compile however running it against its internal test suite it seemed to fail a decent amount of its tests
2015-04-03 04:25:01 <systmkor> at a glance it seemed like something that Alpine community might like
2015-04-03 04:29:07 <dalias> eventually toybox will probably be better than busybox in all ways
2015-04-03 04:29:14 <dalias> but it still has a ways to go
2015-04-03 04:29:28 <dalias> the ls is already a good bit better though
2015-04-03 04:29:50 <systmkor> dalias: okay cool :D I'll put on my queue of things to keep on
2015-04-03 04:29:57 <dalias> busybox ls wastes lots of space and miscalculates columns so that it wraps funny with some dir contents
2015-04-03 04:29:59 <systmkor> if I have tome would love to audit code
2015-04-03 04:30:14 <systmkor> *some time I would
2015-04-03 04:31:07 <systmkor> dalias: I remember looking at the gnu coreutils a while back and I forget which tool it was but it was something really basic and it had gotos everywhere
2015-04-03 04:31:30 <systmkor> dalias: my reaction http://images.cryhavok.org/d/21396-1/Nope+Button.gif
2015-04-03 04:32:01 <dalias> :)
2015-04-03 04:34:08 <systmkor> has anyone messed with aboriginal linux seems like a clean way to bootstrap Alpine to various systems, however if people know of a better method I'm all ears
2015-04-03 04:34:20 <systmkor> http://landley.net/aboriginal/
2015-04-03 04:46:07 <dalias> i suspect you could just as easily cross-compile the base to get alpine bootstrapped, but i'm not sure
2015-04-03 04:46:37 <dalias> it would be nice to get an answer on how easy that is from an alpine developer
2015-04-03 04:48:42 <dalias> zz
2015-04-03 04:49:02 <systmkor> well i remember seeing one Alpine wiki article of getting bootstrapped
2015-04-03 04:49:09 <systmkor> but it seemed like they used arch to build base then move over
2015-04-03 04:49:30 <systmkor> which if I have to do is fine
2015-04-03 04:49:59 <systmkor> just was curious  if it works since it would be bootstrapping from musl libc source rather than glibc to then musl which probably doesn't make much of a difference
2015-04-03 04:50:33 <systmkor> but this project seems to also have a variety of workins with qemu to build emulated native builds which seemed interesting as well
2015-04-03 09:55:17 <engblom> I just noticed there is no openshot package for Alpine Linux. What is the preferred video editor for Alpine Linux?
2015-04-03 10:12:07 <rpu3uO8PEVZZ> i don't think alpinelinux primary goal is to host video editors
2015-04-03 10:13:19 <engblom> :/
2015-04-03 10:16:38 <rpu3uO8PEVZZ> small embedded cramped devices are the primary goal as far as i understand, videoediting is on the opposite spectrum of hw i guess. macs?
2015-04-03 10:17:31 <rpu3uO8PEVZZ> that some of us abuse this for their desktops is a bit of a diversion from that
2015-04-03 10:20:17 <rpu3uO8PEVZZ> why "prefered", why not supported? alpine is not quite a mass-product like ubuntu, where people are told what to use.
2015-04-03 10:22:36 <engblom> It begins to look like I will have to give up on alpine-linux for my desktop. Too much important tools are missing: unison, thunderbird, blender, openshot...
2015-04-03 10:26:10 <rpu3uO8PEVZZ> alpine is not really meant as a consumer-grade user desktop
2015-04-03 10:27:52 <rpu3uO8PEVZZ> it is much more some raw clay out of which you can create beautiful things.
2015-04-03 10:30:50 <rpu3uO8PEVZZ> (which also implies i guess, that if you want, you can build a beautiful consumer-grade desktop, given enough determination)
2015-04-03 10:33:59 <clandmeter> rpu3uO8PEVZZ: i have to agree to your last senctence.
2015-04-03 10:34:23 <clandmeter> alpine is whatever you make it.
2015-04-03 10:35:40 <engblom> What I am searching for is a clean and light general purpose distro without systemd with ready made packages for easy upgrades.
2015-04-03 10:36:36 <clandmeter> the advantage of alpine for me, is its simplicity. Adding things is not that hard.
2015-04-03 10:37:18 <engblom> clandmeter: But it adds complexity for each upgrade as you have to track upstream yourself, compiling stuff yourself etc...
2015-04-03 10:37:34 <clandmeter> engblom: ready made package for desktop is not what our focus currently is.
2015-04-03 10:37:51 <clandmeter> atleast not for me.
2015-04-03 10:38:05 <clandmeter> but i see more and more users using it on his/hers desktop
2015-04-03 10:38:15 <clandmeter> so landscape is changing
2015-04-03 10:39:29 <engblom> I really liked everything I saw so far with the exception of some, for me, important packages missing. I definitely will look at alpine again later
2015-04-03 10:40:54 <clandmeter> alpine is not easy life, not untill we have more developers helping out.
2015-04-03 10:41:30 <agd> does alpine come with a DE by default?
2015-04-03 10:41:47 <engblom> agd: You will find xfce
2015-04-03 10:42:18 <coredumb> i find alpine pretty much easy life ^^
2015-04-03 10:42:22 <agd> engblom; no, i mean that if I boot it up...will I be greeted with a nice login screen on a DE?
2015-04-03 10:42:29 <agd> or...a terminal?
2015-04-03 10:42:44 <agd> coredumb; what do you use alpine for?
2015-04-03 10:42:46 <clandmeter> it would be nice if we had one or two dev's who would target desktop development.
2015-04-03 10:43:18 <clandmeter> agd: there is no desktop by default
2015-04-03 10:43:23 <clandmeter> just bb shell
2015-04-03 10:44:03 <coredumb> agd: servers and laptop
2015-04-03 10:44:54 <coredumb> and i'm happy login in a console at bootup ^^
2015-04-03 10:45:04 <agd> coredumb; you seem to be a very happy person
2015-04-03 10:45:21 <clandmeter> this is a happy channel
2015-04-03 10:45:25 <agd> more happier than i could have ever previously imagined
2015-04-03 10:45:27 <rpu3uO8PEVZZ> \o/
2015-04-03 10:45:27 <coredumb> yep it's a bank holiday here as well
2015-04-03 10:45:39 <coredumb> so double the happiness
2015-04-03 10:45:42 <coredumb> \o/
2015-04-03 10:45:54 <agd> shave your armpits!
2015-04-03 10:46:04 <engblom> For me the log on screen is the least important thing of a distro... That part you only see a few seconds
2015-04-03 10:46:28 <coredumb> engblom: why don't contribute the missing packages ?
2015-04-03 10:46:29 <agd> engblom; are you on alpine, at this very moment?
2015-04-03 10:46:36 <rpu3uO8PEVZZ> well, for the systemd folks it's very important, they even dedicated a daemon to it.
2015-04-03 10:46:48 <coredumb> uhuh
2015-04-03 10:47:01 <engblom> agd: I have an alpine install running on a kvm virtual machine for testing alpine out.
2015-04-03 10:47:14 <agd> engblom; can you give me a guest ssh login, for venturing around
2015-04-03 10:47:46 <agd> engblom; remember, this is a happy channel. upon rejecting my request, the happiness present in this channel will be greatly reduced
2015-04-03 10:47:54 <coredumb> agd: alpine takes roughly 3mn to install...
2015-04-03 10:48:01 <engblom> agd: I do not dare to do it... That means messing with my work place firewall. I have the vm running at my workplace
2015-04-03 10:48:04 <agd> engblom; and besides, do I look like a malicious cybercriminal?\
2015-04-03 10:48:24 <rpu3uO8PEVZZ> agd if you cant setup alpine on your vm, you'll have trouble using it later
2015-04-03 10:48:31 <rpu3uO8PEVZZ> it takes about 3 min as coredumb said
2015-04-03 10:48:56 <clandmeter> and make a video about it, so we can send happy visitors to it.
2015-04-03 10:49:10 <agd> rpu3uO8PEVZZ; what do you use alpine for?
2015-04-03 10:49:38 <engblom> agd: What linux do you use at this moment?
2015-04-03 10:49:48 <agd> engblom; make a guess
2015-04-03 10:50:13 <agd> arch
2015-04-03 10:50:47 <engblom> agd: Then you should not have difficulties with alpine.
2015-04-03 10:52:26 <engblom> agd: the alpine wiki is good for getting you started...
2015-04-03 10:52:53 <agd> can somebody run paxtest on alpine and share the results?
2015-04-03 10:52:57 <agd> i'd like to see how secure it is
2015-04-03 10:53:05 <agd> paxtest takes just four seconds to complete or so
2015-04-03 10:54:15 <coredumb> i'm not on grsec kernel so it wouldn't be that usefull
2015-04-03 11:55:33 <bOSKE> i installed alpine on virtualbox when i startx the screen on virtualbox freeze,i cant move mouse (i can outside of virtualbox) it wont work ctrl+alt+backspace i need to restart it hard over vb
2015-04-03 11:55:48 <bOSKE> when i try rc-service lxdm start the screen flesh
2015-04-03 11:55:54 <bOSKE> just black and white
2015-04-03 11:55:56 <bOSKE> flash*
2015-04-03 11:59:16 <rpu3uO8PEVZZ> have you used setup-xorg-base?
2015-04-03 11:59:24 <rpu3uO8PEVZZ> or something similar called?
2015-04-03 12:02:04 <engblom> bOSKE: Regardless if you install xfce or just a plain wm, this page will help you to get X running: http://wiki.alpinelinux.org/wiki/XFCE_Setup
2015-04-03 12:02:39 <engblom> bOSKE: I had no problem to run alpine with X inside of a KVM VM.
2015-04-03 12:06:09 <AmatCoder> bOSKE: lxdm was very buggy last time I tried (on real thing)...
2015-04-03 12:06:52 <AmatCoder> bOSKE: ...so to avoid it would be better ;)
2015-04-03 12:07:17 <bOSKE> i folowed that page and do everything it says i think the problem is in drivers which drivers for video should i use xf86-video-ati or xf86-video-vmware?
2015-04-03 12:07:27 <xoritor> has anyone managed to port i3?
2015-04-03 12:07:37 <xoritor> that is my wm of choice
2015-04-03 12:07:44 <xoritor> i see i3lock
2015-04-03 12:07:45 <rpu3uO8PEVZZ> i3 is in testing for quite some time
2015-04-03 12:07:45 <bOSKE> AmatCoder, i started it only once just to see if my mouse and keyboard are working i use startx every time..
2015-04-03 12:07:52 <xoritor> hmmm
2015-04-03 12:08:10 <rpu3uO8PEVZZ> also instead of *dms i do this in my .profile: [[ $(tty) = /dev/tty1 ]] && exec startx
2015-04-03 12:08:20 <xoritor> awesome is there... i could use that i guess ...
2015-04-03 12:08:47 <rpu3uO8PEVZZ> xoritor?
2015-04-03 12:09:01 <xoritor> rpu3uO8PEVZZ, ?
2015-04-03 12:09:20 <xoritor> awesome wm...   http://awesome.naquadah.org/
2015-04-03 12:09:29 <rpu3uO8PEVZZ> xoritor: 14:07:41     rpu3uO8PEVZZ | i3 is in testing for quite some time
2015-04-03 12:09:36 <xoritor> yea
2015-04-03 12:10:03 <xoritor> i said awesome is there so i could use that until i3 is done and "release" for us mere lusers
2015-04-03 12:10:14 <xoritor> :-D
2015-04-03 12:10:19 <AmatCoder>  bOSKE: 'apk add xf86-video-vmware'
2015-04-03 12:12:25 <bOSKE> AmatCoder, im using that driver right now.. then it is not that .. :/
2015-04-03 12:15:33 <xoritor> i guess i COULD put testing in there... but any issues with stability doing that?
2015-04-03 12:16:51 <AmatCoder> bOSKE: try xf86-video-vesa
2015-04-03 12:19:32 <bOSKE> hm i get kernel panic now :)
2015-04-03 12:19:50 <bOSKE> it didnt work vesa when i hard reset it kernel panic :)
2015-04-03 12:21:40 <bOSKE> i will delete this image and install new one :) then i will try again :)
2015-04-03 12:22:07 <AmatCoder> hmm I think that you must delete all xf86-video-* (except xf86-video-vesa)
2015-04-03 12:23:00 <AmatCoder> and you can see what happens in '/var/log/Xorg.0.log'
2015-04-03 12:23:10 <bOSKE> i will try now just to install vesa
2015-04-03 12:23:21 <rpu3uO8PEVZZ> xoritor i have no idea what you're talking about. i3 is in testing and compeletely usable, why do assume otherwise?
2015-04-03 12:23:51 <bOSKE> i looked there before kernel panic
2015-04-03 12:24:03 <bOSKE> and there are some errors about it cant load
2015-04-03 12:24:08 <xoritor> rpu3uO8PEVZZ, sorry... i am just leary of repos listed as "testing" as they tend to make things a bit unstable
2015-04-03 12:24:09 <bOSKE> drivers
2015-04-03 12:24:28 <xoritor> rpu3uO8PEVZZ, so i tend to not add them to my listing of available packages
2015-04-03 12:24:58 <rpu3uO8PEVZZ> ok. but i think you're misunderstanding things in this case.
2015-04-03 12:25:02 <xoritor> rpu3uO8PEVZZ, thanks for the heads  up on that though!
2015-04-03 12:25:13 <xoritor> rpu3uO8PEVZZ, probably...
2015-04-03 12:26:01 <xoritor> rpu3uO8PEVZZ, i sometimes misunderstand air so i forget to breath ;-)
2015-04-03 12:26:06 <rpu3uO8PEVZZ> :)
2015-04-03 12:26:49 <xoritor> not really...
2015-04-03 12:26:59 <xoritor> but i have really bad humor
2015-04-03 12:27:25 <rpu3uO8PEVZZ> just do what you remember and understand. ;)
2015-04-03 12:27:37 <xoritor> then i would do NOTHING
2015-04-03 12:27:50 <xoritor> LOL
2015-04-03 12:27:57 <xoritor> 
ACTION drools

2015-04-03 12:27:59 <rpu3uO8PEVZZ> maybe thats better than most people do :)
2015-04-03 12:28:08 <rpu3uO8PEVZZ> 
ACTION also likes drooling :)

2015-04-03 12:28:15 <xoritor> at least i understand i know nothing
2015-04-03 12:29:41 <xoritor> i may have to pull the plug soon and put alpine bare metal on my laptop
2015-04-03 12:29:51 <xoritor> im becomeing a super fan
2015-04-03 12:29:52 <xoritor> lol
2015-04-03 12:30:09 <xoritor> actually using it in lots of places
2015-04-03 12:31:27 <xoritor> i do struggle with the non FHS based directory structure though
2015-04-03 13:46:05 <xoritor> anyone porting hhvm?
2015-04-03 14:39:21 <BitL0G1c> was someone working on elasticsearch ?
2015-04-03 15:17:50 <Psi|4ward> Error relocating /usr/bin/php-fpm: __flt_rounds: symbol not found
2015-04-03 15:18:00 <Psi|4ward> anyone seen this or give me a hint ?
2015-04-03 15:19:28 <dalias> need to upgrade musl package
2015-04-03 15:19:43 <Psi|4ward> how do i do this?
2015-04-03 15:19:49 <dalias> apk add -u musl
2015-04-03 15:20:04 <dalias> and possibly "apk update" first
2015-04-03 15:20:11 <Psi|4ward> yeehaa
2015-04-03 15:20:11 <dalias> the cause is this:
2015-04-03 15:20:24 <dalias> the FLT_ROUNDS macro is supposed to reflect current rounding mode but traditionally it was just defined as 1 always
2015-04-03 15:20:40 <dalias> so we had to add a function __flt_rounds() for defining it that returns the current mode
2015-04-03 15:20:58 <dalias> and now newly linked programs using FLT_ROUNDS will depend on having that function
2015-04-03 15:21:18 <Psi|4ward> okay, didnt understand you but it works
2015-04-03 15:21:26 <dalias> it's surprising anyone was using FLT_ROUNDS considering it was not reliable in the past (or on glibc or most other systems)
2015-04-03 15:21:30 <dalias> ok no prob
2015-04-03 15:21:35 <dalias> glad you got it fixed :)
2015-04-03 15:21:55 <Psi|4ward> bash-4.3# /etc/init.d/php-fpm start
2015-04-03 15:21:57 <Psi|4ward>  * WARNING: php-fpm is already starting
2015-04-03 15:22:21 <Psi|4ward> but this dont work like described in the wiki
2015-04-03 15:22:26 <BitL0G1c> rc-service php-fpm restart
2015-04-03 15:22:50 <Psi|4ward> is there a autostart on boot for this?
2015-04-03 15:23:05 <dalias> i always wondered why alpine has /etc/init.d if that's not the right way to do things
2015-04-03 15:23:10 <BitL0G1c> rc-update add php-fpm default
2015-04-03 15:23:26 <Psi|4ward> rc-service also sais "already starting"
2015-04-03 15:23:50 <BitL0G1c> install htop & find the php process & kill it
2015-04-03 15:24:20 <BitL0G1c> install htop - F4 "php" - F9    9    <enter>
2015-04-03 15:24:43 <Psi|4ward> ps aux
2015-04-03 15:25:04 <Psi|4ward> but kill -9 <pid> dont kill it
2015-04-03 15:25:09 <Psi|4ward>   165 root       0:00 [php-fpm]
2015-04-03 15:25:55 <Psi|4ward> btw. im in a docker container
2015-04-03 15:27:27 <BitL0G1c> php-fpm works fine in LXC - I don't have much experience of docker
2015-04-03 15:28:55 <Psi|4ward> okay just to be sure, i dont have to use rc-service and theres no autostart on boot by default?
2015-04-03 15:29:05 <Psi|4ward> so ill write a little start.sh script to run php-fpm and nginx
2015-04-03 15:29:19 <BitL0G1c> to autostart:
2015-04-03 15:29:29 <BitL0G1c> rc-update add nginx default
2015-04-03 15:29:41 <BitL0G1c> rc-update add php-fpm default
2015-04-03 15:30:15 <Psi|4ward> yeah i know but its not done while installing the pkgs
2015-04-03 15:30:23 <BitL0G1c> correct
2015-04-03 15:30:48 <Psi|4ward> perfect
2015-04-03 15:31:36 <Psi|4ward> alpine is so damn small and cool for containers ;)
2015-04-03 15:33:15 <BitL0G1c> :) yes
2015-04-03 16:53:24 <Psi|4ward> i think ive found a bug in nginx package, where to report?
2015-04-03 16:57:45 <Psi|4ward> im nginx user but i cant cd into /var/lib/nginx (permission denied) nginx has 0777
2015-04-03 16:57:47 <ScrumpyJack> bugs.alpinelinux.org
2015-04-03 17:28:02 <BitL0G1c> Psi|4ward - it's not a bug - nginx user does not have root permissions
2015-04-03 17:28:17 <BitL0G1c> grep nginx /etc/passwd
2015-04-03 17:28:47 <Psi|4ward> of course and he needs no root perms
2015-04-03 17:29:22 <Psi|4ward> nginx runs under nginx and the /var/lib/nginx belongs to nginx and has 0777 but nginx cannot write here
2015-04-03 17:29:31 <BitL0G1c> no
2015-04-03 17:29:45 <BitL0G1c> /var/www/localhost/htdocs is the home directory for nginx
2015-04-03 17:29:55 <Psi|4ward> nginx's temp folder
2015-04-03 17:30:04 <Psi|4ward> client_body_temp_path etc
2015-04-03 17:30:11 <Psi|4ward> not the html stuff
2015-04-03 17:31:34 <BitL0G1c> /var/lib/nginx s/be owned by root
2015-04-03 17:32:06 <Psi|4ward> its owned by default by root
2015-04-03 17:32:38 <Psi|4ward> [crit] 18#0: *9 open() "/var/lib/nginx/tmp/client_body/0000000001" failed (13: Permission denied)
2015-04-03 17:32:45 <Psi|4ward> this happens for fileuploads
2015-04-03 17:32:48 <Psi|4ward> php-fpm BE
2015-04-03 17:33:19 <BitL0G1c> yes - because uploads s/be in /var/www/.....
2015-04-03 17:33:33 <BitL0G1c> not var/lib
2015-04-03 17:33:35 <Psi|4ward> thats something PHP decieds
2015-04-03 17:33:48 <Psi|4ward> nginx only buffers the stream into tmp files
2015-04-03 17:34:22 <BitL0G1c> you need to change php.ini
2015-04-03 17:36:16 <Psi|4ward> PHP works fine, its a Nginx problem
2015-04-03 17:36:24 <Psi|4ward> i solved it telling nginx to use
2015-04-03 17:36:27 <Psi|4ward>   client_body_temp_path /tmp/nginx/body 1 2;
2015-04-03 17:36:29 <Psi|4ward>   fastcgi_temp_path /tmp/nginx/fastcgi_temp 1 2;
2015-04-03 18:37:02 <Psi|4ward> how to populate /etc/ssl/certs ?
2015-04-03 18:38:33 <Psi|4ward> got it! ca-certificates
2015-04-03 18:39:10 <BitL0G1c> or install openssl
2015-04-03 18:39:55 <Psi|4ward> nope that dosnt load the certs
2015-04-03 18:40:21 <dalias> does ca-certificates get updated to remove the discredited CAs like CNNIC?
2015-04-03 18:40:38 <BitL0G1c> don't know
2015-04-03 18:40:46 <dalias> it really should
2015-04-03 18:40:49 <Psi|4ward> hmmm the ca-cert.pem is missing
2015-04-03 18:41:04 <BitL0G1c> musl64 [~/aports]$ ls /etc/ssl/certs|grep CNNIC
2015-04-03 18:41:04 <BitL0G1c> ca-cert-CNNIC_ROOT.pem
2015-04-03 18:41:45 <dalias> there really should be some common policy for this
2015-04-03 18:43:03 <dalias> like 1-3 strikes allowed for CAs that fall victim to attacks, but instant and permanent removal of any CA found to have purposefully violated trust (issuing MITM certs commercially or to governments, etc.)
2015-04-03 18:44:13 <BitL0G1c> https://okturtles.com/
2015-04-03 18:47:32 <_mjones> Policy is for each distribution to decide. Some distros could leave out certain state-controlled CAs or NICs from the strart; others might be excessively forgiving.
2015-04-03 18:47:39 <_mjones> Diversity of choice is strength.
2015-04-03 18:48:11 <dalias> right now the browser vendors lead the choices
2015-04-03 18:48:22 <dalias> but google seems to be the only one taking first moves
2015-04-03 18:48:26 <dalias> then firefox follows
2015-04-03 18:48:29 <_mjones> Yes.
2015-04-03 18:48:45 <dalias> in the past firefox has been very hesitant to remove CAs even ones that proved themselves malicious
2015-04-03 18:49:09 <_mjones> I follow Google's tls and pki work in some detail. They've been at the forefront since 2010 or so, as many people know.
2015-04-03 18:49:16 <dalias> i think it would make sense for them to at least agree on the policies that will lead to removal
2015-04-03 18:49:36 <_mjones> dalias: of course we all know it's political. Like peering policy, there's diplomacy involved.
2015-04-03 18:49:55 <dalias> even if they don't make the same judgements
2015-04-03 18:50:05 <_mjones> If you remove something and your users rebel, it's a problem. If you keep something and keep getting abused by CAs, it's a problem. You need solomonic compromise.
2015-04-03 18:50:18 <dalias> it would lead to more transparency and reduce perception of bias/politics
2015-04-03 18:50:31 <_mjones> dalias: all it takes is for a big player to make an announcement about the objective facts that will get a CA removed.
2015-04-03 18:50:45 <_mjones> the problem is that then you have to abide by it.
2015-04-03 18:50:47 <dalias> right
2015-04-03 18:50:51 <dalias> that's what i
2015-04-03 18:50:54 <dalias> that's what i'd like to see
2015-04-03 18:51:08 <_mjones> this might be the impetus.
2015-04-03 18:51:09 <dalias> or see a joint effort by multiple big players to make such an 'objective' policy
2015-04-03 18:51:28 <_mjones> Have to be careful with that, lest perception or accusation of cartelisation.
2015-04-03 18:51:30 <nsz> browsers should just implement the ssh semantics: first time the cert is saved (with whatever verification), then only that cert is accepted later
2015-04-03 18:52:01 <_mjones> Then people like PRC and Turkey might get to paint themselves as oppressed, politically deprecated orgs.
2015-04-03 18:52:09 <dalias> i don't think ssh policy is very useful
2015-04-03 18:52:33 <dalias> the ssh system would be utterly broken with large numbers of hosts you interact with and if it were a mainstream attack target
2015-04-03 18:52:41 <_mjones> dnssec isn't perfect (what's perfect?), but DANE has a lot of possibilities to disrupt CAs.
2015-04-03 18:52:54 <dalias> i would very much like to see DANE replace CAs
2015-04-03 18:52:58 <_mjones> Might be largely unnecessary now that the cost of regular certs is sloping to zero.
2015-04-03 18:53:13 <dalias> but there's so much stupid policital whining about US control of the root
2015-04-03 18:53:25 <_mjones> I hear some, not a lot.
2015-04-03 18:53:39 <_mjones> I think icann has shown itself to be a good steward over the decades.
2015-04-03 18:53:49 <_mjones> And even so, anyone can switch to altroot, any damn time they want.
2015-04-03 18:53:54 <_mjones> Any dist can choose altroot.
2015-04-03 18:54:10 <dalias> "one superpower can theoretically seize control in a hostile way that would be immediately visible to everyone" is a much less bad situation than "any state or private CA can screw you whenever they want"
2015-04-03 18:54:16 <_mjones> Been tried before, failed, because few actually want altroot.
2015-04-03 18:54:43 <_mjones> cert transparency will be a big help. DANE will take a long time to be truly viable.
2015-04-03 18:55:00 <dalias> yes :/
2015-04-03 18:55:21 <_mjones> certain strategies for multiple certs are useful, too.
2015-04-03 18:55:48 <_mjones> Find a trust-chain that the client likes, possibily for political reasons.
2015-04-03 18:56:05 <dalias> yes
2015-04-03 18:56:22 <_mjones> I use the ssl observatory as an early type of ca transparency vehicle.
2015-04-03 18:56:42 <dalias> i've often joked that PKI should be based on needing 2/3 valid signatures from eurasia, east asia, and oceania ;-)
2015-04-03 18:57:15 <_mjones> nsz: the biggest problem with trust-on-first-encounter is that you encounter so many new certs, and anyone inside a hostile network may never see legit certs.
2015-04-03 18:57:43 <_mjones> dalias: yes, a certain degree of optional web-of-trust like that would be most welcome.
2015-04-03 18:58:56 <_mjones> nsz: if the situation is bad now, think about teaching users that they can't _really_ trust certs they see for the first time on a new machine through free wireless, but _probably_ they can trust it if they've seen it in the last (HSTS, whatever) number of seconds after first seing it through a network they mostly-trust.
2015-04-03 18:59:37 <nsz> for me trust-first-time would be much better
2015-04-03 18:59:49 <_mjones> nsz: then, what does trust mean? Some of these people are never going to 'trust' comcast or whatever. And even though I think most of that is politically misguided, it's not a bad idea to have a system that doesn't require any such trust.
2015-04-03 18:59:53 <nsz> there are only a few sites where i care about security
2015-04-03 19:00:04 <_mjones> nsz: you can publish SSH public keys in DNS with SSHFP record.
2015-04-03 19:00:09 <nsz> and i could easily verify those certs via side channels
2015-04-03 19:00:29 <_mjones> nsz: the problem is that as a browser or dist vendor one has to make policy for all of one's users.
2015-04-03 19:02:43 <_mjones> nsz: I like to use https almost everywhere, and there's no practical way for me to validate identity on them manually.
2015-04-03 19:03:37 <_mjones> recent TLS and PKI issues have convinced some not to use the systems even though they're very mature, very well understood, and their weaknesses are very well known and have countermeasures or workarounds.
2015-04-03 19:04:06 <nsz> recent?
2015-04-03 19:04:34 <nsz> tls/ssl was never an acceptable option..
2015-04-03 19:04:52 <_mjones> from diginotar to date, including implementation issues from the last year in particular.
2015-04-03 19:05:01 <nsz> when i asked the tls public key fingerprint at my bank the clerk didnt know what i was talking about
2015-04-03 19:05:02 <_mjones> nsz: what do you mean by never an acceptable option?
2015-04-03 19:05:25 <nsz> (ie every client of the bank before me blindly trusts whatever their browser accepts)
2015-04-03 19:05:33 <nsz> (that's complete nonsense)
2015-04-03 19:05:45 <_mjones> nsz: he probably doesn't know about the FINCEN anti-money-laundering control procedures either. This means nothing.
2015-04-03 19:06:04 <dalias> nsz, i think you should be glad. the alternative would not be the bank teller knowing the fingerprint, but instead them giving you a dvd to install with a custom root CA on it... ;-)
2015-04-03 19:06:19 <_mjones> no kidding.
2015-04-03 19:06:20 <dalias> i'm told there are plenty of messed up banks that do stuff like that
2015-04-03 19:06:27 <dalias> the situation in the US is so much better
2015-04-03 19:06:52 <_mjones> Everybody is looking for the magic that lets semi-trained users benefit from all of the advances.
2015-04-03 19:07:33 <nsz> anyway i dont think https protects users who dont know how it works in any way
2015-04-03 19:07:35 <dalias> my impression is that our legal system and/or private insurance systems relieve both banks and their clients from liability for fraud to a sufficient extent that banks aren't doing all sorts of security snake-oil to cover their own asses
2015-04-03 19:08:33 <dalias> visa, mastercard, the fdic, and/or insurance companies end up covering the losses from various types of fraud
2015-04-03 19:08:57 <_mjones> nsz: I disagree. Scarcely any systems are perfect, and hypothesising systems that might be perfect is counterproductive in many ways. I became a fan of tls in 2012 after 15 years of avoiding the mess.
2015-04-03 19:09:12 <dalias> we don't even have embedded smartcards in debit cards, just magnetic strip
2015-04-03 19:09:17 <dalias> and 4-digit pin
2015-04-03 19:09:30 <_mjones> chip and pin has been coming to the states for years.
2015-04-03 19:09:53 <_mjones> Lots of american cards don't work in automated systems overseas. Specifically public-transit systems, I find.
2015-04-03 19:09:58 <dalias> nobody wants to run debit cards as debit/atm anyway
2015-04-03 19:10:11 <dalias> because you get better rewards and fraud protection by running it as visa
2015-04-03 19:10:32 <_mjones> yes. rewards and protection are separate issues, though.
2015-04-03 19:10:47 <nsz> the problem with tls is not that it's not perfect but it gives no protection while pretends that it does
2015-04-03 19:11:00 <dalias> i'm saying even if we got chips everyone would still want to run as 'credit' (without chip) instead of debit (with chip/pin)
2015-04-03 19:11:23 <dalias> nsz, it gives large degrees of protection even with the flaws
2015-04-03 19:11:24 <_mjones> nsz: I strongly disagree, and I worry that such opinions are having a big negative impact on adoption and security in general.
2015-04-03 19:11:43 <nsz> dalias: only if you know about its workings
2015-04-03 19:11:48 <nsz> otherwise it's useless
2015-04-03 19:11:52 <_mjones> TLS all the things. Once you do the work of making it work, it's 'free' for everything after.
2015-04-03 19:12:00 <dalias> if a site is using a cipher with forward secrecy, then it prevents all _retroactive_ invasion of privacy
2015-04-03 19:12:10 <_mjones> dalias: chip and pin applies to credit too.
2015-04-03 19:12:18 <nsz> (you get a popup window and click next twice and security is over)
2015-04-03 19:12:22 <dalias> _mjones, how does that work for online purchases?
2015-04-03 19:12:36 <dalias> most of my purchases are online
2015-04-03 19:12:44 <_mjones> dalias: falls back to old style, I believe.
2015-04-03 19:12:59 <_mjones> Merchants get different rates for 'card physically present' and 'card not present' transactions.
2015-04-03 19:13:11 <_mjones> They also get different rates based on their own size, risk, security level and history.
2015-04-03 19:13:30 <dalias> _mjones, yes. so chip is just used for validating 'card physically present' ?
2015-04-03 19:14:01 <dalias> doesn't sound very useful
2015-04-03 19:14:12 <_mjones> I'm not sure dalias. There are readers, like the one in my thinkpad into which I have a card plugged right now, but the main market I see for those is U.S. department of defense CAC id cards.
2015-04-03 19:14:36 <_mjones> Oh, they're used overseas too, for national ID cards.
2015-04-03 19:14:52 <_mjones> Like if you want to file your taxes in scandanavia or things like that.
2015-04-03 19:15:06 <nsz> (in cloud computing era, https makes even less sense since your communication is not protected end-to-end)
2015-04-03 19:15:18 <_mjones> nsz: it mostly is now.
2015-04-03 19:15:32 <nsz> how?
2015-04-03 19:15:52 <_mjones> Google and others used not to protect the backend connections for reasonable reasons, but Nation State Actor reasons caused them to TLS all of the backend in a hurry.
2015-04-03 19:16:13 <nsz> i'm talking about protection against google reading my mails
2015-04-03 19:16:17 <_mjones> plus google has a js library called 'end to end'. It may not be perfect, but it seems damn useful.
2015-04-03 19:16:41 <nsz> end-to-end is not me-google but me-someone
2015-04-03 19:16:48 <_mjones> nsz: that requires encryption in storage, which is a lot uglier right now than encryption in flight, IMO.
2015-04-03 19:16:53 <dalias> i think nsz means the data is decrypted on the 'cloud servers'
2015-04-03 19:17:08 <_mjones> HTML5 has zero-knowledge capabilities.
2015-04-03 19:17:32 <_mjones> Like the HTML5 based password lockers. It's decrypted on the front-end only; backend couldn't get your passwords back even if they wanted.
2015-04-03 19:18:09 <_mjones> Yes, the HTML5 could be compromised, but there are workaround there too (like local clients).
2015-04-03 19:18:46 <dalias> i suspect we're moving towards major cultural changes that will drop a large part of the 'expectation of privacy' that has always existed
2015-04-03 19:18:56 <dalias> and if that's universal it's not necessarily such a horrible thing
2015-04-03 19:19:16 <_mjones> sousveillance is not unlikely.
2015-04-03 19:19:34 <nsz> well in china (which is supposed to be an oppressive system) ppl dont encrypt/auth any of their communication
2015-04-03 19:19:36 <_mjones> We're already used to cameras and cheap digital storage.
2015-04-03 19:19:53 <nsz> (at least all chinese chat things i've looked at sent clear text)
2015-04-03 19:20:06 <dalias> i think there are really two big remaining issues even if that happens though
2015-04-03 19:20:14 <_mjones> nsz: they often use windows xp, many institutions required ie6 until recently, and it's unclear how the 'great firewall' treats https.
2015-04-03 19:21:26 <dalias> one is authentication & authorization -- even if what you're doing can't be private, there still needs to be some way verify the source of information and commands
2015-04-03 19:21:41 <_mjones> I said the other day that Poettering's syslog over http was interesting as a discrete standard/idea and appeals to my sensibilities. Instead of inventing new crypto, we can use tls over any https or tcp connection without further drama.
2015-04-03 19:22:08 <dalias> the other is whether lack of privacy is universal or only affects the non-elite class
2015-04-03 19:22:17 <_mjones> tls enables, for me, the zero-trust network.
2015-04-03 19:22:38 <_mjones> dalias: things aren't so simple or clear-cut as class distinctions.
2015-04-03 19:23:10 <_mjones> anyone involved in criminal or civil disputes has those subject to public knowledge, for instance.
2015-04-03 19:24:37 <nsz> i think the problem is not lack of privacy but asymmetric access to private information
2015-04-03 19:24:40 <_mjones> money and state-political power can buy privacy, but that doesn't mean there are class divides in privacy. cameras get everyone in view. malware doesn't discriminate.
2015-04-03 19:24:41 <dalias> if the elite can still make secret laws and trade agreements in private while the rest of the world is caught up in having their love affairs, nasty things they say about their coworkers and relatives, etc. put on display for everyone to see....
2015-04-03 19:24:48 <dalias> then the world is doomed :-p
2015-04-03 19:24:57 <nsz> i think if everything is public to everyone that should be ok
2015-04-03 19:25:12 <nsz> but if everything is public to a small group that's not ok
2015-04-03 19:25:19 <_mjones> secret laws and trade agreements is an issue of state authoritarianism, not privacy.
2015-04-03 19:25:35 <dalias> _mjones, well that was more of an example
2015-04-03 19:25:55 <dalias> the general phenomenon is "secret negotiations to seize control of economic and political power"
2015-04-03 19:26:02 <_mjones> I don't think the wealthy or connected are generally any more or less vulnerable, except that more people are interested in the nude selfies of attractive celebrities than of random others.
2015-04-03 19:26:42 <nsz> i dont think privacy issues has anything to do with nude selfies
2015-04-03 19:27:00 <_mjones> nsz: private information is more democratised than ever. Anyone can buy access to a police reacord from a website cheaper than it used to be to hire a private detective to pull it for you.
2015-04-03 19:27:26 <_mjones> nsz: it's not clear to me which type of privacy you mean.
2015-04-03 19:27:36 <dalias> nsz, it does as long as they affect your ability to get a job, be taken seriously in your field, be accepted by your family, etc...
2015-04-03 19:28:11 <_mjones> I don't think dalias is primarily concerned about anti-state agitation or whatever.
2015-04-03 19:28:12 <dalias> different people have different things they care about being private for different reasons
2015-04-03 19:28:29 <nsz> i think the problem is that of 1984: protecting private data against central authority
2015-04-03 19:28:39 <nsz> and i dont think any authority is interested in nude selfies..
2015-04-03 19:29:05 <dalias> e.g. i don't particularly care about my banking being private. it's all really boring and not something that could be used against me, and if somebody used data from it to steal my money, it's insured anyway
2015-04-03 19:29:29 <_mjones> nsz: central authority has always had powers of legal compulsion and warrants or writs on its side. There is more data to be had now, but one might say that applies equally for and against central authority.
2015-04-03 19:29:58 <nsz> central authority should obey laws
2015-04-03 19:30:10 <nsz> of course if the laws are broken you cannot do anything
2015-04-03 19:30:15 <nsz> (like in the us)
2015-04-03 19:30:26 <_mjones> dalias: whereas others are hugely concerned about banking privacy, often against state action. The U.S. has a habit of seizing money without arrest, and not giving it back later, for instance.
2015-04-03 19:30:50 <nsz> (which has absolutely no privacy protection to any non-us citizen)
2015-04-03 19:30:56 <dalias> nsz, actually abusive police in the US have used nude pics to blackmail people into not exercising their legal rights
2015-04-03 19:31:00 <_mjones> nsz: it's easy to misunderstand the U.S. because Americans are rather more prone to criticising their own state than are many others.
2015-04-03 19:31:05 <nsz> (eu is not that racist, here we think privay is human right)
2015-04-03 19:31:38 <nsz> _mjones: i'm talking about FISA
2015-04-03 19:31:40 <dalias> _mjones, right. i was making an example about how people's privacy needs/expecations differ
2015-04-03 19:31:41 <_mjones> Euro nations can be unreasonably proud of their own states compared to what they know of the U.S.
2015-04-03 19:32:39 <_mjones> fisa is obviously a big problem, and non-transparent. However, it's not remotely unique. How many secret authoritarian laws has your own state?
2015-04-03 19:33:02 <nsz> my state does not respect the law in the first place :)
2015-04-03 19:33:11 <_mjones> It might be that you know more about U.S. authoritarianism than you know about the authoritarianism of places closer to home.
2015-04-03 19:33:28 <_mjones> nsz: Hungary?
2015-04-03 19:34:07 <nsz> but if you come here and have some medical checkup and the records get used by the govmt without your consent then you can get protection from eu data protection laws and sue the govmt
2015-04-03 19:34:12 <nsz> _mjones: yes
2015-04-03 19:34:38 <nsz> if i go to usa and have the same issue i have no protection if the govt got to my data through fisa
2015-04-03 19:34:47 <_mjones> nsz: governments only allow themselves to be sued at their own discretion, and for their own reasons.
2015-04-03 19:37:33 <nsz> (btw hungary has very strict data protection laws.. because communism.. i think it's the only country without any central id to make it harder for the govmt to collect data about citizens in a central db)
2015-04-03 19:38:00 <nsz> s/central id/personal identification number/
2015-04-03 19:38:57 <_mjones> only communist country? The U.S. has no central id except passport. (Passport regime is incidentally a product of WWI. There was a free-trade and free-travel golden age in the decades before WWI.)
2015-04-03 19:39:33 <_mjones> Oh, the U.S. has numbers. FDR's regime resembled nothing so much as Mussolini's Italy.
2015-04-03 19:42:25 <nsz> i thought in usa ssn was tied to a person forever, but wikipedia says it's optional
2015-04-03 19:42:59 <_mjones> it's not really optional.
2015-04-03 19:44:40 <_mjones> it's theoretically optional in a number of ways, but everyone has one. Parents can't receive a tax credit if children don't have one, and I believe government schools used to require them before being admitted.
2015-04-03 19:45:01 <_mjones> There are ways to avoid using it, in many cases.
2015-04-03 19:46:54 <_mjones> Member of certain designated religious groups can legally avoid paying social security taxes, but almost nobody knows that.
2015-04-03 21:43:03 <ponyo> What's the smallest amount of ram alpine linux can run with on x86 to be the most basic of firewall/nat router for one client?
2015-04-03 21:43:57 <ponyo> I guess a better follow up question, is 64mb reasonable enough?  Were not even talking dhcp/dns or anything else here
2015-04-03 21:46:49 <_mjones> 64 is _probably_ enough.
2015-04-03 21:47:25 <_mjones> Base system takes up 24-31 on 64-bit archs. iptables with nat is a kernel service, so you shouldn't need much more ram.
2015-04-03 21:47:35 <_mjones> or nftables I guess.
2015-04-03 21:47:43 <ponyo> Ty :)
2015-04-03 21:48:41 <ponyo> My stupid ISP gave us 10 IP addresses, but won't support IP aliasing and it's been a real PITA to come up with a sensible workaround
2015-04-03 21:49:44 <jomat> what's IP aliasing?
2015-04-03 21:50:22 <ponyo> more than one IP address to a MAC address, usually in the same subnet :)
2015-04-03 21:50:36 <_mjones> yes, what's ip aliasing? 10 might be quite a lot of ipv4, all things considered.
2015-04-03 21:50:41 <ponyo> I should be able to assign all addresses to the same physical interface, but my ISP blocks it
2015-04-03 21:50:48 <_mjones> ah
2015-04-03 21:50:50 <ponyo> or at least more than 1
2015-04-03 21:50:56 <_mjones> that's policy. They want to prevent wasting IPs.
2015-04-03 21:51:04 <_mjones> I bet you're outside north america?
2015-04-03 21:51:06 <ponyo> Which they have at least confirmed, but can't seem to fix or explain why it's happening :(
2015-04-03 21:51:13 <ponyo> Nope, Seattle, Wa
2015-04-03 21:52:03 <_mjones> the don't support it as in it simply won't function?
2015-04-03 21:52:30 <_mjones> s/the/they/
2015-04-03 21:52:46 <ponyo> Anytime I give more than 1 ip to a network interface, both quit getting response from the upstream router, however, anything else on the same subnet has zero issue with the situation of ip aliasing
2015-04-03 21:53:16 <_mjones> It's just that any machine old enough to only have 64mb of dram I would worry about maxing out the cpu if you're not careful.
2015-04-03 21:53:22 <ponyo> They've confirmed it, but can't explain it, we have a fancy business account with our ISP and even their sales people say their engineers lack
2015-04-03 21:53:36 <ponyo> I'm putting 9 vms in a ESXI box
2015-04-03 21:53:45 <ponyo> and i use vlans to pipe everything around
2015-04-03 21:54:05 <ponyo> I'd rather not have 9x 256mb ram, but 9x64mb is much more reasonable
2015-04-03 21:54:26 <_mjones> You should only need one box to do this.
2015-04-03 21:54:34 <ponyo> I completely agree
2015-04-03 21:54:53 <ponyo> But routing tables get screwy with more than 1 interface in the same subnet
2015-04-03 21:54:57 <_mjones> I haven't used nftables yet because it's rather new, but I would look to do it there first -- one stop shopping.
2015-04-03 21:55:09 <ponyo> I've tried to do policy based routing and had zero luck
2015-04-03 21:55:27 <ponyo> I assume that's what nftables is, some new iptables + routing goodies?
2015-04-03 21:56:11 <_mjones> on the ISP side, how long does it take to stop responding? I'm trying to decide if they have a purposeful mechanism they aren't communicating properly, or some weird technical inability to support a basic config.
2015-04-03 21:56:42 <_mjones> nftables came out in Linux 3.13 and unifies iptables, ebtables, arptables and something else.
2015-04-03 21:56:48 <ponyo> Instantly, as soon as it sees the arp going to a different ip, it freaks out and has to reset
2015-04-03 21:57:41 <ponyo> Or at least that's what I seem to see.  I've had them confirm it for me that it is happening, but they can't explain why or tell me when / if they are going to fix it
2015-04-03 21:57:50 <_mjones> I had equipment with ASIC insufficiencies a long time ago, and I suppose it' possible they're doing some tag-in-tag vlan stuff that might expose such an insufficiency, and not really want to admit it.
2015-04-03 21:57:52 <ponyo> It's been kind of a joke trying to get a response out of them
2015-04-03 21:58:18 <ponyo> That's the first reasonable attempt at an explanation I've heard yet
2015-04-03 21:58:28 <ponyo> I figured they would use the Docsis 2.0/3.0 cop out
2015-04-03 21:58:51 <_mjones> At first I assumed they were doing it on purpose, but that seems less likely under the circumstances.
2015-04-03 21:58:56 <ponyo> As it is an old Arris modem, but in bridged mode, I don't see how it should even be an issue
2015-04-03 21:59:00 <_mjones> it coul dbe DOCSIS.
2015-04-03 21:59:38 <ponyo> It's Wave Broadband, the bastard off shoot of Comcast that claims to not really be Comcast so they say they don't have a monopoly in Washington
2015-04-03 21:59:42 <ponyo> Kind of sad actually
2015-04-03 21:59:49 <_mjones> I've done both l2tp to DSLAMs and PPPoE to (Redback) dslams in the past, and I bet DOCSIS has similar concerns.
2015-04-03 22:00:22 <ponyo> Hmmmm
2015-04-03 22:01:11 <_mjones> my old situation was where a switch would detect a MAC on a certain interface, then detect it on another interface and cut off the first one. Which is what you want....unless the two interfaces are in different VLANs.
2015-04-03 22:02:18 <_mjones> the ASIC on that (Cisco hardware, a Cisco acquisition from combinet or cascade or someone) didn't have tables for any situation except interface. It couldn't handle the condition with the same mac on two interfaces at the same time, even when on different VLANs.
2015-04-03 22:03:16 <ponyo> Right, and just for the record here, we are trying to do IP aliasing which is the inverse of that, many IP's on one interface, which is commonly supported it seems everywhere else
2015-04-03 22:03:24 <_mjones> actually, it very well could be a limitation on their provisioning system too.
2015-04-03 22:03:47 <ponyo> I've given up after pfsense, bsd, mikrotik, and alpine with policy based routing and decided to macro scale and be done with it
2015-04-03 22:03:56 <_mjones> Sure that was an old situation, but it was an example how something could be a hardware problem.
2015-04-03 22:04:00 <ponyo> Right
2015-04-03 22:04:02 <ponyo> :)
2015-04-03 22:04:33 <_mjones> sounds like you tried it all.
2015-04-03 22:04:54 <ponyo> I completely agree its something on their side and I had their engineer in my office agreeing with me.  Two days later he resigned, and now the sales person is being all flaky about responding to emails.
2015-04-03 22:05:06 <_mjones> lol.
2015-04-03 22:05:07 <ponyo> I want to resolve this
2015-04-03 22:05:16 <ponyo> It's not that hard of a problem
2015-04-03 22:05:19 <_mjones> At least someone came out.
2015-04-03 22:05:36 <_mjones> I used to be in the isp business, but that was a long time ago.
2015-04-03 22:06:28 <ponyo> Yea, I've been wanting to get into virtual servers and services providing, but I keep getting roped into office deployments
2015-04-03 22:06:44 <_mjones> consulting?
2015-04-03 22:06:44 <ponyo> and public web stack deployments
2015-04-03 22:06:47 <ponyo> Basically
2015-04-03 22:07:04 <ponyo> Consulting with stock options?
2015-04-03 22:07:15 <ponyo> I dunno if it still counts if i'm getting extra benefits too
2015-04-03 22:07:38 <_mjones> doing different things that people want is a lot harder work than doing fewer things and finding people who need that, I've found.
2015-04-03 22:07:48 <_mjones> ponyo: this for a s.o.?
2015-04-03 22:07:57 <ponyo> S.o?
2015-04-03 22:08:23 <ponyo> Significant other?
2015-04-03 22:08:37 <_mjones> I don't know which extra benefits.
2015-04-03 22:09:50 <ponyo> Oh yea stock options, basically I get paid what they can afford and the rest comes in options.  I don't know if that is considered consulting but i've been acting at the CTO for one company the head of System Administration at another.  I also do network engineering consulting for small to medium business deployments
2015-04-03 22:10:11 <ponyo> and Tutor :)
2015-04-03 22:11:54 <Mo0O> is it safe  to create symlink in aport?
2015-04-03 22:12:28 <_mjones> Mo0O: what's the context?
2015-04-03 22:12:46 <_mjones> ponyo: it's usually easier when you get to make the decisions, and not just implement someone else's decisions.
2015-04-03 22:13:02 <Mo0O> create a symlink to configuration files in /etc
2015-04-03 22:13:14 <ponyo> Yea, except trying to explain those decisions to gamers that think they are techs that run companies can be a challenge sometimes :)
2015-04-03 22:13:30 <_mjones> during the build or during the install?
2015-04-03 22:13:59 <Mo0O> during install
2015-04-03 22:14:07 <_mjones> ponyo: I'm always interested in examples, although I have plenty of my own.
2015-04-03 22:14:49 <_mjones> Mo0O: normally you want to use the install program; I think it will do symlinks. Otherwise, yes, the install() function should do symlinks ok.
2015-04-03 22:15:09 <Mo0O> ok, thx ;)
2015-04-03 22:15:11 <ponyo> I got a great one, vlans.  Took me days to get him to wrap his head around the idea
2015-04-03 22:15:55 <_mjones> Mo0O: looks like I was wrong and install won't do symlinks. But ln -s should work fine.
2015-04-03 22:16:26 <ponyo> Or the one that wanted me to do reverse proxy back to his CRM install, but his CRM installed used relative paths for all it's helper files starting with ..
2015-04-03 22:16:42 <ponyo> and it was a subdir in, trying to come out
2015-04-03 22:16:48 <ponyo> So much fun!
2015-04-03 22:16:49 <_mjones> ponyo: ah. fairly simple one. When you're used to routing only, vlans seems slightly pointless. VLANs can be overused, too.
2015-04-03 22:17:01 <ponyo> Oh sure, but we had a mess without them
2015-04-03 22:17:06 <Mo0O> _mjones: ok, noted
2015-04-03 22:17:11 <ponyo> 30+ vms in the same subnet with no real order to anything
2015-04-03 22:17:47 <_mjones> I like reverse proxies. the base webserver qorked but the reverse-proxy choked? Normally there are plenty of rewriting options.
2015-04-03 22:17:52 <ponyo> It's so much cleaner now, mind you the mikrotik is doing a lot more work, but I took vpn away from it so I don't think it will be an issue and so far it's been working very well for me
2015-04-03 22:18:20 <_mjones> ponyo: I'm perfectly enthusiastic about running a /22 on one broadcast domain if there's no reason to do otherwise.
2015-04-03 22:18:34 <_mjones> s/qorked/worked/
2015-04-03 22:18:54 <ponyo> Yea I'm sure I could have re-wrote that one somehow, I use the crap out of haproxy and love it, but they had a url http://xx.xx.xx.xx:4000/CRM/ they wanted to reverse to from http://yy.yy.yy.yy/ and using relative paths kind of messed that up big time
2015-04-03 22:19:54 <_mjones> I'm a fan of that technique but I can't recall having a problem I couldn't fix, although sometimes it was harder to fix than you'd think.
2015-04-03 22:19:55 <ponyo> That's not a bad idea, but keeping them all in /24's has made it easy with the site to site vpn crap and the user vpn crap.  I needed vlans for the public side of the cable to the vm's anyways
2015-04-03 22:20:05 <_mjones> VPNs are crap, period.
2015-04-03 22:20:27 <ponyo> Yea I still haven't found a solution I really like
2015-04-03 22:20:31 <_mjones> IPsec works as a bridge under normal circumstances, and this is somewhat painful, somewhat unintuitive, and definitely nonscalable.
2015-04-03 22:20:38 <ponyo> Openvpn is so much hackery it's annoying with win32
2015-04-03 22:21:00 <ponyo> Right but IPSec with dynamic ip addresses is not something I look forward to dealing with anytime soon
2015-04-03 22:21:01 <_mjones> I'm currently a fan of TLS everywhere and everything outside of the locked datacenter is treated the same.
2015-04-03 22:21:16 <_mjones> VPNs are hard to support, too.
2015-04-03 22:21:27 <ponyo> Were SSL + TLS + user/pass/priv key
2015-04-03 22:21:42 <_mjones> split-horizon DNS is great until you have mobility meaning people are moving onto and off the LAN without knowing it
2015-04-03 22:22:08 <ponyo> I use *.lan for my internal .com's
2015-04-03 22:22:11 <_mjones> ponyo: sounds like a good setup. Maybe still ipsec for site-to-site....maybe...
2015-04-03 22:22:16 <ponyo> I agree :)
2015-04-03 22:22:28 <ponyo> People love to scold me about it
2015-04-03 22:22:36 <ponyo> but for private ip's I don't see the fuss
2015-04-03 22:22:42 <_mjones> Eh, you don't want to use anything fake. Many reasons, but consider this: you can't get signed a x.509 cert for anything you don't actually own.
2015-04-03 22:23:13 <_mjones> scold you about the .lan? That part I agree with. Other stuff is fine.
2015-04-03 22:23:27 <ponyo> Constantly I hear crap from people about it
2015-04-03 22:23:43 <ponyo> But anything internal I don't mind self signed certs
2015-04-03 22:24:04 <_mjones> it's very confusing for users.
2015-04-03 22:24:21 <_mjones> self-signed on admin interfaces might be acceptable, but users...have enough to deal with.
2015-04-03 22:24:22 <ponyo> Yea with the new breaks in the browsers, it's very annoying
2015-04-03 22:24:26 <Mo0O> do you know where can I found documentation on how to write openrc initscript?
2015-04-03 22:24:38 <ponyo> Gentoo's wiki I believe Mo0O
2015-04-03 22:24:50 <Mo0O> nice thx
2015-04-03 22:24:53 <_mjones> Mo0O: uh...I think Gentoo has a wiki article? Not certain. Let me know if you find a goo one.
2015-04-03 22:24:53 <ponyo> np :)
2015-04-03 22:25:11 <ponyo> https://wiki.gentoo.org/wiki/OpenRC
2015-04-03 22:25:59 <Mo0O> _mjones: sure
2015-04-03 22:26:22 <_mjones> Plus the commercial VPNs can be pretty nasty overall.
2015-04-03 22:26:43 <_mjones> Checkpoint's (perhaps older one) needs Java, and maybe only supported Windows. Problems on both counts.
2015-04-03 22:27:31 <_mjones> Cisco spent a while trying to push everyone to 'SSL' based (really dTLS), but instead of being unrestricted like ipsec, each ssl client session cost money. They wouldn't release a 64-bit ipsec windows client either.
2015-04-03 22:27:36 <ponyo> Yea I've been looking all over for a good turn key vpn
2015-04-03 22:27:43 <Mo0O> _mjones: this one looks realy cool: https://wiki.gentoo.org/wiki/Handbook:X86/Working/Initscripts
2015-04-03 22:27:53 <_mjones> Like I said, my new strategy is no VPNs, and especially no client VPNs.
2015-04-03 22:28:05 <ponyo> :)
2015-04-03 22:28:33 <_mjones> You're going to have to do competent PKI and TLS anyway, so once you internalise that, it's free after and it lets you get rid of a lot of other junk.
2015-04-03 22:28:39 <ponyo> I personally feel if I can't get it in in SSH, then i'm doing something wrong
2015-04-03 22:29:53 <_mjones> you mean crypto connections? I changed my mind on TLS a few years ago, and it feels so good.
2015-04-03 22:34:30 <ponyo> "VPN"
2015-04-03 22:36:42 <_mjones> Having distinct inside and outside addresses also leads to DNS resolution issues, even when split-horizon is properly implemented. This was especially bad with the old Microsoft PPTP connections, which seemed never to be able to consistently use 'internal' resolvers before external.
2015-04-03 22:38:28 <ponyo> Fun
2015-04-03 22:38:40 <ponyo> reminds me of my fun with NT 4 and RAS
2015-04-03 22:38:59 <_mjones> same thing, more or less.
2015-04-03 22:40:16 <_mjones> So yeah. Now the idea is just to access control everything by ID and never by IP, put TLS on it, and treat clients on the LAN the same as any other connection from elsewhere.
2015-04-03 22:40:49 <_mjones> Mostly what you have punching through traditional firewalls is https, http, ssh, and database connections. http and database go over TLS just fine.
2015-04-03 23:32:02 <_mjones> postgres-bdr vs. postgres-bdr-extension?
2015-04-03 23:45:28 <Mo0O> do you know how can I have the start-stop-deamon man page?
2015-04-03 23:56:53 <_mjones> It should have a manpage in base 'man-pages', in my opinion. Here's one: http://www.calculate-linux.org/main/en/start-stop-daemon
2015-04-04 00:59:27 <Mo0O> I'm trying to make an simple initscript, but it looks that openrc doesn't create the pid file, also running stop doesn't kill the process, here is my script: http://sprunge.us/KVdP?sh
2015-04-04 01:02:04 <_mjones> I'm not familiar with openrc yet, but that seems very reasonable init script.
2015-04-04 01:02:49 <_mjones> it doesn't stop because you have no stop() function, I'm reasonably sure.
2015-04-04 01:03:52 <_mjones> stop() { kill `cat ${pidfile}` }   # would probably work, but that's not a graceful shutdown!
2015-04-04 01:04:26 <Mo0O> _mjones: tested with a stop function, and doesn't work too, because the pid file is not created
2015-04-04 01:04:56 <_mjones> change the position of the pidfile command -- make it first, and see if that makes a difference.
2015-04-04 01:05:12 <_mjones> also, what happens if you run start-stop-daemon with those args by hand?
2015-04-04 01:05:54 <Mo0O> 
ACTION try

2015-04-04 01:09:04 <Mo0O> I've try to run: `start-stop-daemon --pidfile /var/run/elasticsearch.pid --background --start --exec /usr/share/elasticsearch/bin/elasticsearch` elasticsearch is launched, but no elasticsearch.pid file is present in /var/run
2015-04-04 01:09:13 <Mo0O> strange, not?
2015-04-04 01:09:36 <Mo0O> s/run/run by hand/
2015-04-04 01:11:02 <_mjones>  /var/run is where it should do, so I tentatively assume no security thing would prevent that.
2015-04-04 01:11:12 <_mjones> running as root, obviously?
2015-04-04 01:11:43 <Mo0O> yes, as root
2015-04-04 01:12:08 <jomat> Mo0O: I run es on alpine with this runscript: https://0.jmt.gr/?3bf17ab5c4ac9103#ybf1MHfhIQxBH9WXqER51d96hShhruqExj6/h1oc/1M=
2015-04-04 01:12:09 <_mjones> add -v for verbose.
2015-04-04 01:12:17 <jomat> perhaps that helps you :-)
2015-04-04 01:12:22 <Mo0O> :)
2015-04-04 01:12:34 <Mo0O> thanks jomat, I try that
2015-04-04 01:13:17 <jomat> oh, and /var/run/elasticsearch/ is owned by elasticsearch:elasticsearch
2015-04-04 01:14:35 <jomat> but checkpath should ensure that that is correct
2015-04-04 01:14:40 <Mo0O> that's what I was thinking after seing the `-user` option :)
2015-04-04 01:14:51 <jomat> :-)
2015-04-04 01:17:34 <jomat> well, obviously --pidfile /var/run/elasticsearch.pid should be replaced by --pidfile ${pidfile}  :-))
2015-04-04 01:17:45 <jomat> oh no
2015-04-04 01:17:59 <jomat> well, obviously I should go to bed or fetch another beer :-)
2015-04-04 01:18:27 <jomat> I mean -p /var/run/elasticsearch/elasticsearch.pid
2015-04-04 01:19:43 <Mo0O> 
ACTION give a beer to jomat

2015-04-04 01:20:00 <jomat> thx %)
2015-04-04 01:27:43 <Mo0O> you rox jomat :)
2015-04-04 01:28:24 <Mo0O> now, I just need to know how to create a user and fix permissions in aport
2015-04-04 01:28:30 <Mo0O> ^^
2015-04-04 01:31:53 <Mo0O> oh, I understand now, I was missing the elasticsearch `-d` option to daemonise... my bad
2015-04-04 01:33:49 <Mo0O> do you have an example aport which create a user and set permissions?
2015-04-04 01:35:24 <_mjones> you can probablt find one this way:
2015-04-04 01:35:51 <_mjones>  find ~/aport -type f -name APKBUILD -exec grep adduser {} \; -print
2015-04-04 01:36:31 <Mo0O> what a good idea, thanks _mjones
2015-04-04 01:36:33 <Mo0O> :)
2015-04-04 01:36:53 <_mjones> Except I see none. :(
2015-04-04 01:37:54 <jomat> Mo0O: There's an example: http://git.alpinelinux.org/cgit/aports/tree/main/mumble/murmur.pre-install
2015-04-04 01:41:49 <Mo0O> this story is going to cost me alot of beer :D
2015-04-04 01:44:54 <_mjones> That was my mistake, I was only looking in APKBUILD. The adduser command is always in the *.pre-install file.
2015-04-04 13:00:48 <rhapsodhy> is it only me who gets a gateway timeout for the forums?
2015-04-04 13:10:06 <AmatCoder> forum does not work since a few days :(
2015-04-04 13:12:03 <Jean-Scotch> forum is down till a new one is decided and deployed
2015-04-04 17:06:28 <darkfader> anyone considered simply putting up a maintenance page?
2015-04-04 17:06:32 <darkfader> just saying
2015-04-04 19:04:24 <_mjones> MMC is not enabled in the default kernel. Would be sort-of nice if it was, considering disk and diskless mode is likely to boot from mmc.
2015-04-04 22:04:29 <owen1>  i try to compile io.js by running 'make' and get: 'deps/cares/cares.target.mk:146: recipe for target '/tmp/io.js/out/Release/obj.target/cares/deps/cares/src/ares_cancel.o' failed' any ideas? - https://github.com/oren/test-iojs-docker-alpine
2015-04-04 22:19:37 <andyshinn> owen1: does iojs require g++?
2015-04-04 22:20:16 <clandmeter> owen1: anything in dmesg? and do you have enough space in /tmp?
2015-04-04 22:21:50 <owen1> andyshinn: let me see
2015-04-04 22:22:19 <owen1> clandmeter: i am not sure what do you mean by 'anything in dmesg'
2015-04-04 22:22:53 <clandmeter> your error doesnt state much, so im just guessing the obvious.
2015-04-04 22:23:01 <owen1> 223.1M  /tmp
2015-04-04 22:23:19 <clandmeter> owen1: you are running from ram?
2015-04-04 22:23:31 <clandmeter> tmpfs
2015-04-04 22:23:57 <owen1> andyshinn: https://github.com/iojs/io.js/#build
2015-04-04 22:24:05 <owen1> require gcc and g++
2015-04-04 22:24:39 <owen1> clandmeter: i don't know. how do i tell?
2015-04-04 22:25:08 <andyshinn> i'm just noting that your Dockerfile doesn't install g++
2015-04-04 22:25:11 <owen1> clandmeter: i follow the steps in the readme - https://github.com/oren/test-iojs-docker-alpine
2015-04-04 22:25:13 <clandmeter> if you installed alpine to hd, its not running fom ram
2015-04-04 22:25:33 <owen1> andyshinn: apk add gcc make python openssl-dev zlib-dev paxctl linux-headers xz
2015-04-04 22:26:02 <owen1> andyshinn: i was not able to install g++.  the build command errored
2015-04-04 22:26:11 <owen1> andyshinn: let me try now
2015-04-04 22:26:16 <clandmeter> bedtime. gn.
2015-04-04 22:26:31 <owen1> clandmeter: thanks!
2015-04-04 22:27:23 <andyshinn> maybe also worth pointing out that iojs is in testing: http://git.alpinelinux.org/cgit/aports/tree/testing/iojs/APKBUILD
2015-04-04 22:29:21 <owen1> andyshinn: adding g++ to Dockerfile. build was fine but 'docker-compose run web sh' => Cannot start container 183f6d22f933f37186d7011fbb2c9d9feaf126f0505a6b4f438122ca761cf2d6: no such file or directory
2015-04-04 22:30:09 <owen1> actualy no
2015-04-04 22:30:16 <andyshinn> well, that appears to be a Docker or Docker Compose issue, not a Alpine issue
2015-04-04 22:30:17 <owen1> the build was not successful: ERROR: busybox-1.23.1-r0.trigger: script exited with error 1
2015-04-04 22:31:14 <owen1> andyshinn: http://paste.ubuntu.com/10740413/
2015-04-04 22:32:08 <andyshinn> try building with --no-cache
2015-04-04 22:33:56 <owen1> wow. it looks better now. what's going on with that cache?
2015-04-04 22:35:53 <andyshinn> well it is based on edge tag and apk index probably gets out of sync pretty quickly
2015-04-04 22:35:54 <owen1> andyshinn: thank you! build is running now!!
2015-04-04 22:36:13 <andyshinn> might ass well just use --update or -U during apk add
2015-04-04 22:36:28 <andyshinn> err, that was an unfortunate extra 's'
2015-04-04 22:36:32 <owen1> ok
2015-04-04 22:36:48 <owen1> this is exciting
2015-04-04 22:46:13 <owen1> andyshinn: i think you are talking about --update-cache
2015-04-04 22:49:36 <andyshinn> yea, but i think --update actually works as well
2015-04-04 23:02:36 <owen1> i had to install curl as well since the tests uses it
2015-04-04 23:03:13 <owen1> my first failure might not be related to alpine - http://paste.ubuntu.com/10740550/
2015-04-04 23:27:11 <owen1> what about this error: AssertionError: { [Error: Command failed: /bin/sh -c ps -p 14517 -o args=
2015-04-04 23:27:13 <owen1> ps: unrecognized option: p
2015-04-04 23:33:17 <_mjones> that's the busybox error to which I was referring, owen1.
2015-04-04 23:33:27 <_mjones> busybox doesn't implement the -p.
2015-04-04 23:33:34 <owen1> _mjones: oh
2015-04-04 23:33:37 <owen1> gotcha
2015-04-04 23:34:06 <owen1> did u solve this issue?
2015-04-04 23:34:09 <_mjones> I'm not concerned about that one, but I'm somewhat concerned about the other errors. I get more on 1.6.3 than 1.6.2., but haven't had a chance to look if it's because of new or modified tests. The total test count seemed about the same, but I didn't check.
2015-04-04 23:34:29 <_mjones> I'm not sure if upstream busybox wants a patch for that. ;)
2015-04-04 23:34:46 <_mjones> In reality we'd probably stub out the check.
2015-04-04 23:35:17 <owen1> _mjones: that's my output: http://paste.ubuntu.com/10740701/
2015-04-04 23:35:19 <_mjones> My 1.6.3. compile is _probably_ fine, but tracking down the failed tests would take time.
2015-04-04 23:35:29 <rpu3uO8PEVZZ> apk add psutils and you have a bloated ps with -p
2015-04-04 23:36:44 <_mjones> owen1: interesting, my 1.6.3 failed more than 6 I think! Let me see again.
2015-04-04 23:38:13 <owen1> rpu3uO8PEVZZ: ho. let me try
2015-04-04 23:38:45 <_mjones> that will just fix the last test owen1.
2015-04-04 23:39:07 <owen1> _mjones: sure, better than nothing, right
2015-04-04 23:39:44 <owen1> rpu3uO8PEVZZ: ERROR: unsatisfiable constraints: psutils (missing): required by: world[psutils]
2015-04-04 23:40:50 <rpu3uO8PEVZZ> hmmz, what's the name of the pkg then? sorry for bad info.
2015-04-04 23:41:48 <rpu3uO8PEVZZ> procps
2015-04-04 23:42:04 <owen1> ok
2015-04-04 23:55:31 <_mjones> [11:45|% 100|+ 830|-   8]: Done    # ( 8 tests failed. I would've sworn it was many more than that.)
2015-04-04 23:55:50 <_mjones> With only 8 tests failing, there's no major regression. I say: ship it!
2015-04-04 23:56:25 <owen1> _mjones: 4 failures on mine
2015-04-04 23:56:35 <owen1> (after adding procps)
2015-04-05 00:02:36 <owen1> https://github.com/iojs/docker-iojs/issues/44#issuecomment-89690405
2015-04-05 01:21:47 <owen1> is there a way to color my script in sh? in bash i use green=`tput setaf 2`
2015-04-05 01:21:56 <owen1> but tput is not available
2015-04-05 02:17:06 <BitL0G1c> owen1 - you will need to use ANSI codes
2015-04-05 02:17:56 <owen1> http://misc.flogisoft.com/bash/tip_colors_and_formatting ?
2015-04-05 02:18:06 <BitL0G1c> see /etc/profile.d/color_prompt
2015-04-05 02:18:37 <BitL0G1c> yes that
2015-04-05 02:19:00 <owen1> BitL0G1c: sweet. thanks
2015-04-05 02:19:19 <BitL0G1c> ;-)
2015-04-05 02:19:32 <owen1> i guess i'll start use that approach so my bash script will be reusable
2015-04-05 02:19:41 <owen1> and maybe it's time to get used to sh
2015-04-05 02:19:57 <owen1> or ash, whatever you people calls it ):
2015-04-05 02:20:00 <owen1> i meant smily
2015-04-05 02:20:14 <BitL0G1c> the only thing ash can't do is arrays
2015-04-05 02:20:33 <owen1> BitL0G1c: interesting. so what do u do when you need one?
2015-04-05 02:20:58 <BitL0G1c> use bash :) if i really need an array
2015-04-05 02:22:03 <owen1> oh
2015-04-05 02:22:40 <BitL0G1c> with ash you could just store the data in a file instead of an array
2015-04-05 02:23:34 <BitL0G1c> & use set -- to access the elements of a line
2015-04-05 02:24:56 <owen1> interesting
2015-04-05 02:28:31 <owen1> i am probably doing something wrong -  echo $GREENHello World$NORMAL
2015-04-05 02:29:06 <owen1> i copied from /etc/profile.d/color_prompt
2015-04-05 02:29:09 <owen1> GREEN="\[\e[1;32m\]"
2015-04-05 02:29:23 <owen1> but it's not coloring anything..
2015-04-05 02:29:25 <BitL0G1c> quote "       " the echo
2015-04-05 02:29:56 <owen1>  echo "$GREENHello World$NORMAL"
2015-04-05 02:29:57 <BitL0G1c> & echo -e "$GREEN xxxxxxxx $NORMAL"
2015-04-05 02:30:33 <BitL0G1c> need the '-e'
2015-04-05 02:31:23 <BitL0G1c> also check echo $GREEN (no quotes) - prints the ANSI codes in your environment
2015-04-05 02:32:23 <owen1> echo $GREEN => \[\e[1;32m\]
2015-04-05 02:32:37 <owen1> echo -e "$GREEN Hello World $NORMAL" => \[\] Hello World \[\]
2015-04-05 02:34:58 <BitL0G1c> do something like this http://hastebin.com/azoveqilos.pl
2015-04-05 02:39:08 <owen1> BitL0G1c: http://i.imgur.com/qOPeugX.png
2015-04-05 02:39:27 <owen1> i don't think it's colorful
2015-04-05 02:39:29 <owen1> maybe bold?
2015-04-05 02:41:12 <BitL0G1c> you should probably put your colour functions in ~/.profile or /etc/profile & log off & on again
2015-04-05 02:44:37 <BitL0G1c> http://i.imgur.com/5PZInoK.png when i run it
2015-04-05 02:48:19 <BitL0G1c> to get a colour prompt 'mv ./etc/profile.d/color_prompt /etc/profile.d/color_prompt.sh'
2015-04-05 02:48:47 <owen1> 1 sec
2015-04-05 02:53:07 <owen1> BitL0G1c: before i move the function to ~./profile, i need to figure out why it's not colorful
2015-04-05 03:02:40 <vkrishn> try, https://github.com/insteps/scripts
2015-04-05 03:03:00 <vkrishn> I usually include, https://github.com/insteps/scripts/blob/master/bash/inc/color.inc
2015-04-05 03:06:58 <vkrishn> source color.inc && echo -e "$cRED xxxxxxxx $cNORMAL"
2015-04-05 03:10:35 <owen1> vkrishn: 1 sec
2015-04-05 03:15:08 <owen1> vkrishn: works!
2015-04-05 03:15:11 <owen1> thanks
2015-04-05 03:20:56 <owen1> vkrishn: looks good, except for yellow. it just look bold
2015-04-05 03:22:18 <vkrishn> replace that from http://misc.flogisoft.com/bash/tip_colors_and_formatting schemes
2015-04-05 03:23:56 <vkrishn> I just needed few colors for scripts output
2015-04-05 03:24:26 <vkrishn> now I use clog, http://tasktools.org/projects/clog.html
2015-04-05 03:43:15 <owen1> thanks for clog. looks nice
2015-04-05 04:51:52 <owen1> BitL0G1c: i would like to avoid 'source /root/.colors' in each of my script files. is there a file like .bashrc or .bash_profile that is loaded automaticaly? (i am using docker-componse run my-app')
2015-04-05 19:41:44 <kahiru1> hey, is there something like mainline kernel package for alpine?
2015-04-05 19:43:05 <jomat> you mean without grsec? that's linux-vanilla :-)
2015-04-05 19:49:39 <kahiru1> i meant more up to date, something 3.19-ish
2015-04-05 19:58:12 <dalias> :)
2015-04-05 19:58:24 <dalias> or 4.1rc :)
2015-04-05 19:58:42 <kahiru1> dalias: im not that adventurous
2015-04-05 20:09:44 <kahiru1> another question, did anyone manage to get zfs working?
2015-04-05 22:27:19 <xenkey> Hi
2015-04-05 22:27:32 <xenkey> This apk has coffee making abilities. This is like apt-get isn't it? :)
2015-04-05 22:31:37 <c355E3B> i think apt-get has super cow powers
2015-04-05 22:59:16 <Mo0O> hi there
2015-04-05 23:01:00 <Mo0O> I've made an aport, the application is functionnal but I have this errors when running `abuild -r`:
2015-04-05 23:01:03 <Mo0O> BFD: ./usr/share/elasticsearch/lib/sigar/libsigar-ia64-linux.so: warning: sh_link not set for section `.IA_64.unwind'
2015-04-05 23:01:05 <Mo0O> strip: Unable to recognise the format of the input file `./usr/share/elasticsearch/lib/sigar/libsigar-ia64-linux.so'
2015-04-05 23:01:17 <Mo0O>  libc.so.6: path not found
2015-04-05 23:01:45 <Mo0O> is it sane to ignore those errors or is there a way to avoid it?
2015-04-05 23:01:49 <dalias> sounds like a glibc-linked binary is involved somewhere..
2015-04-05 23:02:04 <Mo0O> yes
2015-04-05 23:39:19 <owen1> how to view all the processes that running on my alpine? 'ps' doesn't show me a lot and i know i have a webserver running.
2015-04-05 23:39:31 <xenkey> owen1: ps ax
2015-04-05 23:39:39 <xenkey> ps faux if you're feeling verbose
2015-04-05 23:44:43 <nsz> those args shouldnt be needed for busybox ps
2015-04-05 23:45:58 <nsz> if you are on grsec kernel you wont see everything in ps unless you are root
2015-04-05 23:48:17 <xenkey> we have a grsec kernel?
2015-04-05 23:50:38 <nsz> alpine uses grsec kernel by default, but vanilla kernel is packaged as well, i dont know what owen uses
2015-04-05 23:54:39 <owen1> xenkey: thanks
2015-04-05 23:57:27 <owen1> xenkey: i run top and than in another terminal i sh again and i only see:     1 root       0:00 sh
2015-04-05 23:57:29 <owen1>     9 root       0:00 ps faux
2015-04-05 23:57:33 <owen1> i don't see top running
2015-04-06 00:05:24 <nsz> grep . /proc/[1-9]*/comm
2015-04-06 00:05:26 <nsz> as root
2015-04-06 00:06:30 <xenkey> my alpine from online labs isn't grsec...
2015-04-06 00:06:43 <clandmeter> 
ACTION loves alpine on QHD

2015-04-06 00:07:54 <xenkey> QHD?
2015-04-06 00:08:32 <clandmeter> QHD (Quad HD), also sometimes advertised as WQHD
2015-04-06 00:08:56 <clandmeter> 2560x1440
2015-04-06 00:14:57 <xenkey> Jesus
2015-04-06 00:16:27 <clandmeter> even jesus looks better on QHD ;-)
2015-04-06 00:17:03 <clandmeter> xenkey: what is `online labs`?
2015-04-06 00:17:12 <xenkey> labs.online.net
2015-04-06 00:17:38 <xenkey> Linux hostname 3.19.1-181 #1 SMP Thu Mar 12 15:57:27 UTC 2015 armv7l Linux
2015-04-06 00:20:22 <clandmeter> heh, https://github.com/scaleway/image-alpine
2015-04-06 00:32:43 <vkrishn> with 80columns set on QHD, I am sure its good for "remote desktop viewing" from 2blocks away withour need of binoculars
2015-04-06 00:33:04 <clandmeter> hi vkrishn
2015-04-06 00:33:09 <vkrishn> is there a keyboard that works at that range ?
2015-04-06 00:33:24 <vkrishn> then we would have a revolutionary remote-desktop
2015-04-06 00:33:26 <vkrishn> hi
2015-04-06 00:33:50 <clandmeter> well, atleast it scales better then on windows
2015-04-06 00:33:55 <vkrishn> :)
2015-04-06 00:34:28 <clandmeter> im running edge now from ram
2015-04-06 00:34:37 <clandmeter> speedy :)
2015-04-06 00:36:20 <clandmeter> about 1g of memory used (including libreoffice)
2015-04-06 00:37:50 <clandmeter> seems we dont have cpupower in aports (linux-tools)
2015-04-06 00:38:21 <clandmeter> my notebook is about to lift off
2015-04-06 00:39:54 <xenkey> notebook?
2015-04-06 00:40:24 <clandmeter> laptop?
2015-04-06 00:40:45 <vkrishn> good there is only one fan, with 4 you have a drone
2015-04-06 00:41:19 <clandmeter> :)
2015-04-06 00:41:36 <clandmeter> i need to get a new desktop
2015-04-06 00:43:18 <clandmeter> i hope my colleagues want to to share with me.
2015-04-06 00:46:45 <clandmeter> bedtime here. gnite.
2015-04-06 00:59:33 <vkrishn> found another virtual keyboard, https://github.com/kikito/lua-uinput
2015-04-06 01:01:25 <vkrishn> gtg
2015-04-06 01:28:28 <vkrishn> hmm... and I forgot I had logged in ;)
2015-04-06 01:28:49 <vkrishn> is testing/opennebula/ still maintained ?
2015-04-06 01:29:16 <vkrishn> now gtg
2015-04-06 02:41:36 <ttilley> how painful would it be to install alpine from within another OS? any chance i can do an unpack-and-chroot like gentoo?
2015-04-06 02:43:27 <ttilley> nevermind, found The Friendly Manual - http://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot
2015-04-06 03:38:22 <Mo0O> do you know if there a way to fix the `<rpc/rpc.h>` missing library?
2015-04-06 03:56:43 <dalias> did you try what i suggested on #musl ?
2015-04-06 03:58:45 <Mo0O> dalias: yes
2015-04-06 14:34:37 <botto> I'm trying to access the forums but I get a 504
2015-04-06 19:45:23 <knite> hi guys. just tried installing node-sass on Alpine and it's failing. it's a pretty common package, so I'm wondering if it might an issue on the alpine side: /app/node_modules/node-sass/vendor/linux-x64-iojs-1.6/binding.node: __strtod_internal: symbol not found
2015-04-06 19:45:23 <knite> before I ping either the iojs or node-sass devs.
2015-04-06 19:45:23 <knite> libstdc++-is version 4.9.2-r4
2015-04-06 19:46:27 <nsz> where did __strtod_internal come from?
2015-04-06 19:46:35 <nsz> it is a glibc thing
2015-04-06 19:46:40 <dalias> sounds like you installed a binary from somewhere random?
2015-04-06 19:46:47 <dalias> build from source and it should work fine
2015-04-06 19:47:03 <nsz> and dont include glibc headers :)
2015-04-06 19:47:21 <knite> dalias: I installed the Alpine build of iojs.
2015-04-06 19:48:42 <nsz> btw why is nodejs binary called 'node' on alpine? i thought the binary is 'nodejs' (iirc that was the name on debian)
2015-04-06 19:49:44 <dalias> it seems node is common
2015-04-06 19:50:22 <nsz> (node is a horrible name for a js interpreter.. there are 9-10 different js interpreters at this point none of them with reasonable name so you cannot write scripts that invoke them reliably)
2015-04-06 19:50:33 <dalias> :)
2015-04-06 19:51:11 <dalias> is npm downloading binary packages or building from source? i'm not familiar with how node.js stuff works
2015-04-06 19:51:29 <dalias> anyway it sounds like knite's problem is coming from having downloaded binaries from somewhere
2015-04-06 19:52:00 <knite> I'm pretty sure npm builds what it needs rather than downloading binaries?
2015-04-06 19:52:54 <knite> either way, this is a bit of an issue. I had a related issue with getting iojs running in the first place a couple of weeks ago. the fix was bumping libstdc++ to the latest version on edge/main at the time.
2015-04-06 19:53:21 <knite> so how can I troubleshoot this issue?
2015-04-06 19:53:53 <nsz> it's not clear to me which binary refers to that symbol
2015-04-06 19:54:11 <nsz> see if that binary comes from the alpine package
2015-04-06 19:54:20 <nsz> if it's from the package then it's a packaging bug
2015-04-06 19:54:49 <nsz> you can use nm binary to see the symbols it references
2015-04-06 19:54:53 <nsz> or nm -D
2015-04-06 19:55:38 <nsz> you should look for a line like 'U __strtod_internal'
2015-04-06 19:55:48 <knite> hm...npm cleaned up after itself, so I can't access that binding.node file...
2015-04-06 20:01:22 <dalias> knite, the error message you saw: where does it come from? can you tell?
2015-04-06 20:05:15 <knite> dalias: I can dig in but it'll be a bit of work. I'm not super familiar with node/npm.
2015-04-06 20:08:11 <dalias> what were you doing when the error appeared?
2015-04-06 20:08:17 <dalias> installing? running? what command? etc.
2015-04-06 20:14:51 <knite> dalias: oh, well that's simple. just a npm install node-sass.
2015-04-06 20:14:58 <knite> on a system with:
2015-04-06 20:15:43 <knite> apk add -U -X http://nl.alpinelinux.org/alpine/edge/main/ git=2.3.5-r0 libstdc++=4.9.2-r4; apk add -U -X http://nl.alpinelinux.org/alpine/edge/testing/ iojs=1.6.2-r0
2015-04-06 20:16:43 <nsz> almost surely npm does something wrong then
2015-04-06 20:18:42 <nsz> knite: ok i could reproduce the issue
2015-04-06 20:21:09 <dalias> perhaps node-sass is distributing binaries in their package rather than correctly having npm build source; i dunno
2015-04-06 20:24:22 <nsz> yes that seems to be happening
2015-04-06 20:24:45 <Mo0O> hi there
2015-04-06 20:25:39 <Mo0O> I don't understand why the $pkgname.pre-install is not executed: http://sprunge.us/ZaTN
2015-04-06 20:26:36 <Mo0O> here is the $pkgname.pre-install: http://sprunge.us/QfCM
2015-04-06 20:27:11 <Mo0O> can you see where I'm wrong?
2015-04-06 20:27:46 <Mo0O> I guess it's a typo error, but can't figure out where is the error
2015-04-06 20:28:36 <knite> Aria: here's a clue that points to a problem with alpine.
2015-04-06 20:29:11 <knite> using iojs on my laptop, the same version of node-sass installs correctly.
2015-04-06 20:30:07 <knite> though I suppose it could be that whatever they're doing builds correctly on OS X but not Linux.
2015-04-06 20:30:14 <knite> will try under Ubuntu, need a minute.
2015-04-06 20:32:29 <knite> also, why has http://forum.alpinelinux.org/packages/ been timing out for like a week?
2015-04-06 20:33:20 <darkfader> from what i read it got stuck mid-upgrade and noone feels like putting up a maintenance page :/
2015-04-06 20:34:31 <Mo0O> well, I've run the script manualy, will test on deployement ;)
2015-04-06 20:35:47 <nsz> knite: http://sprunge.us/FbXP
2015-04-06 20:36:11 <nsz> this is what happens in strace -f npm install node-sass
2015-04-06 20:36:29 <knite> what does it mean?
2015-04-06 20:37:10 <nsz> it downloads an elf binary (from github?) over https then dlopens it i think (mmap as executable)
2015-04-06 20:37:53 <nsz> i assume it is some extension module written in c and precompiled
2015-04-06 20:39:41 <Mo0O> nsz: npm install node-sass@alpha  # will build the binary
2015-04-06 20:39:57 <Mo0O> wait a minute I'll give you the related issue
2015-04-06 20:40:34 <Mo0O> https://github.com/sass/node-sass/issues/808 here it is
2015-04-06 20:41:35 <nsz> Mo0O: as clearly visible in the strace the binary is downloaded and when it fails the build exits with error
2015-04-06 20:41:41 <Mo0O> looks now stable version v2.1.1 is released with the fix
2015-04-06 20:41:59 <knite> er...so their solution is to add a dependendy on *python*?
2015-04-06 20:42:04 <knite> dependency*
2015-04-06 20:42:08 <Mo0O> no
2015-04-06 20:42:52 <knite> that's my reading of the issue. they're building on CentOS. prebuilt fails due to musl. fallback is to manually build (which was broken and then fixed), which requires python.
2015-04-06 20:42:55 <Mo0O> nsz: which version are you installing in the sprunge strace ?
2015-04-06 20:43:12 <dalias> <Mo0O> nsz: npm install node-sass@alpha  # will build the binary
2015-04-06 20:43:14 <knite> which means my nice tiny Alpine container will need to be pretty big.
2015-04-06 20:43:16 <dalias> what does the @alpha mean?
2015-04-06 20:44:02 <nsz> Mo0O: i have nodejs-0.10.33-r0 and i just did 'npm install node-sass'
2015-04-06 20:44:24 <dalias> it's disturbing that npm is downloading binaries with no way to know if they match the system it's running on
2015-04-06 20:44:41 <Mo0O> it normaly build alpha release dalias, I personaly set it using package.json
2015-04-06 20:45:14 <dalias> nsz, and it worked?
2015-04-06 20:45:23 <dalias> perhaos knite manually configured it to download incompatible binaries? :-p
2015-04-06 20:45:40 <knite> dalias: ha, not at all.
2015-04-06 20:45:45 <nsz> no it didnt work
2015-04-06 20:45:50 <nsz> i got the same error as knite
2015-04-06 20:45:57 <nsz> so i straced the whole thing
2015-04-06 20:46:03 <knite> I'm adding a comment to that issue even though it's closed.
2015-04-06 20:46:30 <nsz> but the strace log is huge because it's some billion process nonsense with epolling and futexing everywhere
2015-04-06 20:46:53 <nsz> http://sprunge.us/FbXP was just the relevant fragment
2015-04-06 20:48:58 <nsz> ok with @alpha it works
2015-04-06 20:49:12 <dalias> ok
2015-04-06 20:49:25 <dalias> anyway @alpha is not the right fix
2015-04-06 20:49:29 <nsz> (by works i mean it detects the error and then starts some build process)
2015-04-06 20:49:36 <dalias> because this will happen again with other packages
2015-04-06 20:50:25 <dalias> alpine may need some patch to the node/npm package to keep it from identifying the system as "GNU/Linux" and trying to download binary packages
2015-04-06 20:50:45 <Mo0O> dalias: no but @v2.1.1 could be
2015-04-06 20:50:54 <dalias> mo0o, ?
2015-04-06 20:51:08 <Mo0O> Mo0O │ looks now stable version v2.1.1 is released with the fix
2015-04-06 20:51:10 <Mo0O> Mo0O │ looks now stable version v2.1.1 is released with the fix
2015-04-06 20:51:12 <Mo0O> Mo0O │ looks now stable version v2.1.1 is released with the fix
2015-04-06 20:51:19 <Mo0O> opps
2015-04-06 20:51:21 <Mo0O>  sorry
2015-04-06 20:51:36 <Mo0O> but the issue tell the fix is available on v2.1.1
2015-04-06 20:51:40 <dalias> ?
2015-04-06 20:51:49 <nsz> Mo0O: the problem is that you cannot guarantee that all packages with binary extensions are fixed
2015-04-06 20:51:51 <Mo0O> which issue?
2015-04-06 20:51:54 <dalias> what fix?
2015-04-06 20:51:57 <nsz> and they uselessly download broken binaries
2015-04-06 20:52:02 <Mo0O> https://github.com/sass/node-sass/issues/808
2015-04-06 20:52:07 <nsz> that's not the issue
2015-04-06 20:52:18 <nsz> the issue is that npm tries to run glibc binaries
2015-04-06 20:52:23 <dalias> *nod* that's not the issue
2015-04-06 20:52:27 <Mo0O> ok sorry, I've miss-understand
2015-04-06 20:52:31 <nsz> that can siletly succeed and do horrible things
2015-04-06 20:53:24 <dalias> trying the binary at all is a bug
2015-04-06 20:53:29 <dalias> it just needs to always get the source
2015-04-06 20:54:11 <dalias> there's no guarantee that the mismatching binary will fail in a detectable way and cause the fallback to building from source
2015-04-06 20:54:19 <dalias> it could silently seem to work but have problems
2015-04-06 21:26:48 <systmkor> dalias: definitely the scarier side of software engineering
2015-04-06 21:30:46 <dalias> systmkor, btw i saw your messages. no problem. but i don't want to get dragged into that topic again right now :-)
2015-04-06 21:33:30 <systmkor> for sure
2015-04-06 21:34:01 <systmkor> note I don't care about talking about systemd just more of the topics behind it and other things like OpenRC, runit, supervisord etc.
2015-04-06 21:34:07 <systmkor> :D
2015-04-06 21:35:15 <dalias> *nod*
2015-04-06 21:35:38 <dalias> but even that level feels like a time sink to me
2015-04-06 21:36:03 <dalias> i just want this stuff to stay out of my way (especially out of the dependencies of apps that should have nothing to do with it)
2015-04-06 21:38:09 <dalias> i don't consider myself an expert on solving all the problems of programmatic management of services, control panels, ... or whatever ppl who want these apis want
2015-04-06 21:38:21 <systmkor> fo sho :D
2015-04-06 21:39:47 <dalias> but i do consider myself qualified to have an opinion that the way my system's init, daemon starting, supervision, etc. works should not be dictated (and forced to be 1000x more complex than it needs to be) by the design that happens to work for someone with very different system needs
2015-04-06 21:40:32 <systmkor> agreed
2015-04-06 21:40:56 <systmkor> is red hat strong arming companies and distros to adopt it?
2015-04-06 21:41:01 <systmkor> like financially or something?
2015-04-06 21:41:13 <dalias> not afaik
2015-04-06 21:41:37 <dalias> rather it's poettering's gang/fdo that's putting pressure
2015-04-06 21:41:46 <systmkor> ahhhh
2015-04-06 21:41:55 <systmkor> so more of socially pressuring
2015-04-06 21:42:17 <dalias> "fdo component X that everybody needs now 'optionally' depends on systemd-whatever"
2015-04-06 21:42:28 <dalias> etc
2015-04-06 21:42:36 <systmkor> got it
2015-04-06 21:42:49 <systmkor> well your meeting go well?
2015-04-06 21:43:09 <dalias> i think so
2015-04-06 21:43:18 <systmkor> :thumbs up:
2015-04-06 21:43:49 <systmkor> oh over the weekend I started going down a list of pam modules and seeing if they compile on Alpine
2015-04-06 21:44:12 <systmkor> so hopefully there will be like 10 system authentication ports coming this weekend
2015-04-06 21:44:28 <systmkor> during compilation I kept seeing sys/cdefs.h is deprecated
2015-04-06 21:44:46 <systmkor> and didn't get a chance to look up what the proper replacement of such was
2015-04-06 21:45:18 <systmkor> those programs finished compiling any way but need to learn so I can patch code if it becomes an issue because of it
2015-04-06 21:45:18 <dalias> normally it's one of two things
2015-04-06 21:45:54 <dalias> either (a) using __BEGIN_DECLS / __END_DECLS in place of explicitly writing #ifdef __cplusplus / extern "C" { / ...
2015-04-06 21:46:27 <dalias> or (b) putting __P() around args in function prototypes out of some misguided notion you need to support pre-ANSI compilers
2015-04-06 21:47:43 <systmkor> ah m'kay
2015-04-06 21:47:47 <systmkor> thnx
2015-04-06 21:48:08 <dalias> the motivation is unclear in both cases
2015-04-06 21:48:35 <dalias> since there's an explicit portable way to do the same thing without depending on __-prefixed macros from a nonstandard header
2015-04-06 21:49:33 <systmkor> what's the explicit portable way
2015-04-06 21:51:59 <systmkor> I didn't get a chance to look at the projects' code yet
2015-04-06 21:56:53 <nsz> #ifdef __cplusplus / extern "C" {
2015-04-06 22:01:51 <systmkor> ah m'kay cool
2015-04-06 22:43:55 <iwearsable> Anybody have any experience getting Alpine working on a laptop with Intel 945GM? I've tried so many different things now that I'm losing my mind.
2015-04-06 22:46:14 <iwearsable> Current state is that screen goes black during boot, and I log in to tty while blind then type startx (xinitrc contains openbox-session). This starts up openbox and I can see mouse cursor but it takes a few minutes for the touchpad buttons to work. If I then open firefox it has graphical corruption but most other things work. I'm also getting "error connecting to system bus" (dbus) spammed in xorg
2015-04-06 22:46:20 <iwearsable> 's log file, same as bug #3691
2015-04-06 22:47:54 <iwearsable> I read on arch wiki that firefox corruption can be fixed by setting UXA accelmethod but I'm not sure how to do this without an xorg.conf file. Mainly I need to get rid of the blank screen while in a tty, get rid of corruption in firefox, and fix that dbus issue which I assume has something to do with the touhpad malfunction
2015-04-06 22:52:37 <iwearsable> /etc/module has intel_agp before i915 and I rebuilt initramfs, that's when I got xorg to work but tty broke (before that I always ot no screens found or no device)
2015-04-06 23:10:02 <iwearsable> Xorg.0.log: http://sprunge.us/BNGL
2015-04-06 23:25:23 <iwearsable> I think the black screen in tty is probably due to no backlight, I think I can fix that but I'm still not sure about the graphical corruption and dbus issue
2015-04-06 23:26:21 <BitL0G1c> did you run 'setup-xorg-base' ?
2015-04-06 23:41:31 <iwearsable> I did do that, when first setting it up I ran though the XFCE setup wiki then just changed xinitrc to openbox-session. I just ran it again though and rebooted with no improvement.
2015-04-06 23:42:00 <iwearsable> Also I thought that setpci -s 00:02.0 F4.B=FF would help my tty issue but it only adjusts backlight while in X, no visible change in tty
2015-04-06 23:42:41 <iwearsable> And the two physical buttons belowmy touchpad don't work after starting X, until I switch to a tty and back
2015-04-06 23:48:12 <iwearsable> DMESG: http://sprunge.us/iMRX
2015-04-07 00:05:33 <dalias> the intel graphics drivers were broken for a long time in alpine; i dunno if they were fixed
2015-04-07 00:05:47 <dalias> i have my X pinned at the version from 3.0 stable
2015-04-07 00:05:52 <dalias> because later versions broke
2015-04-07 00:12:06 <iwearsable> Yeah, I've spent enough time on this over the past weeks. I think I'll try Void on it instead. Nothing more frustrating than graphics/X issues I tell you what :)
2015-04-07 00:13:11 <iwearsable> Always end up reading ancient or irrelevant fixes for other distros and trying them too, which usually just makes things worse because you can't remember all of the things you changed
2015-04-07 00:14:43 <nsz> it will get worse as gpu things get more complex and nobody seem to care about sane abi and stability
2015-04-07 00:16:18 <iwearsable> I actually have video problems on all of the computers I've tried with Alpine. My desktop had a GTX 750 which doesn't have working nouveau drivers in LTS, switched it to an 8800GT but 8-series nvidia also has graphical corruption issues with nouveau, thankfully infrequently.
2015-04-07 00:17:29 <iwearsable> I still love Alpine though that's why I tried to work through this intel thing
2015-04-07 00:18:01 <dalias> iwearsable, try installing X from 3.0
2015-04-07 00:21:58 <iwearsable> How do I get X from 3.0? I don't know how to do that
2015-04-07 00:23:25 <iwearsable> Actually, nevermind Ithink I figured it out
2015-04-07 00:26:11 <dalias> the wiki documents adding a tagged repo and pinning packages to it
2015-04-07 00:26:27 <dalias> i forget the details and always have to go find them again when i need them
2015-04-07 00:37:06 <iwearsable> I got it to downgrade but X won't start. The backlight flickers but then it returns to the broken black tty, so I can't see the error. It was worth a try but I'm done screwing with it.
2015-04-07 00:37:31 <iwearsable> Thanks for the help though, dalias and BitL0G1c
2015-04-07 00:44:20 <BitL0G1c> 3.0 was the last xorg i had working too
2015-04-07 08:38:24 <ncopa> dalias: i use xorg with intel chipset on my laptop and on my desktop
2015-04-07 10:10:49 <clandmeter> could anyone test if forum.alpinelinux.org is working?
2015-04-07 10:12:23 <HarryHaaren> clandmeter, yep its loaded here
2015-04-07 10:12:47 <clandmeter> dank u
2015-04-07 10:29:42 <_ikke_> ncopa: Already have someone testing redmine in docker?
2015-04-07 10:32:38 <clandmeter> we dont
2015-04-07 10:34:47 <clandmeter> _ikke_: if you are up for the task, go for it.
2015-04-07 10:44:26 <_ikke_> Any preference for an alpine image? There are several available
2015-04-07 10:45:17 <_ikke_> Ah, there is also just "Alpine"
2015-04-07 10:48:46 <clandmeter> it makes sence to have docker images for the apps we currently use in our infra
2015-04-07 10:49:18 <clandmeter> redmine, drupal, mediawiki
2015-04-07 11:01:37 <ncopa> hi
2015-04-07 11:01:45 <xenkey> hi
2015-04-07 11:02:08 <ncopa> _ikke_: the main point is to have an easy way to test if redmine packages are actually working or not
2015-04-07 11:02:24 <ncopa> so please use the packages and not bundler
2015-04-07 11:02:37 <ncopa> and make it use latest edge packages
2015-04-07 12:19:11 <_ikke_> ncopa: yeah, I got that
2015-04-07 14:16:29 <Tsutsukakushi> alpinelinux.org down?
2015-04-07 14:53:58 <clandmeter> NL infra isp is currently down, which means www,nl master mirror, bugs,forum, and probably more
2015-04-07 15:05:47 <dalias> :/
2015-04-07 15:12:29 <clandmeter> soon ill be going home, hoping it returs asap.
2015-04-07 16:17:12 <Psi|4ward> Hi there, where can i get the mariadb client headers ?
2015-04-07 16:25:09 <_ikke_> apk add mariadb-dev ?
2015-04-07 16:36:46 <_ikke_> Psi|4ward: ^
2015-04-07 16:38:11 <Psi|4ward> yes worked but "make" gives me now 100000 errors :\
2015-04-07 16:38:21 <Psi|4ward> think ill stick to ubuntu for this image
2015-04-07 16:39:28 <xenkey> :/
2015-04-07 16:40:08 <Psi|4ward> https://gist.github.com/psi-4ward/829f1e04ac9ba2e12448
2015-04-07 16:40:10 <Psi|4ward> any ideas?
2015-04-07 16:41:28 <_ikke_> seems like nss is missing
2015-04-07 16:42:10 <Psi|4ward> should be bundled here? http://libnss-mysql.sourceforge.net/
2015-04-07 16:42:51 <_ikke_> perhaps apk add nss-dev
2015-04-07 16:43:00 <Psi|4ward> apk search nss
2015-04-07 16:43:04 <Psi|4ward> Mozilla Security stuff
2015-04-07 16:43:16 <BitL0G1c> apk search nss|grep dev
2015-04-07 16:43:29 <_ikke_> nss-dev should exist
2015-04-07 16:43:37 <_ikke_> http://git.alpinelinux.org/cgit/aports/tree/main/nss/APKBUILD
2015-04-07 16:43:44 <BitL0G1c> musl64 [~]$ apk search nss|grep dev
2015-04-07 16:43:44 <BitL0G1c> nss-dev-3.17.4-r0
2015-04-07 16:44:01 <Psi|4ward> nss-dev-3.17.4-r0 description:
2015-04-07 16:44:02 <Psi|4ward> Mozilla Network Security Services
2015-04-07 16:44:04 <Psi|4ward> nss-dev-3.17.4-r0 description:
2015-04-07 16:44:05 <Psi|4ward> Mozilla Network Security Services
2015-04-07 16:44:14 <Psi|4ward> its something completly different
2015-04-07 16:44:29 <_ikke_> Psi|4ward: different from what?
2015-04-07 16:44:41 <Psi|4ward> has nothing todo with NameServiceSwitch
2015-04-07 16:44:47 <_ikke_> ah ok
2015-04-07 16:46:17 <_ikke_> Psi|4ward: what are you building?
2015-04-07 16:47:10 <Psi|4ward> libnss-mysql to authenticate system users against MySQL
2015-04-07 16:47:43 <_ikke_> ah, ok
2015-04-07 16:51:38 <_ikke_> https://github.com/PADL/nss_ldap/blob/master/nss_common.h
2015-04-07 16:52:49 <Psi|4ward> nss.h should be part of glibc
2015-04-07 16:52:57 <Psi|4ward> is there a glibc in alpine?
2015-04-07 16:53:02 <_ikke_> not anymore
2015-04-07 16:53:09 <_ikke_> they changed to musl
2015-04-07 16:53:24 <BitL0G1c> looking at the readme you probably have to disable PAM support to build it
2015-04-07 16:53:53 <Psi|4ward> user auth without PAM?
2015-04-07 16:54:27 <Psi|4ward> whatever, lets use ubuntu for this container, it has alle neede pkgs
2015-04-07 17:01:38 <BitL0G1c> you are missing nss.h - #include <nss.h>
2015-04-07 17:01:56 <Psi|4ward> probably but where to get it?
2015-04-07 17:03:33 <BitL0G1c> nss-dev
2015-04-07 17:04:23 <Psi|4ward> http://pkgs.alpinelinux.org/package/main/x86/nss-dev
2015-04-07 17:04:56 <Psi|4ward> but i think Mozilla Network Security Services has nothing todo with NameServiceSwitch
2015-04-07 17:06:03 <_mjones> Yes, those are two different things with the same initialism!
2015-04-07 17:06:21 <_mjones> Mozilla NSS is a TLS/SSL/PKI implementation.
2015-04-07 17:07:08 <BitL0G1c> on debian:
2015-04-07 17:07:12 <BitL0G1c> stuart@deb64-vmware:~$ grep -r "NSS_STATUS" /usr/include
2015-04-07 17:07:13 <BitL0G1c> /usr/include/nss.h:  NSS_STATUS_TRYAGAIN = -2,
2015-04-07 17:07:13 <BitL0G1c> /usr/include/nss.h:  NSS_STATUS_UNAVAIL,
2015-04-07 17:07:13 <BitL0G1c> /usr/include/nss.h:  NSS_STATUS_NOTFOUND,
2015-04-07 17:07:13 <BitL0G1c> /usr/include/nss.h:  NSS_STATUS_SUCCESS,
2015-04-07 17:07:13 <BitL0G1c> /usr/include/nss.h:  NSS_STATUS_RETURN
2015-04-07 17:07:15 <_mjones> NameSwitchService is libc routines (or a discrete library) to abstract name lookups (DNS, uid, gid, but also NIS, NIS+, LDAP, others if supported)
2015-04-07 17:07:43 <_mjones> musl has simplified nss routines; for instance NIS is not supported.
2015-04-07 17:08:11 <dalias> not nss
2015-04-07 17:08:20 <_mjones> misnomer. ;)
2015-04-07 17:08:27 <dalias> nss is a sun/glibc implementation detail/backend
2015-04-07 17:08:31 <_mjones> "name lookup functions", correct?
2015-04-07 17:08:43 <dalias> something like that
2015-04-07 17:10:10 <_mjones> 
ACTION suspects dalias's irc client beeps whenever it regexp matches 'musl' or 'dalias'.

2015-04-07 17:11:24 <dalias> using an nss module for user db makes sense (tho ldap would probably be better if you need huge centrally managed user db, and if you don't, flat files are better)
2015-04-07 17:11:47 <dalias> but using it for _authentication_ purposes does not make sense
2015-04-07 17:12:06 <dalias> for that you'd probably want PAM
2015-04-07 17:14:07 <_mjones> Yes. Most 'AAA' implementations need both.
2015-04-07 17:14:19 <_mjones> especially on unix.
2015-04-07 17:15:14 <_mjones> today we have key-value stores that might be ideal. LDAP is nice and fairly standardised, but it can be cantankerous.
2015-04-07 17:15:36 <nsz> AAA?
2015-04-07 17:15:44 <_mjones> Long ago there were a few systems that used to use DNS for this. It workd well except for the security. I still love DNS as a key-value store.
2015-04-07 17:16:03 <_mjones> nsz: 'Authentication, Authorisation, and Accounting'. Term probably popularised by Cisco.
2015-04-07 17:16:13 <nsz> i see
2015-04-07 17:16:43 <BitL0G1c> Psi|4ward - http://arthurdejong.org/nss-pam-ldapd/ is in musl & is actively maintained
2015-04-07 17:16:45 <_mjones> In terms of cisco and similar gear, AAA protocols would be RADIUS and TACACS+ and similar (has anyone ever actually used DIAMETER?).
2015-04-07 17:17:31 <Psi|4ward> i need myql be :\
2015-04-07 17:17:38 <_mjones> iirc, the DNS-using systems were Telebit Trailblazers or something in that vein. Probably before your time. I'm not as old as I sound, I swear.
2015-04-07 17:18:54 <_mjones> nsz: for instance you can't use RADIUS as NSS on unix because RADIUS will only authenticate username/password pairs, not let you look up UIDs and GIDs forward and backward.\
2015-04-07 17:19:42 <_mjones> 
ACTION wonders if someone has hooked NSS into redis/memcached yet. Probably.

2015-04-07 17:20:34 <_mjones> Psi|4ward: probably more users are using NSS->LDAP->MySQL than NSS->MySQL straight.
2015-04-07 17:22:08 <_mjones> LDAP suffers from being OSI x.500. Most committee specifications suffer from conflicting goals and end up overly complicated with onerous minimum specifications.
2015-04-07 17:22:33 <Psi|4ward> possibly yes but ive everything running in ubuntu right now
2015-04-07 17:24:10 <_mjones> Psi|4ward: ldap is rather well supported in linux/unix. It's a generic directory service.
2015-04-07 17:26:53 <_mjones> Psi|4ward: I\m not saying I recommend adding an extra component, just that I'm making an educated guess that there are far more people doing nss->ldap->mysql than nss->mysql because that's a standardised approach and the tooling is strongly supported.
2015-04-07 17:28:15 <Psi|4ward> yea i got it, but ive an old installation wich uses libnss-mysql for years now and i only want to put this into a docker container
2015-04-07 17:28:29 <_ikke_> ncopa: What is your goal with the redmine docker instance? How do you want to verify everything is still working?
2015-04-07 17:28:56 <_mjones> psi|4ward: I think I missed that part of the conversation.
2015-04-07 17:29:01 <Psi|4ward> huh, im interested in redmine docker too!
2015-04-07 17:34:27 <clandmet1r> nice, nl infra is back online
2015-04-07 17:42:15 <Jean-Scotch> I for one use ldap everywhere I can...
2015-04-07 17:42:20 <Jean-Scotch> even with AL
2015-04-07 17:44:23 <Jean-Scotch> I still use a debian box for sftp access becaus of pam-ldap not so good on AL
2015-04-07 17:44:35 <Jean-Scotch> ;)
2015-04-07 17:47:19 <Jean-Scotch> And thank you clandmeter for the testing/nbd package ;)
2015-04-07 17:47:34 <Jean-Scotch> I will test it as soon as I receive my new lab machine
2015-04-07 17:48:01 <Jean-Scotch> I have 7 NBD servers in production
2015-04-07 17:48:41 <Jean-Scotch> Is tests are successull, I will migrate those 7 to AL ;)
2015-04-07 17:49:14 <clandmet1r> Jean-Scotch: i tried to get that other nbd pkg build, but its a pita.
2015-04-07 17:49:48 <Jean-Scotch> I spent also severall days trying :(
2015-04-07 17:50:59 <clandmet1r> it probably needs patching to support musl. probably thinking glibc == linux
2015-04-07 17:51:16 <Jean-Scotch> I will keep a few NBD server in debian for xNBD proxy feature. But most of the stock will soon run on full AL boxes
2015-04-07 17:52:59 <Jean-Scotch> I have read about consul package to be available soon ;) I will also be an avid tester ;)
2015-04-07 17:55:02 <clandmeter> yes, i think coredump was working on it.
2015-04-07 17:58:00 <clandmeter> i installed alpine on my home desktop. its working nicely.
2015-04-07 17:58:14 <clandmeter> i just need to get better control of power/fans
2015-04-07 17:58:47 <clandmeter> laptop as desktop that is. it sounds like its taking off.
2015-04-07 18:34:49 <Jean-Scotch> I still use AL only for headless servers
2015-04-07 18:35:13 <Jean-Scotch> but trying x2go ;)
2015-04-07 18:35:48 <Jean-Scotch> it's already working for sigle app
2015-04-07 18:36:02 <Jean-Scotch> I'm now trying to have full desktop
2015-04-07 18:36:17 <Jean-Scotch> with minimalistic approch
2015-04-07 18:36:56 <Jean-Scotch> my goal is to have lot of concurrent x2go sessions
2015-04-07 18:37:49 <Jean-Scotch> with full desktop capabilities but in a very thight and controlled environment
2015-04-07 18:38:29 <_mjones> Jean-Scotch: x2go client on Alpine desktops, _vice_versa_, or both?
2015-04-07 18:38:52 <Jean-Scotch> for now x2goserver on AL server
2015-04-07 18:39:18 <Jean-Scotch> client could be anything running x2goclient
2015-04-07 18:39:51 <Jean-Scotch> my customer are using windows desktop
2015-04-07 18:40:25 <Jean-Scotch> And the team is using linux desktop
2015-04-07 18:40:59 <Jean-Scotch> and laptops
2015-04-07 18:41:27 <Jean-Scotch> soon android clients on tablets...
2015-04-07 18:42:30 <Jean-Scotch> _mjones:ceph will be a must have ;)
2015-04-07 18:43:42 <Jean-Scotch> I was in the proces off deploying a ceph cluster debian based. I frezzed that project, lurking that topic ;)
2015-04-07 18:45:18 <Jean-Scotch> I have big hopes in ceph, consul, combined with PXE to managed my fleet very lightly ;)
2015-04-07 18:49:32 <Jean-Scotch> I think I will manage to have the full fleet in only one 32 gb usb stick including full stable and edge repos, and only plugged for a couple of seconds/minutes .
2015-04-07 18:50:47 <_mjones> Jean-Scotch: ncopa and I made progress on ceph. I haven't tried to finish yet.
2015-04-07 18:51:27 <_mjones> Jean-Scotch: I decided to do some easy aports before I went back to Ceph ;)
2015-04-07 18:52:06 <Jean-Scotch> I'm still struggling with mounting NFS share inside LXC
2015-04-07 18:53:21 <_mjones> Jean-Scotch: ceph was actually changing code that we had to port. pthreads stuff. Based on this I think the codebase is not stable enough for my taste. I wouldn't run ceph production on data that needed to stay complete and accurate.
2015-04-07 18:54:13 <Jean-Scotch> On the web sit, I think they recommand some cherry picked versions as stable...
2015-04-07 18:55:41 <_mjones> I was working on 0.87.1 and git HEAD, then ncopa and I got past all the problems I had solved and gotten to a new problem, and that problem was in his 0.93 but not in my 0.87.1 not my latest HEAD. :(
2015-04-07 18:55:43 <Jean-Scotch> I want to run ceph in parallel to the current production 7 SAN for 6 or 12 month
2015-04-07 18:56:40 <Jean-Scotch> I'm very conservative mith my customers data ;)
2015-04-07 18:58:35 <spew> is there a package that includes uxterm?
2015-04-07 18:59:18 <_mjones> spew: seems not. I don't see one in aports.
2015-04-07 18:59:36 <clandmeter> spew you can check pkgs.a.o
2015-04-07 19:01:42 <spew> where is pkgs.a.o?
2015-04-07 19:02:28 <clandmeter> http://pkgs.alpinelinux.org/contents
2015-04-07 19:04:38 <clandmeter> xbmc/kodi mess :|
2015-04-08 05:52:53 <_mjones> These fitlet micro fanless PCs have dramatically better specs and formfactor than anything else I've found, and the cost is insanely reasonable: http://www.fit-pc.com/web/products/fitlet/
2015-04-08 05:55:13 <_mjones> 5yr warranty, nice AMD processor, intel giganics and wlan. I could use 4k on the digital video, but still. Not sure they're making any profit on these.
2015-04-08 06:04:48 <__number5__> _mjones: why not just go with NUC?
2015-04-08 06:25:49 <_mjones> NUCs are awfully nice, it's true. But the i5 and i7 models aren't incredibly cheap. They're also not fanless. Intel has a fanless NUC but it's closely designed for thin-client computing (ECC disabled on board, 8gb ram, one nic, one core one-thread E3810 cpu.
2015-04-08 06:26:22 <_mjones> That one most likely does 4K though, and the higher-end NUCs definitely do. I need to find out if the i3 broadwell nuc does.
2015-04-08 06:28:56 <_mjones> I probably need to get one NUC for now and give up on ecc ram and open firmware for the moment. This all started because I was trying to find a cheap server for a spare 9mm SSD I have, and it would be nice if that machine had 2xUSB3 and 4K digital out also.
2015-04-08 06:29:33 <_mjones> I can do all that but it requires a lga 1150 socket server/workstation board and an i3 and ecc ram, btw.
2015-04-08 06:35:32 <_mjones> I'm being silly, really. The ssd is a 32gb model worth a fraction of my last bar tab. I decided tonight I'll probably use it to alt-boot freebsd 10.1 from the esata port on my t420 and use the internal msata for linux.
2015-04-08 07:17:09 <ScrumpyJack> morning
2015-04-08 08:25:54 <systmkor> ScrumpyJack: evening :D
2015-04-08 08:33:26 <_mjones> ScrumpyJack: bedtime...
2015-04-08 08:33:34 <ScrumpyJack> :(
2015-04-08 08:33:52 <ScrumpyJack> no 32bit mplayer
2015-04-08 08:34:57 <_mjones> arch="x86_64"
2015-04-08 08:35:04 <ScrumpyJack> yup
2015-04-08 08:35:17 <_mjones> Hmm, no comments. I was looking at this earlier and I'm guessing it's probably just an oversight.
2015-04-08 08:35:44 <_mjones> let's see before I go to bed. Shouldn't take a minute.
2015-04-08 08:36:34 <_mjones> no, but it won't compile on x86-64 will it?
2015-04-08 08:36:40 <_mjones> You try.
2015-04-08 08:36:56 <_mjones> (56/90) Installing libxshmfence (1.2-r0) (57/90) Installing libxshmfence-dev (1.2-r0) Cough.
2015-04-08 08:37:18 <_mjones> I'm a big mplayer fan too.
2015-04-08 08:37:29 <ScrumpyJack> why wont it compile on x86_64?
2015-04-08 08:37:55 <_mjones> I mean it won't crosscompile x86 on my x86-64s
2015-04-08 08:38:19 <_mjones> whoa, lots of warnings.
2015-04-08 08:38:27 <_mjones> Too bad those mplayer2 blokes never got anywhere.
2015-04-08 08:40:16 <_mjones> I'd need a x86 vm to test the package build.
2015-04-08 08:40:32 <ScrumpyJack> setting one up no
2015-04-08 08:40:34 <ScrumpyJack> now
2015-04-08 08:43:39 <ScrumpyJack> 123 build dependancies :(
2015-04-08 08:44:17 <ScrumpyJack> arg, abuild doesn't work behind a PROXY
2015-04-08 08:45:08 <clandmeter> make[1]: Leaving directory '/home/clandmeter/aports/main/mplayer/src/MPlayer-1.1.1/ffmpeg'
2015-04-08 08:45:11 <clandmeter> Makefile:787: recipe for target 'ffmpeg/libavcodec/libavcodec.a' failed
2015-04-08 08:45:13 <clandmeter> make: *** [ffmpeg/libavcodec/libavcodec.a] Error 2
2015-04-08 08:47:12 <ScrumpyJack> do we have Realtek ALC272 Audio support?
2015-04-08 08:47:20 <_mjones> Our slogan: " The only shell that should get started during boot is gnome-shell! " -- Yes, the slogan needs a bit of work, but you get the idea.
2015-04-08 08:47:37 <_mjones> ScrumpyJack: uh, that would be a kernel question.
2015-04-08 08:47:52 <_mjones> Realteks are pretty popular, but I don't know models. I'd guess yes.
2015-04-08 08:48:37 <_mjones> My build from aports failed :(
2015-04-08 08:49:10 <ScrumpyJack> it'll be in the kernel.config
2015-04-08 08:49:28 <_mjones> libavcodec/rv34.c:723:35: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]         (unsigned)(src_x - !!lx*2) > s->h_edge_pos - !!lx*2 - (width <<3) - 4 ||
2015-04-08 08:50:10 <ScrumpyJack> CONFIG_SND_HDA_CODEC_REALTEK=m
2015-04-08 08:51:29 <_mjones> loadable kernel module.
2015-04-08 08:52:47 <_mjones> ./pci/hda/snd-hda-codec-realtek.ko in default kernel.
2015-04-08 08:53:52 <_mjones> About 7% of all LKM storage space is consumed by sound drivers. Interesting.
2015-04-08 08:56:54 <ScrumpyJack> no sound on my dell mini 10
2015-04-08 09:27:12 <ScrumpyJack> got it
2015-04-08 09:27:16 <ScrumpyJack> thanks alsamixer
2015-04-08 09:27:26 <ScrumpyJack> bloody mute was on
2015-04-08 14:07:36 <Orochimarufan> Is there no pulseaudio package on Alpine?
2015-04-08 14:11:27 <nsz> there is this http://git.alpinelinux.org/cgit/aports/plain/unmaintained/pulseaudio/
2015-04-08 14:11:31 <nsz> in /unmaintained/
2015-04-08 14:13:29 <Orochimarufan> how would I go about using that?
2015-04-08 14:14:55 <nsz> using abuild, but since it's not maintained it most likely wouldnt work
2015-04-08 14:15:25 <Orochimarufan> Well, I'll just have to give it a shot
2015-04-08 14:31:35 <__number5__> quit
2015-04-08 19:47:09 <Orochimarufan> does musl have differing function sigs in some cases?
2015-04-08 19:50:14 <dalias> there are a couple glibc functions that have the same name as standard functions but with wrong signature/behavior. glibc remaps these to other names in standards-conforming profiles
2015-04-08 19:50:26 <dalias> musl does not offer functions that clash with the standard ones
2015-04-08 19:50:46 <Orochimarufan> hmm
2015-04-08 19:51:32 <Orochimarufan> it seems to be the case with ioctl stat and fopen >.<
2015-04-08 19:52:06 <dalias> i think glibc has a wrong signature for ioctl just as a historic mistake
2015-04-08 19:52:16 <dalias> it's not incompatible semantically, just a minor detail
2015-04-08 19:52:25 <dalias> for stat and fopen i don't see what the difference could be
2015-04-08 19:52:29 <dalias> they should not mismatch
2015-04-08 19:55:49 <lrvick> Is there a way to get a 'useradd' command in alpine?
2015-04-08 19:55:57 <Orochimarufan> use adduser
2015-04-08 19:56:44 <_mjones> lrvick: /usr/sbin/adduser -> /bin/busybox
2015-04-08 19:56:58 <lrvick> ah. thank
2015-04-08 19:57:00 <Orochimarufan> Hrm.. never mind, it's just a redefinition, not type mismatch
2015-04-08 19:57:16 <_mjones> lrvick: there's no useradd because 'adduser' is implemented as part of busybox.
2015-04-08 19:57:19 <Orochimarufan> only ioctl has a mismatch
2015-04-08 19:57:46 <Orochimarufan> lrvick: Watch out thoug, the usage is slightly different
2015-04-08 19:59:13 <clandmeter> anybody know why udev and dbus wont start at boot? it was working fine when running from tmpfs, but fresh hd install not.
2015-04-08 20:01:54 <lrvick> I want to create a user with no pasword, that I can still log in with via ssh keys.
2015-04-08 20:02:04 <lrvick> Normally i would just set a random password with /dev/random and uuencode
2015-04-08 20:02:29 <lrvick> uuencode does not seem to be a thing
2015-04-08 20:02:36 <lrvick> is there a more sane way to do this?
2015-04-08 20:03:16 <_mjones> @clandmeter: stupid question, but do you have them installed? udev is not in base, and d-bus must not be either.
2015-04-08 20:03:53 <_mjones> lrvick: not sure if sshd has a parameter to allow one to login to locked accounts\
2015-04-08 20:04:03 <_mjones> lrvick: using openssh or dropbear?
2015-04-08 20:04:12 <lrvick> openssh
2015-04-08 20:04:37 <lrvick> or is there a way to unlock an account without a password somehow?
2015-04-08 20:04:45 <lrvick> 
ACTION has been unable to find a solution to this

2015-04-08 20:04:57 <_mjones> lrvick: anyway, sha1sum or md5sum would seem to be a drop-in replacement for uuencode for this.
2015-04-08 20:05:41 <lrvick> hmm. well i guess i can continue with a variation of that hack then
2015-04-08 20:06:12 <_mjones> exactly how do you do it that every password doesn't end up as "begin 644 /dev/random" ?
2015-04-08 20:06:25 <clandmeter> _mjones: they are installed and running fine if i start them manual
2015-04-08 20:06:33 <_mjones> also, always use /dev/urandom on Linux and every platform that supports it.
2015-04-08 20:06:59 <lrvick> _mjones: usermod -p `dd if=/dev/urandom bs=1 count=30 | uuencode -m - | head -2 | tail -1` git
2015-04-08 20:07:06 <_mjones> otherwise you're likel to overspend your ration of random and have problems.
2015-04-08 20:08:01 <BitL0G1c> haveged is good for making sure there is plenty of entropy
2015-04-08 20:08:45 <lrvick> In this case, I won't even know what the password is. I just want password login disabled and ssh key only
2015-04-08 20:08:49 <lrvick> its for a git server
2015-04-08 20:09:02 <_mjones> lrvick: sha1sum is basically a drop-in replacement for 'uuencode -m'.
2015-04-08 20:09:42 <_mjones> @clandmeter: then blocking on something that isn't yet started but not known as a dependency, or somehow relying on a terminal.
2015-04-08 20:10:36 <_mjones> @clandmeter: hmm, the udev on my devbox isn't starting automatically either. :( I didn't notice because it's only there to build ceph.
2015-04-08 20:12:38 <_mjones> BitL0G1c: that one's new to me. But I think I'd prefer a hardware prng over having another daemon around. Too many daemons are typically running even without Lennart's efforts.
2015-04-08 20:13:08 <_mjones> I know inetd's out of style and daemons are good and obviate much suid/guid, but still.
2015-04-08 20:13:34 <BitL0G1c> haveged is useful to run if you are creating LUKS devices
2015-04-08 20:13:37 <lrvick> since usermod does not exist
2015-04-08 20:13:41 <lrvick> I need some equivilant to:
2015-04-08 20:13:44 <lrvick> usermod -p `dd if=/dev/urandom bs=1 count=30 | sha1sum` git
2015-04-08 20:15:40 <BitL0G1c> usermod is in shadow but may be broken in musl
2015-04-08 20:17:28 <_mjones> lrvick: 'passwd' would generally be what you want, but implementations are so damn eager to prevent anyone from putting the password in $@ (argv[]) that this might be unnecessarily difficult.
2015-04-08 20:18:26 <lrvick> Yeah I need to be able to script it.
2015-04-08 20:18:45 <_mjones> BitL0G1c: have you seen Ubuntu's cloud PRNG 'pollen' and 'pollinate'? I thought it was a jest at first, but read about it and it's pretty slick.
2015-04-08 20:18:48 <lrvick> Does alpine use pretty standard /etc/shadow lines?
2015-04-08 20:18:57 <_mjones> lrvick: yes
2015-04-08 20:19:00 <lrvick> I could just generate one...
2015-04-08 20:19:03 <lrvick> heh
2015-04-08 20:19:07 <_mjones> until we get TCB ;)
2015-04-08 20:19:34 <nsz> od -An -N20 -v -w20 -tx1 /dev/urandom | tr -d ' '
2015-04-08 20:19:36 <BitL0G1c> _mjones - no new to me - thx :)
2015-04-08 20:19:45 <nsz> ..is just as good as using sha1sum
2015-04-08 20:20:38 <Orochimarufan> http://paste.ee/p/vLJ5x Any Idea what's going wrong?
2015-04-08 20:20:45 <_mjones> nsz: then, least-bad way to script it as a password change with Alpine base install?
2015-04-08 20:21:25 <nsz> (od+tr are standard tools available on every posix platform)
2015-04-08 20:21:46 <nsz> i dont know how to change passwords on alpine
2015-04-08 20:21:55 <_mjones> ;)
2015-04-08 20:21:59 <nsz> i set trivial passwords on my laptop
2015-04-08 20:22:27 <nsz> the only difficult passwd is the disk encryption pass wd and the ssh key passwd
2015-04-08 20:22:36 <nsz> but those are not managed by alpine
2015-04-08 20:24:59 <_mjones> lrvick: I'm trying to think of the best way to do it. Writing directly to /etc/shadow is a last resort.
2015-04-08 20:25:19 <_mjones> ah mkpasswd should do it, if you use -P
2015-04-08 20:25:49 <Orochimarufan> Why does it claim that stat64 is a redefinition of stat >.<
2015-04-08 20:25:54 <_mjones> lrvick: try to generate the passwd then pipe it to "mkpasswd git --password-fd=1"
2015-04-08 20:26:52 <Orochimarufan> I guess this would be why:  99 #define stat64 stat
2015-04-08 20:27:29 <_mjones> lrvick: oops, I mean "mkpasswd git --password-fd=0"
2015-04-08 20:27:41 <_mjones> lrvick: this would seem to be portable to other Linux, at least.
2015-04-08 20:27:50 <nsz> yes stat64 is a macro on musl
2015-04-08 20:28:13 <Orochimarufan> which means I won't get around modifying the source >.<
2015-04-08 20:28:43 <_mjones> Orochimarufan: not necessarily.
2015-04-08 20:28:44 <nsz> ah you are trying to build the horible pulseaudio oss dsp emulation layer
2015-04-08 20:28:48 <nsz> that cannot possibly work
2015-04-08 20:28:55 <Orochimarufan> why not?
2015-04-08 20:29:03 <nsz> it is broken even on glibc
2015-04-08 20:29:06 <_mjones> way to be nonjudgemental, nsz.
2015-04-08 20:29:09 <Orochimarufan> (Not that I particularly need it)
2015-04-08 20:29:09 <_mjones> ;)
2015-04-08 20:29:19 <nsz> it's an ldpreload hack
2015-04-08 20:29:34 <Orochimarufan> question is, can I disable it
2015-04-08 20:29:36 <lrvick> mkpasswd just generates the hash, so i still need to generate a /etc/shadow line
2015-04-08 20:29:36 <nsz> that replaces a lot of libc apis to hijack access to /dev/dsp
2015-04-08 20:29:40 <lrvick> yeah?
2015-04-08 20:30:11 <nsz> they replace as-signalsafe functions with non-as-safe ones.. and introduce other brokenness
2015-04-08 20:30:28 <Orochimarufan> I guess this is it: --disable-oss-wrapper   Disable optional OSS wrapper support
2015-04-08 20:30:44 <nsz> it should be possible to make it compile though
2015-04-08 20:30:52 <nsz> not sure where it is used
2015-04-08 20:30:59 <nsz> but it is mostly broken
2015-04-08 20:31:17 <_mjones> lrvic: oh, damn. You're right.
2015-04-08 20:31:17 <lrvick> echo git:$(dd if=/dev/urandom bs=1 count=30 | sha1sum | mkpasswd -P0):16533:0:99999:7:::
2015-04-08 20:31:21 <lrvick> lulz
2015-04-08 20:32:08 <Orochimarufan> let's try this...
2015-04-08 20:32:38 <_mjones> let me google. Forcing people to keep passwords off of argv[] can't possibly be better than directly appending to /etc/shadow. :P
2015-04-08 20:33:40 <nsz> lrvick: dont use sha1sum on /dev/urandom
2015-04-08 20:33:47 <nsz> that makes no sense
2015-04-08 20:34:02 <nsz> and breaks when sha1sum is not available for some reason
2015-04-08 20:34:59 <nsz> and dd bs=1 count=30 will make 30 read syscalls..
2015-04-08 20:35:14 <nsz> ..and 30 write ones
2015-04-08 20:35:24 <_mjones> lrvick: ah, I missed chpasswd. That will do it from stdin.
2015-04-08 20:35:34 <Orochimarufan> it built :D
2015-04-08 20:35:35 <_mjones> bs=30c
2015-04-08 20:35:45 <Orochimarufan> now I just need to get my APKBUILD fixed
2015-04-08 20:35:58 <_mjones> lrvick: nsz's pipeline is more portable; I'd use that
2015-04-08 20:36:34 <lrvick> sed -i '$ d' /etc/shadow && echo git:$(od -An -N20 -v -w20 -tx1 /dev/urandom | tr -d ' '| mkpasswd -P0):16533:0:99999:7:::" >> /etc/shadow
2015-04-08 20:36:41 <lrvick> There -must- be a better way to do this
2015-04-08 20:36:57 <_mjones> nsz: I suggested sha1sum as a drop-in replacement for uuencode; of course the hashing is unnecessary.
2015-04-08 20:36:57 <nsz> actually you dont need 20bytes about 10 should be enough :)
2015-04-08 20:37:06 <Orochimarufan> let me gues... {} expansion is a bash-ism D:
2015-04-08 20:37:11 <nsz> (but that does not matter)
2015-04-08 20:37:45 <_mjones> I think (( )) isn't.
2015-04-08 20:37:55 <Orochimarufan> how's that work?
2015-04-08 20:38:01 <_mjones> Never use bash for scripting and your scripts will almost always be portale, grasshopper.
2015-04-08 20:38:02 <nsz> what is {} expansion?
2015-04-08 20:38:11 <nsz> ${foo} works in posix sh
2015-04-08 20:38:15 <Orochimarufan> mv {src,dst}/filename
2015-04-08 20:38:17 <Orochimarufan> for example
2015-04-08 20:38:18 <nsz> ah
2015-04-08 20:38:27 <nsz> that's non-posix
2015-04-08 20:38:34 <_mjones> lrvick: chpasswd will change it from stdin.
2015-04-08 20:38:34 <nsz> or hm
2015-04-08 20:38:50 <Orochimarufan> 'twould've been too convenient^^
2015-04-08 20:40:05 <lrvick> `od -An -N20 -v -w20 -tx1 /dev/urandom | tr -d ' ' | chpasswd git` ?
2015-04-08 20:40:12 <lrvick> 
ACTION tries

2015-04-08 20:40:20 <Orochimarufan> can I just have the APKBUILD run in bash?
2015-04-08 20:40:40 <nsz> lrvick: od writes a \n too
2015-04-08 20:41:10 <nsz> although chpasswd should handle that
2015-04-08 20:41:30 <lrvick> chpasswd: missing new password
2015-04-08 20:41:46 <lrvick> sure busybox chpasswd supports that?
2015-04-08 20:41:51 <_mjones> Orochimarufan: ffs, do it right.
2015-04-08 20:42:08 <Orochimarufan> but I'm lazy >.<
2015-04-08 20:42:29 <_mjones> That kind of thing is why it was a stupid idea to make /bin/sh bash by default and why it was so hard to fix.
2015-04-08 20:43:03 <_mjones>  ash (Almquist) predates the Linux kernel by years, but misguided distro-makers wouldn't use it, no.
2015-04-08 20:43:04 <Orochimarufan> /bin/sh should not have any nonstandard features enabled, no matter what it's symlinked to...
2015-04-08 20:43:23 <_mjones> That too.
2015-04-08 20:43:30 <lrvick> echo "git:$(od -An -N20 -v -w20 -tx1 /dev/urandom | tr -d ' ')" | chpasswd
2015-04-08 20:43:32 <lrvick> nailed it
2015-04-08 20:43:32 <Orochimarufan> well, debian uses dash (Debian Almquist SHell)
2015-04-08 20:43:43 <_mjones> but bash is much, much heavier footprint than ash or dash.
2015-04-08 20:43:50 <Orochimarufan> true
2015-04-08 20:44:02 <_mjones> dash is modified ash. Added command-line editing I guess, probably more.
2015-04-08 20:44:26 <Orochimarufan> I'm not saying bash should be the default
2015-04-08 20:44:47 <_mjones> that bash nonsense is one of several contraindications I got when I first tried Linux, and didn't run it for 10 more years.
2015-04-08 20:45:50 <nsz> mv src/filename dst/filename does not seem that bad to me compared to mv {src,dst}/filename
2015-04-08 20:46:04 <_mjones> mv -nv, yes
2015-04-08 20:46:17 <_mjones> there are harder bashisms to fix, no?
2015-04-08 20:47:03 <nsz> i thought it's a short script that Orochimarufan is writing
2015-04-08 20:47:04 <_mjones> lrvick: excellent. Notice that scarcely resembles the string with which you started....
2015-04-08 20:47:06 <Orochimarufan> nsz: when filename is /usr/share/man/man1/<file> that's a different story tho
2015-04-08 20:47:57 <nsz> Orochimarufan: no it's not, you just assign a variable and use $FOO
2015-04-08 20:48:08 <Orochimarufan> right...
2015-04-08 20:48:19 <Orochimarufan> or I just make a function that does what I want
2015-04-08 20:48:28 <_mjones> Orochimarufan: show us the portion of the script.
2015-04-08 20:48:54 <Orochimarufan> it's the APKBUILD subpackage split function, what more do you need to know?
2015-04-08 20:48:55 <_mjones> for-in-do loop with basename even.
2015-04-08 20:49:29 <_mjones> that function isn't already in bash. I thought you were porting an existing script.
2015-04-08 20:49:54 <Orochimarufan> I kinda "ported" this: https://projects.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/pulseaudio#n80
2015-04-08 20:50:13 <Orochimarufan> not 100% but I took it as a template
2015-04-08 20:50:49 <_mjones> fair enough. Arch isn\t using ash?
2015-04-08 20:51:03 <Orochimarufan> nah, they use bash
2015-04-08 20:51:19 <lrvick> _mjones: indeed. Thanks for the help guys :)
2015-04-08 20:51:50 <Orochimarufan> Arch is quite a lot bigger than Alpine in it's minimal setup
2015-04-08 20:52:21 <Orochimarufan> even if it's still tiny compared to most others (I'm sure you know what I'm talking about^^)
2015-04-08 20:52:26 <_mjones> I know they use systemd. That aside, one of their tenets is correctness, and another is simplicity. I assumed they followed those with resp[ect to /bin/sh.
2015-04-08 20:53:01 <_mjones> Now I'm wondering if some whinger has justified Bash because anything else is complicated and not compatible with a 1999 edition of Red Hat. :P
2015-04-08 20:53:40 <Orochimarufan> well, they probably just never saw the need to switch to *ash
2015-04-08 20:53:48 <Orochimarufan> $ ls -l /bin/sh
2015-04-08 20:53:50 <Orochimarufan> lrwxrwxrwx 1 root root 4 Dec 30 22:08 /bin/sh -> bash
2015-04-08 20:53:58 <_mjones> so you don't know how to port those or you just looking for another way?
2015-04-08 20:54:13 <Orochimarufan> I'm just being lazy
2015-04-08 20:54:18 <_mjones> Because it's longer characters but extremely straightforward.
2015-04-08 20:54:25 <Orochimarufan> I know
2015-04-08 20:54:58 <_mjones> one needs to look at it not as a switch to ash, but as a correct reversion to Bourne (clone).
2015-04-08 20:55:07 <_mjones> bash was the switch.
2015-04-08 20:56:19 <Orochimarufan> well, a switch is, by definition, an alteration of some status quo, which going (back) to ash would qualify as
2015-04-08 20:57:11 <_mjones> I should see if I can find any documentation on the original decision. For the longest time I thought it was because bash was the only encumbrance-free Bourne available at the time, but I've since learned better.
2015-04-08 20:57:53 <Orochimarufan> it was probably because "everyone does it that way" or something like that
2015-04-08 20:58:01 <_mjones> And it doesn't explain why not use only bourne functionality if [argv[0]=="sh"
2015-04-08 20:58:18 <Orochimarufan> because GNU?
2015-04-08 20:58:20 <_mjones> But nobody did it that way.
2015-04-08 20:58:27 <Orochimarufan> zsh does it that way
2015-04-08 20:59:53 <_mjones> No I mean using bash as /bin/sh, not argv[0]
2015-04-08 20:59:59 <_mjones> I think you're right though.
2015-04-08 21:00:14 <Orochimarufan> well symlinking bash to /bin/sh
2015-04-08 21:00:16 <_mjones> ash is BSD licensed I'm pretty sure. So 'because GNU' is probably correct.
2015-04-08 21:00:50 <Orochimarufan> GNU doesn't seem to be too concerned about POSIX in general from what I gather
2015-04-08 21:01:25 <Orochimarufan> (Not less than systemd of course, but then that'd be quite the feat to accomplish, wouldn't it? :P)
2015-04-08 21:01:41 <_mjones> the original ash license is basically BSD 2-clause, so yeah.
2015-04-08 21:01:54 <_mjones> rms, you jerk.
2015-04-08 21:11:26 <Orochimarufan> all right, lets try this again
2015-04-08 21:18:45 <Orochimarufan> I must be particularly stupid today...
2015-04-08 21:25:04 <Orochimarufan> >>> ERROR: pulseaudio*: Libtool archives (*.la) files found and $options has no 'libtool' flag:
2015-04-08 21:25:14 <Orochimarufan> what does the "libtool" flag do?
2015-04-08 21:26:58 <BitL0G1c> there is a pulseaudio package in unmaintained - use that APKBUILD
2015-04-08 21:27:04 <Orochimarufan> I did
2015-04-08 21:27:53 <_mjones> libtool helps build libraries. I'm not sure if it's functionality isn't duplicated elsewhere.
2015-04-08 21:28:22 <Orochimarufan> is there something like https://www.archlinux.org/pacman/PKGBUILD.5.html#_options_and_directives for APKBUILDs?
2015-04-08 21:28:24 <BitL0G1c> I thought 'rm -f "$pkgdir"/usr/lib/*.la' fixed the libtool message
2015-04-08 21:28:53 <Orochimarufan> except /usr/lib/pulseaudio/libpulsecommon-6.0.la
2015-04-08 21:30:02 <vkrishn> `apk update` with /testing in repo,
2015-04-08 21:30:20 <vkrishn> `apk add newsbeuter`
2015-04-08 21:30:59 <vkrishn> create a /tmp/urls file with entry,
2015-04-08 21:30:59 <vkrishn> http://pear.php.net/feeds/latest.rss
2015-04-08 21:31:09 <vkrishn> newsbeuter -u /tmp/urls
2015-04-08 21:31:21 <vkrishn> apk add strace
2015-04-08 21:31:28 <vkrishn> strace newsbeuter -u /tmp/urls
2015-04-08 21:37:23 <Orochimarufan> mv: can't rename '/home/build/aports/unmaintained/pulseaudio/pkg/pulseaudio/usr/lib/cmake': Directory not empty
2015-04-08 21:37:25 <Orochimarufan> wtf
2015-04-08 21:38:42 <Orochimarufan> Ah I think I know what it's complaining about
2015-04-08 21:40:03 <vkrishn> and also....
2015-04-08 21:40:05 <vkrishn> newsbeuter -u /tmp/urls -x reload
2015-04-08 21:40:05 <vkrishn> terminate called after throwing an instance of 'std::logic_error'
2015-04-08 21:40:05 <vkrishn>   what():  basic_string::_S_construct null not valid
2015-04-08 21:40:05 <vkrishn> Aborted
2015-04-08 21:40:30 <vkrishn> ...though... reload goes ok
2015-04-08 21:41:39 <Orochimarufan> Anyway, I have a working PA 6.0 APKBUILD :D
2015-04-08 21:41:40 <BitL0G1c> Orochimarufan - it builds for me - but will update the APKBUILD to 6.0 - this is the APKBUILD I just used http://hastebin.com/isucuboliv.parser3 successfully
2015-04-08 21:43:22 <Orochimarufan> are you on 3.x? cause that can't have built with musl (it's building the dsp wrapper which is broken with musl)
2015-04-08 21:44:29 <Orochimarufan> not to mention that it doesn't do the pulseaudio-dev
2015-04-08 21:47:22 <BitL0G1c> I'm using an LXC container with edge repos to build
2015-04-08 21:47:48 <Orochimarufan> I'm in a 3.1.3 LXC container^^
2015-04-08 21:47:59 <Orochimarufan> http://hastebin.com/ubolojorax.coffee that's mine
2015-04-08 21:49:06 <Orochimarufan> It's not perfect but it's not too bad IMO
2015-04-08 21:50:00 <Orochimarufan> Also, who thought up a pastebin replacement without a monospace font >.<
2015-04-08 21:53:28 <BitL0G1c> I also removed "$pkgdir"/usr/lib/*/*.la
2015-04-08 21:53:38 <pcarrier> Orochimarufan: looks monospaced to me
2015-04-08 21:53:57 <Orochimarufan> doesn't seem to be
2015-04-08 21:54:06 <pcarrier> Orochimarufan: maybe they use webfonts and you don't support them?
2015-04-08 21:54:15 <Orochimarufan> maybe
2015-04-08 21:54:21 <pcarrier> cause it's monospaced here :)
2015-04-08 21:54:47 <Orochimarufan> this is a fresh system, I haven't done much with it yet
2015-04-08 21:56:13 <Orochimarufan> right now I need to figure out why my container template stopped working >.<
2015-04-08 21:56:53 <BitL0G1c> speexdsp-dev also needs to be added to depends_dev
2015-04-08 21:57:27 <pcarrier> and I need to package openconnect ;)
2015-04-08 21:57:38 <BitL0G1c> package() needs rm -f "$pkgdir"/usr/lib/*/*.la
2015-04-08 21:57:41 <_mjones> Orochimarufan: I'd generally recommend Edge if you're working with aports.
2015-04-08 21:58:16 <Orochimarufan> well, I'm building it to be used in a different container which won't be running edge
2015-04-08 22:00:49 <BitL0G1c> 6.0 builds for me
2015-04-08 22:03:15 <BitL0G1c> instead of find in package() just use: rm -f "$pkgdir"/usr/lib/*.la \
2015-04-08 22:03:15 <BitL0G1c>                 "$pkgdir"/usr/lib/*/*.la
2015-04-08 22:03:52 <Orochimarufan> why?
2015-04-08 22:06:59 <BitL0G1c> to remove the libtool error you are seeing at the end
2015-04-08 22:06:59 <Orochimarufan> finally, the template works again >.<
2015-04-08 22:07:10 <Orochimarufan> the find works just fine
2015-04-08 22:07:34 <BitL0G1c> you need to find recursively then
2015-04-08 22:07:44 <Orochimarufan> indeed that's what it does
2015-04-08 22:11:34 <Orochimarufan> is speexdsp only in edge or something? :O
2015-04-08 22:21:55 <_mjones> I don't feel like firing up 3.1.3 to check. oh, alright.
2015-04-08 22:23:05 <Orochimarufan> not that it matters particularly much
2015-04-08 22:23:19 <Orochimarufan> though... why isn't init.d/networking working >.<
2015-04-08 22:25:11 <BitL0G1c> by default networking is not started
2015-04-08 22:25:22 <_mjones> Yeah, my 3.1.3 vm never has networking when it comes up.
2015-04-08 22:25:47 <Orochimarufan> well yeah but /etc/init.d/networking start doesn't work
2015-04-08 22:26:06 <BitL0G1c> rc-service networking start
2015-04-08 22:26:20 <Orochimarufan> same result
2015-04-08 22:26:34 <Orochimarufan> wait...
2015-04-08 22:26:42 <BitL0G1c> or sshd (& networking will start)
2015-04-08 22:27:31 <_mjones> service networking start
2015-04-08 22:27:45 <Orochimarufan> ifup <device> doesn't even work
2015-04-08 22:27:46 <_mjones> anyway, seems speexdsp is indeed only in edge, although it's in main.
2015-04-08 22:28:07 <Orochimarufan> though setting it up manually with iproute2 works just fine
2015-04-08 22:28:36 <BitL0G1c> /sbin/setup-interfaces
2015-04-08 22:28:47 <doughdemon> vkrishn: the problem lies in newsbeuter's use of //TRANSLIT for iconv. musl iconv does not support transliteration.
2015-04-08 22:30:36 <Orochimarufan> BitL0G1c: what's that do?
2015-04-08 22:30:50 <BitL0G1c> sets up eth0
2015-04-08 22:30:59 <BitL0G1c> so you can ifup
2015-04-08 22:31:59 <Orochimarufan> shouldn't setting up /etc/network/interfaces suffice?
2015-04-08 22:33:57 <BitL0G1c> yes
2015-04-08 22:34:07 <Orochimarufan> It even errors with just lo
2015-04-08 22:34:29 <Orochimarufan> how do I debug this?
2015-04-08 22:37:58 <BitL0G1c> you're in a docker container ?
2015-04-08 22:38:04 <Orochimarufan> No, LXC
2015-04-08 22:40:22 <BitL0G1c> here's the container config i use http://hastebin.com/desepibora.avrasm
2015-04-08 22:41:09 <BitL0G1c> with a libvirt bridge
2015-04-08 22:41:55 <BitL0G1c> for running services inside LXC openvswitch is less hassle
2015-04-08 22:42:40 <Orochimarufan> well, I'd like to set it up from inside the container
2015-04-08 22:43:03 <Orochimarufan> I might just write up a small runscript calling the appropriate iproute2 tools
2015-04-08 22:43:15 <Orochimarufan> (or rather their busybox equivalents)
2015-04-08 22:53:00 <Orochimarufan> how do I handle a non-forking service in openrc?
2015-04-08 22:59:34 <_mjones> Find another one and look through aports for its openrc script ;)
2015-04-08 22:59:49 <_mjones> there are a couple options, but I'd look for examplars, myself.
2015-04-08 23:00:02 <Orochimarufan> I'm not aware of any other nonforking services off the top of my head
2015-04-08 23:01:53 <Orochimarufan> I guess start-stop-daemon -x <thing> -b -m ?
2015-04-08 23:05:22 <_mjones> looks likely.
2015-04-08 23:05:44 <_mjones> -b is the option of which I was thinking.
2015-04-08 23:06:27 <Orochimarufan> how does it handle interpreted services? do I need to -x the interpreter or the script?
2015-04-08 23:07:49 <_mjones> the script
2015-04-08 23:08:01 <_mjones> anything with #! should exec correctly
2015-04-08 23:08:17 <_mjones> separate-interpreter syntax is ugly
2015-04-08 23:08:41 <Orochimarufan> I guess I'll be symlinking python -> python3 then :P
2015-04-08 23:10:28 <systmkor> has anyone recently installed MATE on Alpine?
2015-04-08 23:11:16 <BitL0G1c> there are recent instructions on the wiki
2015-04-08 23:11:45 <BitL0G1c> they helped me get a bit further installing xfce (adding dbus / consolekit)
2015-04-08 23:19:26 <Orochimarufan> ... it seems the testing python3 is broken
2015-04-08 23:19:43 <Orochimarufan> ImportError: No module named '_socket'
2015-04-08 23:21:05 <Orochimarufan> Getting my stuff running on alpine is turning out harder than anticipated >.<
2015-04-08 23:23:11 <_mjones> but is it making your code better yet?
2015-04-08 23:23:20 <Orochimarufan> as in?
2015-04-08 23:23:45 <_mjones> in getting your stuff working, have you fixed any bugs (including portability) in your code, yet?
2015-04-08 23:23:57 <Orochimarufan> no (not yet)
2015-04-08 23:24:06 <_mjones> I assumed when you said 'my stuff' that it included some app or novel code.
2015-04-08 23:24:13 <Orochimarufan> it does
2015-04-08 23:24:21 <Orochimarufan> but it not working isn't my fault
2015-04-08 23:24:45 <Orochimarufan> I can't even start fixing my own bugs when the interpreter doesn't even work :P
2015-04-08 23:25:16 <_mjones> Orochimarufan: ok, that one must not be a bug in your code.
2015-04-08 23:26:26 <Orochimarufan> Shibuya:/usr/local/lib/cdev# python
2015-04-08 23:26:28 <Orochimarufan> Python 3.4.3 (default, Apr  8 2015, 18:44:37)
2015-04-08 23:26:30 <Orochimarufan> >>> import socket
2015-04-08 23:26:32 <Orochimarufan> ImportError: No module named '_socket'
2015-04-08 23:27:44 <Orochimarufan> python2 seems to work
2015-04-08 23:29:35 <systmkor> BitL0G1c: yah, starting with xfce at moment for a friend
2015-04-08 23:30:08 <systmkor> BitL0G1c: however even though it's a ThinkPad like mine(however a newer model) when we startup XFCE the keyboard and mouse don't work
2015-04-08 23:30:22 <systmkor> we do have xf86-input-[mouse,keyboard] installed
2015-04-08 23:30:28 <systmkor> and evdev
2015-04-08 23:30:33 <systmkor> BitL0G1c: any suggestions?
2015-04-08 23:30:55 <BitL0G1c> I was just trying it in KVM - but did not succeed - even with a vanilla kernel
2015-04-08 23:31:11 <systmkor> for MATE or XFCE?
2015-04-08 23:31:22 <systmkor> I currently am running XFCE on my laptop
2015-04-08 23:31:55 <BitL0G1c> XFCE
2015-04-08 23:31:59 <systmkor> strange
2015-04-08 23:33:59 <BitL0G1c> don't set the display manager to autostart - setup ssh so you can check the logs for errors
2015-04-08 23:34:49 <systmkor> the manager doesn't auto start luckily
2015-04-08 23:35:00 <systmkor> it boots up into console login
2015-04-08 23:37:26 <__number5__> why so many people trying to use alpine for desktop
2015-04-08 23:37:44 <BitL0G1c> try with linux-vanilla so you can rule out grsecurity as the problem
2015-04-08 23:38:29 <BitL0G1c> you will need to select the vanilla kernel when it boots
2015-04-08 23:38:41 <systmkor> yesh, I'm familiar with that
2015-04-08 23:38:52 <systmkor> was doing custom kernels for a little while since default grsec kernel for Alpine
2015-04-08 23:39:00 <BitL0G1c> __number5__: - no systemd most probably
2015-04-08 23:39:02 <systmkor> doesn't have all the cgroups, chroot, and pax features
2015-04-08 23:41:29 <__number5__> BitL0G1c: a very good reason indeed
2015-04-08 23:42:14 <BitL0G1c> i'm still using sysvinit on wheezy - but that won't last forever
2015-04-08 23:43:02 <systmkor> okay LOOOL
2015-04-08 23:43:08 <systmkor> so it's a touch screen lenovo
2015-04-08 23:43:19 <systmkor> the touch screen works but not the keyboard, clit mouse, or touchpad
2015-04-08 23:43:21 <systmkor> LOOOOOL
2015-04-08 23:44:56 <BitL0G1c> any idea how to get rid of "grsec: denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/bin/Xorg"
2015-04-08 23:45:54 <BitL0G1c> nm - ulimit -c 4096
2015-04-08 23:53:47 <Orochimarufan> anyway, cya later I need to get some zzZs
2015-04-09 00:00:09 <BitL0G1c> cirrus driver is broken
2015-04-09 00:03:25 <BitL0G1c> what is "AlpProbe" ?
2015-04-09 00:09:30 <_mjones> systmkor: I haden't yet heard anybody but myself use that name for the pointer stick. I have to catch myself before I say it because someone will take it the wrong way eventually.
2015-04-09 02:37:35 <BitL0G1c> a desktop with spice-server is fantastic
2015-04-09 02:48:58 <_mjones> I considered running a fanless thin-client and spicing to a big server for everything.
2015-04-09 02:50:00 <BitL0G1c> I'm impressed how quick it is - copy & pasting between the client & host needs a bit of work
2015-04-09 02:51:39 <_mjones> Using spicec I was never able to do that.
2015-04-09 02:53:03 <_mjones> So, a number of national governments have sponsored versions of Linux or other open-source operating systems for national use.
2015-04-09 02:53:38 <_mjones> Yet I have not found evidence to date that any of this top-down effort has produced any significant contributions back upstream, or created anything new of note.
2015-04-09 02:54:21 <BitL0G1c> I'm using spicec & setup a virtio serial device in libvirt
2015-04-09 02:54:45 <BitL0G1c> I think spice-vdagentd needs to run
2015-04-09 02:55:04 <_mjones> I thought the Cubans and the Venezuelans with their Bolivarian revolution were supposed to be all advanced and productive. Yet nothing seems to come from them.
2015-04-09 02:56:09 <BitL0G1c> "oppressed" but have quite a lot of free time to not work
2015-04-09 03:26:50 <systmkor> _mjones: that's the only word I knew for that type of mouse until like 2 years ago
2015-04-09 04:27:43 <_mjones> systmkor: the technical name is 'Trackpoint'. Some people say 'nipple mouse' but that's clearly not correct. Makes a good backup to hardware or driver problems with trackpad.
2015-04-09 04:28:32 <_mjones> It's sort of interesting the evolution that has resulted in the dominance of trackpads. I really hated them at first and I still don't love them, but I use them before I use the stick.
2015-04-09 04:34:16 <systmkor> agreed
2015-04-09 07:38:04 <clandmeter> good morning
2015-04-09 07:38:14 <clandmeter> ncopa, i have a problem with openrc
2015-04-09 07:38:34 <clandmeter> it seems it doesnt want to start udev dbus for whatever reason
2015-04-09 07:38:48 <clandmeter> rc.log doesnt say anything
2015-04-09 07:40:47 <ncopa> lxc? vserver? on the iron?
2015-04-09 07:40:53 <ncopa> rc --sys
2015-04-09 07:41:35 <ncopa> there might be a keyword in init.d script that tells it to not start
2015-04-09 07:41:37 <clandmeter> iron
2015-04-09 07:42:04 <clandmeter> i first set  it up via tmpfs which worked fine. bit weird
2015-04-09 07:42:09 <_mjones> I seem tentatively to have the same problem, in a KVM fullvirt instance. Nothing from rc --sys.
2015-04-09 07:42:20 <_mjones> Edge, x86-64.
2015-04-09 07:43:17 <clandmeter> ncopa: like _mjones nothing
2015-04-09 07:43:26 <_mjones>         # udev does not work inside vservers         keyword novserver nolxc
2015-04-09 07:43:54 <clandmeter> it looks like alsa also didnt start
2015-04-09 07:44:31 <clandmeter> http://sprunge.us/FHeL my run levels
2015-04-09 07:45:01 <clandmeter> rc-status -a http://sprunge.us/XILX
2015-04-09 07:45:07 <_mjones> it's a big init script (rc script).
2015-04-09 07:45:47 <_mjones>  * You should add udev-postmount service to your default runlevel.
2015-04-09 07:46:15 <clandmeter> i did
2015-04-09 07:46:23 <clandmeter> /etc/runlevels/default/udev-postmount
2015-04-09 07:46:50 <clandmeter> ncopa: its edge, if that makes a diff
2015-04-09 07:47:43 <ncopa> clandmeter: i think you need remove mdev
2015-04-09 07:48:41 <ncopa> rc-update -q del mdev sysinit
2015-04-09 07:48:41 <ncopa> rc-update -q add udev sysinit
2015-04-09 07:48:42 <ncopa> rc-update -q add udev-postmount default
2015-04-09 07:49:09 <clandmeter> hmm
2015-04-09 07:49:19 <clandmeter> i was thinking about doing that, but didnt dare :)
2015-04-09 07:49:33 <ncopa> hwdrivers
2015-04-09 07:49:41 <ncopa> remove hwdrivers too i think
2015-04-09 07:49:55 <clandmeter> box is remote
2015-04-09 07:50:04 <clandmeter> hope i dont stop it :)
2015-04-09 07:50:07 <ncopa> i think both udev and mdev provides /dev or so
2015-04-09 07:50:12 <ncopa> shouldnt
2015-04-09 07:50:18 <_mjones> @clandmeter: that was advice emitted from 'service udev start'; I didn't have it.
2015-04-09 07:51:18 <clandmeter> ncopa: i dont think have mdev in my run levels?
2015-04-09 07:51:36 <ncopa> but you have it in http://sprunge.us/XILX
2015-04-09 07:51:55 <clandmeter> but not in http://sprunge.us/FHeL
2015-04-09 07:52:09 <ncopa> it could be hwdrivers that pulls it it?
2015-04-09 07:52:12 <ncopa> something has pulled it it
2015-04-09 07:52:13 <_mjones> It's in runlevel sysinit by default I think.
2015-04-09 07:52:32 <ncopa> you need manually stop it if you dont want reboot
2015-04-09 07:52:42 <clandmeter> i can reboot
2015-04-09 07:53:00 <ncopa> this is how i do it from script: http://git.alpinelinux.org/cgit/alpine-conf/tree/setup-xorg-base.in
2015-04-09 07:53:03 <clandmeter> i wonder why it worked on tmpfs
2015-04-09 07:53:10 <clandmeter> i used that script
2015-04-09 07:53:19 <ncopa> hum psmouse is no longer needed
2015-04-09 07:53:39 <ncopa> it sounds like something is pulling in mdev
2015-04-09 07:53:45 <ncopa> grep mdev /etc/init.d/
2015-04-09 07:53:53 <_mjones> @clandmeter: is this an old install upgraded continuously for years?
2015-04-09 07:54:04 <ncopa> inpsect /etc/init.d/mdev and see what it provides
2015-04-09 07:54:19 <ncopa> iirc it provides dev or similar
2015-04-09 07:54:21 <ncopa> grep for that too
2015-04-09 07:54:28 <clandmeter> http://sprunge.us/gGBS
2015-04-09 07:56:57 <clandmeter> grep dev is a bit verbose :)
2015-04-09 07:57:54 <clandmeter> ncopa: sprunge dev is here http://sprunge.us/dYWB
2015-04-09 08:03:28 <ncopa> what happens if you manually tries to /etc/init.d/udev start
2015-04-09 08:03:40 <clandmeter> works fine
2015-04-09 08:03:49 <clandmeter> i can start them all fine
2015-04-09 08:04:03 <clandmeter> udev dbus and alsa
2015-04-09 08:04:35 <clandmeter> udev could be mdev related, but the other 2?
2015-04-09 08:07:01 <ncopa> clandmeter> it seems it doesnt want to start udev dbus for whatever reason
2015-04-09 08:07:04 <ncopa> <clandmeter> i can start them all fine
2015-04-09 08:07:04 <ncopa> <clandmeter> udev dbus and alsa
2015-04-09 08:07:18 <ncopa> do they start or do they dont start?
2015-04-09 08:07:31 <ncopa> if they start, what is the problem?
2015-04-09 08:07:35 <clandmeter> they all starts fine manually
2015-04-09 08:07:42 <clandmeter> but not from init
2015-04-09 08:07:46 <ncopa> at boot?
2015-04-09 08:07:54 <clandmeter> no not at boot
2015-04-09 08:08:12 <clandmeter> the 2 sprunges show what has started and what is added to runlevels
2015-04-09 08:08:31 <ncopa> this shows thats mdev was started: http://sprunge.us/XILX
2015-04-09 08:08:39 <ncopa> and mdev-mount
2015-04-09 08:08:44 <clandmeter> yes
2015-04-09 08:08:51 <ncopa> Dynamic Runlevel: needed
2015-04-09 08:08:53 <clandmeter> i dont know what causes that
2015-04-09 08:08:56 <ncopa> means something pulled it in
2015-04-09 08:09:18 <ncopa> what are the dependencies for hwdrivers?
2015-04-09 08:09:25 <ncopa> my guess is that hwdrivers pulls it in
2015-04-09 08:09:26 <clandmeter> ./tmpfiles.dev:before dev
2015-04-09 08:09:42 <ncopa> devfs
2015-04-09 08:10:22 <clandmeter> devfs?
2015-04-09 08:12:34 <clandmeter> ncopa: can i safely remove hddrivers and reboot?
2015-04-09 08:12:39 <clandmeter> hwdrivers
2015-04-09 08:13:05 <ncopa> if you have udev there it should work
2015-04-09 08:13:07 <ncopa> i can check
2015-04-09 08:13:52 <ncopa> hum
2015-04-09 08:14:28 <ncopa> this is my desktop: http://sprunge.us/jCeK
2015-04-09 08:14:46 <ncopa> and i have hwdrivers there
2015-04-09 08:15:31 <clandmeter> i had this issue before when i tested edge months ago
2015-04-09 08:15:47 <clandmeter> but last week i tried it, and everything worked on usb tmpfs
2015-04-09 08:16:01 <clandmeter> now i wanted to convert to hd install, and its back again.
2015-04-09 08:17:33 <clandmeter> ncopa: your desktop is edge?
2015-04-09 08:25:18 <ncopa> yes
2015-04-09 08:25:28 <ncopa> same with my laptop
2015-04-09 08:27:55 <_mjones> the man is committed if nothing else.
2015-04-09 08:39:51 <clandmeter> hmm
2015-04-09 08:40:22 <clandmeter> ncopa: this fixed it: for i in `apk info`;do sudo apk fix $i; done
2015-04-09 08:40:25 <clandmeter> :|
2015-04-09 08:42:17 <ncopa> heh
2015-04-09 08:42:19 <ncopa> strange
2015-04-09 10:11:24 <Orochimarufan> morning
2015-04-09 11:05:24 <Orochimarufan> does musl miss some AF_CAN structures or something?
2015-04-09 11:09:37 <Orochimarufan> hmm those are in linux-headers, not libc, apparently
2015-04-09 11:13:28 <Orochimarufan> Note to whoever is packaging python3: it needs linux-headers to build correctly
2015-04-09 11:36:28 <clandmeter> Orochimarufan: its already in edge
2015-04-09 11:36:41 <clandmeter> but there is some breakage currently
2015-04-09 11:36:46 <Orochimarufan> exactly
2015-04-09 11:36:51 <clandmeter> will be resolved shortly
2015-04-09 11:37:00 <Orochimarufan> As I said, it needs linux-headers to build
2015-04-09 11:37:14 <clandmeter> it has
2015-04-09 11:37:22 <fabled> Orochimarufan, yes, it's already committed. build box is just catching up
2015-04-09 11:37:23 <Orochimarufan> didn't yesterday
2015-04-09 11:37:30 <Orochimarufan> I see
2015-04-09 11:37:32 <fabled> was fixed earlier today
2015-04-09 11:37:44 <fabled> i think x86_64 is up-to-date already
2015-04-09 11:37:53 <fabled> x86 buildbox is still compiling stuff
2015-04-09 11:37:55 <clandmeter> kind of
2015-04-09 11:37:58 <Orochimarufan> well, I have a custom package for now
2015-04-09 11:38:31 <Orochimarufan> Once I have everything live, I'll try to pull back in the official packages
2015-04-09 11:40:12 <Orochimarufan> FileNotFoundError: [Errno 2] No such file or directory
2015-04-09 11:40:34 <Orochimarufan> Don't we all love it when it tells you "File not found" but fails to mention WHAT file wasn't found >.<
2015-04-09 11:47:57 <Orochimarufan> finally.. I got it to work :)
2015-04-09 12:43:22 <Orochimarufan> Eureka :D
2015-04-09 12:43:57 <Orochimarufan> In the end, it was just good old unix permissions >.<
2015-04-09 13:30:35 <Orochimarufan> samba-dc is missing a dependency on py-tdb in edge
2015-04-09 14:16:33 <ncopa> is it samba-dc that needs it or is it py-samba that needs it?
2015-04-09 14:17:31 <ncopa> pkg/py-samba/usr/lib/python2.7/site-packages/samba/samba3/__init__.py:import tdb
2015-04-09 14:17:50 <ncopa> look like its py-samba that needs it as depends
2015-04-09 14:51:46 <andyshinn> anyone know if there is a way to ignore automatically adding dependencies when building packages? i'm trying to avoid having libgcc automatically added to a package:
2015-04-09 14:51:50 <andyshinn> https://www.irccloud.com/pastebin/zCGHqXHM
2015-04-09 15:02:34 <BitL0G1c> I think you can put !pkgname in the APKBUILD
2015-04-09 16:07:39 <gandaro> Hi!  Is it possible to install Alpine Linux on a Raspberry Pi without having to plug in a monitor?
2015-04-09 16:09:23 <gandaro> Ah, I just got a brilliant idea: instead of bringing my monitor to the Raspberry Pi, I will bring the Raspberry Pi to my monitor
2015-04-09 16:09:30 <gandaro> Then it will work :P
2015-04-09 16:11:34 <jomat> gandaro: though i've never done it, it should be possible. Your root-fs has to start up networking and everything you need to access the pi... I've done that once with a botstrapped debian w/o monitor
2015-04-09 16:13:42 <gandaro> I was asking because you can go through the Raspbian installation via SSH
2015-04-09 16:19:22 <andyshinn> BitL0G1c: that seems to mean that it conflicts with that package, not just removing the dependency
2015-04-09 16:20:19 <jomat> gandaro: oh, and I usually install my servers on dedicated hardware with alpine w/o the ability to access the vga port. that also works :-)
2015-04-09 16:29:32 <Orochimarufan> ncopa: quite possible
2015-04-09 16:29:44 <Orochimarufan> ncopa: quite likely, in fact
2015-04-09 16:33:33 <Orochimarufan> Is there a way to get rid of kernel messages in /var/log/messages?
2015-04-09 16:48:55 <_mjones> Orochimarufan: the default syslog is busybox syslog, and I'm not sure if it has a syslog.conf configurability (there is no syslog.conf by default).
2015-04-09 16:49:21 <_mjones> While it should be possible to remove them, I'd give serious thought against doing so.
2015-04-09 16:50:09 <_mjones>  /var/adm/messages is often your only durable (across reboots; over time as the kernel log fills its cyclic buffer) log of kernel messages and errors.
2015-04-09 16:51:13 <_mjones> rsyslog is in the packages, but it's not as good as I'd like: the default "-c5" argument needs to be removed from the package because rsyslog segfaults on alpine (probably due to pax/pie/aslr and/or musl) with it.
2015-04-09 16:54:54 <Orochimarufan> I have it logged on the lxc host
2015-04-09 16:55:02 <Orochimarufan> don't need it in the container
2015-04-09 16:55:40 <BitL0G1c> inetutils-syslogd is in testing too
2015-04-09 16:55:51 <Orochimarufan> what's it do?
2015-04-09 16:56:06 <BitL0G1c> gives the same logging layout as in debian
2015-04-09 16:56:26 <BitL0G1c> remote logging disabled
2015-04-09 16:56:29 <Orochimarufan> hmm
2015-04-09 16:56:49 <Orochimarufan> well, I really only need busybox syslog, minus kernel messages
2015-04-09 16:59:05 <BitL0G1c> you will have a syslog.conf you can edit http://hastebin.com/oqoqaqisid.vala
2015-04-09 17:02:19 <_mjones> Orochimarufan: ah; the containers situation would be a time you can drop such logging.
2015-04-09 17:02:54 <_mjones> BitL0G1c: you mean to confirm that the standard syslog.conf syntax works with busybox syslog? I assume it does.
2015-04-09 17:03:06 <Orochimarufan> busybox doesn't use syslog.conf at all
2015-04-09 17:03:21 <_mjones> he means inetdutils-syslog then.
2015-04-09 17:03:28 <Orochimarufan> System logging utility
2015-04-09 17:03:30 <Orochimarufan> (this version of syslogd ignores /etc/syslog.conf)
2015-04-09 17:03:55 <_mjones> rsyslog can be configured not to listen remotely, if that's a reason why someone wouldn't use it.
2015-04-09 17:04:27 <_mjones> <Orochimarufan> (this version of syslogd ignores /etc/syslog.conf)
2015-04-09 17:04:28 <_mjones> ah, so.
2015-04-09 17:05:53 <BitL0G1c> busybox ignores syslog.conf
2015-04-09 18:39:54 <Orochimarufan> is there no APKBUILD for (non open-) ntpd?
2015-04-09 18:43:27 <andyshinn> is it possible to have triggers for subpackages?
2015-04-09 18:43:41 <_mjones> Orochimarufan: correct, the reference implementation doesn't seem to be in aports.
2015-04-09 18:44:06 <Orochimarufan> Another APKGBUILD to write >.<
2015-04-09 18:44:27 <Orochimarufan> Just gonna piggyback off the PKGBUILD
2015-04-09 18:44:49 <andyshinn> it is pretty easy to translate PKGBUILD to APKBUILD
2015-04-09 18:45:07 <Orochimarufan> Indeed
2015-04-09 18:50:16 <_mjones> You don't like openntpd? They've both got disadvantages.
2015-04-09 18:50:28 <Orochimarufan> pkgver=${_pkgver/p/.p} << That isn't POSIX is it?
2015-04-09 18:50:38 <Orochimarufan> does openntpd do signed ntp?
2015-04-09 18:50:56 <_mjones> Don't know. If so it's probably a recent addition.
2015-04-09 18:51:05 <_mjones> You have chrony in the main repo too.
2015-04-09 18:51:34 <Orochimarufan> is that another alternative implementation?
2015-04-09 18:51:52 <_mjones> Despite trying to like chrony, I don't really, despite its claimed advantages. Something about it I can't put my finger on that causes me not to love it or have full faith in it.
2015-04-09 18:52:08 <_mjones> Yes, it's a clean-sheet implementation of the protocol.
2015-04-09 18:52:55 <_mjones> And I'm sure you saw sntpc which is a stub minimal (SNTP or simple ntp) implementation that doesn't run as a daemon
2015-04-09 18:53:29 <_mjones> sntp is perfectly adequate for many uses. I'd be thrilled if many of my embedded systems even implemented sntp client.
2015-04-09 18:53:39 <Orochimarufan> well, I need the server portion
2015-04-09 18:54:03 <_mjones> http://arstechnica.com/information-technology/2015/04/next-windows-server-offer-new-even-smaller-nano-server-footprint/
2015-04-09 18:54:52 <Orochimarufan> yeah... I don't think anyone but stock ntpd does signed ntp >.<
2015-04-09 18:55:21 <Orochimarufan> that link made otter crash lol
2015-04-09 18:55:32 <_mjones> Following in the footsteps of Alpine Linux, Windows is getting a minimal-footprint distro suitable for virt-servers, firewalls, and routers, and also getting containers and Docker-like isolation containers.
2015-04-09 18:55:53 <Orochimarufan> well what did you expect
2015-04-09 18:56:04 <Orochimarufan> The question is "will it be as good" ;=)
2015-04-09 18:56:55 <_mjones> Windows manages to be orders of magnitude more lines of code deployed, while not having until very late things like usable scripting or GUI-less servers.
2015-04-09 18:57:29 <_mjones> I've had to manage more than a few Windows servers alongside Linux and others, and that world is totally different than the posix world.
2015-04-09 18:57:35 <Orochimarufan> yep
2015-04-09 18:58:38 <Orochimarufan> the sad part is, first thing you do with one of their "headless" servers is bang your head against a wall cause the commandline is next to useless or at least complicated enough that it's useless without 4 hrs of reading stuff first
2015-04-09 18:58:44 <_mjones> What's funny is the new trend for windows-world people to want to get some of that devops. They're about four years behind, still, but with enough effort they'll be able to implement enough vague equivalents to claim parity.
2015-04-09 18:59:54 <_mjones> powershell is obtuse and foreign if you've used any other command-line, true. But the bigger tragedy is that so few things work 100% without the GUI. Virtually nothing was ever designed not to need the GUI, and everything was supposed to be 'easy to use'.
2015-04-09 18:59:57 <Orochimarufan> (not to mention that it's backwards anyway, like running X11 in kernel just so you can use XTerm. Long live VTs, no matter what Poettering says...)
2015-04-09 19:01:13 <Orochimarufan> anyway, is ${_pkgver/p/.p} valid POSIX? if not, what do I substitute?
2015-04-09 19:01:44 <Orochimarufan> I don't remember what it does
2015-04-09 19:01:55 <_mjones> I don't know if I understand that syntax. You mean can you use sed syntax inside a variable?
2015-04-09 19:02:11 <_mjones> I suspect that's a bashism.
2015-04-09 19:02:19 <Orochimarufan> also, https://technet.microsoft.com/en-us/library/cc773013(v=ws.10).aspx [NTP Security] section :(
2015-04-09 19:02:47 <Orochimarufan> Now the question is what does it do and how to replace it -.-
2015-04-09 19:04:26 <Orochimarufan> ${parameter/pattern/string} performs pattern replacement, which is a bashism. See the next section for suggestions on fixing this.
2015-04-09 19:04:29 <Orochimarufan> aha
2015-04-09 19:04:29 <_mjones> That NTP Security section doesn't mean much. However, normally I'd run the domain controllers as NTP clients of four or more reliable timesources, then run the domain/forest with its internal timesync (type nt5domain or something).
2015-04-09 19:04:55 <_mjones> That was my guess on both counts. I'm so clever I deserve another espresso.
2015-04-09 19:05:12 <Orochimarufan> well, I'm not setting up 4 time servers for a mockup virtual network :P
2015-04-09 19:06:01 <Orochimarufan> also, what's that "internal timesync" part?
2015-04-09 19:06:03 <_mjones> There's a new trend to ssldate over ntp. I understand it, but I'm not hugely persuaded that ntp has so many disadvantages. The biggest disad would be ddos potential, which is a shame but we've known this for many years.
2015-04-09 19:06:46 <_mjones> ad domains do internal timesync by default. They must. Regular kerberos will fail if time becomes off by >5 minutes, and this applies to AD Kerberos as well.
2015-04-09 19:07:13 <_mjones> Just use 1 for a mockup, or 4 hostnames pointing to 1 IP or something. You do want 4 or 5 in production, though.
2015-04-09 19:07:48 <_mjones> Also if it's virtual you might be able to ignore it. KVM seems to do time ok -- I need to check more thoroughly -- but vmware needs ntp clients on all the guests badly.
2015-04-09 19:07:51 <Orochimarufan> Indeed that 5min problem has screwed me over multiple times now
2015-04-09 19:08:22 <Orochimarufan> my kvm doesn't seem to get it right
2015-04-09 19:08:33 <_mjones> vmware used to recommend timesync through vmware tools (open sourced as open-vm-tools) but that's a poor choice and a couple of years ago they reversed their recommendation entirely. Which was good, because they were wrong before.
2015-04-09 19:09:10 <Orochimarufan> it's been acting up since I redid the system over all
2015-04-09 19:09:10 <_mjones> kvm has a tsc option for this but I haven't needed to mess with it as long as I do -rtc based=utc
2015-04-09 19:10:07 <_mjones> timezone off or generic drift? I haven't seen a drift issue with kvm.
2015-04-09 19:11:15 <systmkor> has anyone worked with buildroot on Alpine? I started futzing with it. Looks great. However having an issue to find were to turn of LOCALE support.
2015-04-09 19:13:33 <Orochimarufan> There... soon as the time was right (pun unintended :P) the join succeeded :)
2015-04-09 19:14:12 <_mjones> oh you're actually running windows? I thought this was an academic discussion.
2015-04-09 19:14:27 <_mjones> I could have told you all about this.
2015-04-09 19:14:33 <Orochimarufan> Using windows for RSAT
2015-04-09 19:14:39 <Orochimarufan> Well I knew
2015-04-09 19:14:49 <Orochimarufan> but I forgot (yet again -.-)
2015-04-09 19:15:48 <_mjones> I just use rdesktop and xfreerdp
2015-04-09 19:16:00 <Orochimarufan> why the fuck is the time off by 5 mins again
2015-04-09 19:16:13 <Orochimarufan> what the F are you doing qemu
2015-04-09 19:21:37 <systmkor> nvm found it I think
2015-04-09 19:28:12 <Orochimarufan> Out of curiosity: how would I get APKBUILDS back into aports?
2015-04-09 19:28:24 <Orochimarufan> like, who do I have to talk to
2015-04-09 19:29:39 <_mjones> There's a good article on wiki.alpinelinux.com.
2015-04-09 19:29:59 <systmkor> http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2015-04-09 19:30:00 <_mjones> Basically, it's a mailing list like linux kernel mailing list. There is no github pull-request system.
2015-04-09 19:30:01 <systmkor> more specifically
2015-04-09 19:30:12 <Orochimarufan> I see
2015-04-09 19:30:32 <_mjones> One uses git then exports the diffs into one unified diff per package into an email, then sends it.
2015-04-09 19:30:35 <systmkor> unless you consider irc a pull-request system :P
2015-04-09 19:31:04 <Orochimarufan> it's more like a push-request system :P
2015-04-09 19:31:05 <systmkor> ncopa: accepts or denies the package (idk who else reviews package requests official)
2015-04-09 19:31:15 <_mjones> I have to redo my setup before I can send in mine.
2015-04-09 19:31:40 <systmkor> well the "pull" request is sent to alpine-aports mailing list
2015-04-09 19:31:44 <systmkor> Orochimarufan: but yah :D
2015-04-09 19:32:50 <Orochimarufan> why :P
2015-04-09 19:33:16 <Orochimarufan> wow... Note to self: don't leave unfinished sentences in the irc input...
2015-04-09 19:34:42 <Orochimarufan> ... ntpd sure was a lot less painful than either python or pulseaudio :O
2015-04-09 19:35:43 <_mjones> python you got unlucky and found a bug that devs missed because they always have that prerequisite installed. (Shouldn't have built, really.)
2015-04-09 19:36:11 <Orochimarufan> well, they DIDNT have it
2015-04-09 19:36:16 <Orochimarufan> because that was the problem
2015-04-09 19:36:26 <_mjones> pulseaudio is pulseaudio. It was invented by someone who was probably correct that the existing audio was imperfect, but who didn't respect the principles of unix nor the _status_quo_.
2015-04-09 19:36:50 <Orochimarufan> python silently fails on non-core modules it can't find deps for
2015-04-09 19:36:56 <_mjones> the APKBUILD didn't have it as a prerequisite. If it was already installed for other reasons it would have worked fine.
2015-04-09 19:37:13 <Orochimarufan> it would, it didn't tho
2015-04-09 19:37:44 <_mjones> If the change is small, I bet you can probably talk someone into submitting it for you. I'm not submitting yet, so I'm not yet that person.
2015-04-09 19:38:11 <Orochimarufan> I think they fixed the worst stuff already
2015-04-09 19:38:38 <_mjones> Just like the iojs package requires xz which isn't on the dependency list. I'm submitting that first (along with other minor changes).
2015-04-09 19:41:44 <Orochimarufan> what should I call the ntpd initscript I wonder... ntpd is already taken by busybox >.<
2015-04-09 19:44:04 <_mjones> You can't run two because they both need to bind to udp/123.
2015-04-09 19:44:15 <_mjones> Therefore you can call it ntpd.
2015-04-09 19:44:18 <Orochimarufan> I know but apk doesn't like conflicts
2015-04-09 19:44:51 <Orochimarufan> ERROR: ntp-4.2.8.p2-r1: trying to overwrite etc/init.d/ntpd owned by busybox-initscripts-2.2-r25.
2015-04-09 19:44:56 <_mjones> The real question is what is the script-etiquette for replacing a default initscript. What do openntpd and chrony do?
2015-04-09 19:48:13 <Orochimarufan> does the ntp group and user exist by default?
2015-04-09 19:49:20 <_mjones> Yes
2015-04-09 19:49:51 <_mjones> but I think you can do a adduser-if-notexist, which would be more robust
2015-04-09 19:50:18 <Orochimarufan> I dont want to ship a install script tho
2015-04-09 19:51:33 <_mjones> you can't do it in the function?
2015-04-09 19:51:54 <Orochimarufan> pretty sure you need to ship an install script for it
2015-04-09 19:52:26 <Orochimarufan> http://wiki.alpinelinux.org/wiki/APKBUILD_Reference#pkgusers
2015-04-09 19:53:22 <Orochimarufan> I think my windows vm is completely broken now >.<
2015-04-09 20:01:29 <clamiax> hi
2015-04-09 20:01:58 <_mjones> Hello.
2015-04-09 20:02:51 <clamiax> I stumbled across the bug #3691 and I really have no idea how to deal with it. Google has not been helpful, yet. Does anyone have a clue?
2015-04-09 20:04:32 <clamiax> Is #alpine-devel a more appropriate channel for such questions?
2015-04-09 20:05:31 <nsz> they wont know more than the answer on the linked issue
2015-04-09 20:05:44 <_mjones> uh, I think that there may have been a commit related to that pushed yesterday, but I'm not sure.
2015-04-09 20:06:20 <nsz> either you fix the module load order, or ldpreload stuff
2015-04-09 20:06:23 <_mjones> Discussion was had, anyway.
2015-04-09 20:06:34 <_mjones> right. what nsz said.
2015-04-09 20:06:37 <clamiax> nsz: this makes me out of luck
2015-04-09 20:07:04 <clamiax> _mjones: I'll give a look at the commit history. Thanks!
2015-04-09 20:08:03 <_mjones> At least one of nsz's stated options should work for you, but it needs to be documented I suppose.
2015-04-09 20:10:44 <clamiax> I also think so, I'll RTFM and play around a bit. Thank you for now.
2015-04-09 20:11:23 <gandaro> Hi!  I am trying to install on a Raspberry Pi.  I copied the tar archives contents onto an FAT32-formatted SD card (one partition that consumes all the space on the card).  I booted it, logged in and ran setup-alpine, but after a reboot everything is like before I have run setup-alpine.
2015-04-09 20:12:04 <nsz> probably you installed it in ram instead of onto the sd card
2015-04-09 20:12:40 <nsz> clamiax: i think alpine should have the loadorder fixed by default
2015-04-09 20:12:51 <gandaro> yes, I am installing it to /, which is a tmpfs
2015-04-09 20:13:27 <nsz> it should only break if you use special modules or disable the load order configs with things like 'Xorg -configure'
2015-04-09 20:14:00 <nsz> if you have to disable the load order config then try to ldpreload all xorg modules
2015-04-09 20:14:22 <_mjones> gandaro: sys, disk, or disless mode?
2015-04-09 20:14:38 <gandaro> _mjones, when I run setup-alpine, I do not get to choose
2015-04-09 20:14:59 <_mjones> I haven't used the Pi installer.
2015-04-09 20:24:39 <gandaro> It only asks me "where to store configs ('floppy', 'mmcblk0p1', 'usb' or 'none') [mmcblk0p1]"
2015-04-09 20:25:43 <gandaro> probably it wants to use diskless mode …
2015-04-09 20:26:33 <_mjones> ah, so when you select mmcblk0p1 it doesn't work?
2015-04-09 20:28:04 <gandaro> probably I have to "lbu commit"
2015-04-09 22:01:50 <systmkor> installing Alpine RPi, easiest install eva
2015-04-09 22:10:47 <gandaro> systmkor, ...
2015-04-09 22:10:57 <systmkor> gandaro: yes?
2015-04-09 22:11:14 <gandaro> systmkor, do you know how to make a sys installation on an RPi?
2015-04-09 22:11:47 <systmkor> meaning to take the release tarball and get a RPi to boot with it?
2015-04-09 22:11:55 <gandaro> no
2015-04-09 22:12:07 <systmkor> as in?
2015-04-09 22:12:07 <gandaro> how to make it so when you reboot everything's how you left it
2015-04-09 22:12:57 <systmkor> oh now not yet, I literally just dumped the data onto the sd card and booted up
2015-04-09 22:13:11 <systmkor> however I will with time
2015-04-09 22:13:14 <gandaro> and if there's a power outage and you reboot also everything's how you left it...
2015-04-09 22:14:02 <systmkor> at the moment I don't, however I will figure it out
2015-04-09 22:14:26 <gandaro> tell me when you did it
2015-04-09 22:14:28 <gandaro> :P
2015-04-09 22:15:14 <systmkor> I will
2015-04-10 04:47:28 <hiu> hello, I'm confused! Does 'apk add <some package>' command automatically checks the package's cryptographic signature before install? Or do I have to run 'apk verify <some package> manually?
2015-04-10 04:52:58 <_mjones> There are no crypto signatures.
2015-04-10 04:53:35 <_mjones> or maybe I'm wrong. Let me look.
2015-04-10 04:56:20 <_mjones> You don't need to do it manually, though. Just 'apk add'.
2015-04-10 04:58:49 <hiu> I see, you mean 'apk add' does the verification automatically? GOOD!
2015-04-10 04:59:10 <hiu> thanks mjones.
2015-04-10 05:41:16 <_ikke_> _mjones: isn't every package signed?
2015-04-10 05:42:09 <_mjones> _ikke_: somehow, but not cryptographically I don't think.
2015-04-10 05:43:13 <_mjones> I started to look at the verify.c in apk-tools, but it's not a casual thing to understand for me.
2015-04-10 05:44:47 <_ikke_> _mjones: look at /etc/apk/keys
2015-04-10 05:45:02 <_ikke_> Those are public keys that packages need to be signed with
2015-04-10 05:45:11 <_ikke_> (rsa keys)
2015-04-10 05:45:41 <_mjones> _ikke_: ah, and I should remember because I put my own in there.
2015-04-10 05:45:57 <_mjones> So the crypto verification is through openssl and not gpg.
2015-04-10 05:46:46 <_ikke_> yeah, probably
2015-04-10 05:46:54 <_ikke_> not gpg at least, because there is no need to sign those keys
2015-04-10 05:46:54 <_mjones> definitely
2015-04-10 05:47:13 <_mjones> apk.c:#include <openssl/crypto.h> apk.c:#ifndef OPENSSL_NO_ENGINE apk.c:#include <openssl/engine.h>
2015-04-10 05:47:14 <_ikke_> Just putting a key in that directory is enough
2015-04-10 05:47:17 <_ikke_> ah ok
2015-04-10 05:47:51 <fabled> yes, it's pretty much raw RSA signatures, and key trust is based on having the key in /etc/apk/keys which is the trust list
2015-04-10 05:47:59 <_mjones> explains why all of the verify functions have _ctx_ in the names too. An openssl trademark.
2015-04-10 05:49:04 <_mjones> tnx fabled. btw, if there is an https URL to pull dist tars, is there any policy reason we wouldn't prefer to use the https URL?
2015-04-10 05:49:27 <fabled> apk does not do https currently
2015-04-10 05:49:51 <_mjones> I thought it did. Let me check something.
2015-04-10 05:50:15 <fabled> mmm. the earlier versions just launched wget, and https probably worked then
2015-04-10 05:50:21 <fabled> now we are using libfetch
2015-04-10 05:52:26 <_mjones> this is edge, updated today, and https works.
2015-04-10 05:57:51 <fabled> oh
2015-04-10 05:57:56 <fabled> libfetch uses openssl
2015-04-10 05:58:05 <fabled> yes it works then :)
2015-04-10 06:05:49 <_mjones> I see it links against openssl but I can't find any tls code in it yet :P
2015-04-10 06:06:34 <_mjones> I rather suspect it doesn't support SNI, and I would prefer that it did support it.
2015-04-10 06:09:41 <_mjones> sweet. I was wrong, it does support SNI.
2015-04-10 06:11:29 <_mjones> Seems safe to use https anywhere we want.
2015-04-10 06:12:28 <dalias> fabled, speaking of tls..
2015-04-10 06:12:54 <dalias> is the ca-certificates package managed to remove compromised/untrustable CAs?
2015-04-10 06:28:56 <_mjones> fabled can answer that, but I'll point out that despite me remembering a commit the other day, we still have CNNIC in Edge. :(
2015-04-10 06:34:03 <_mjones> Upstream package hasn't been revised.
2015-04-10 06:38:10 <fabled> ncopa is the maintainer. but yes, seems debian has not update it's lists for a while
2015-04-10 06:39:18 <_mjones> I assumed they had, but at the given source there is no newer version.
2015-04-10 06:40:14 <Guest32443> alpine uses openrc right?
2015-04-10 06:40:35 <fabled> Guest32443, yes
2015-04-10 06:40:39 <Guest32443> can someone do a strace echo 1 > /sys/fs/cgroup/openrc/notify_on_release and pastbin the output?
2015-04-10 06:40:52 <Guest32443> something is buggy with musl
2015-04-10 06:41:02 <Guest32443> or openrc
2015-04-10 06:41:52 <fabled> Guest32443, http://sprunge.us/LRYf
2015-04-10 06:42:03 <Guest32443> wtf
2015-04-10 06:42:45 <fabled> coffee. brb.
2015-04-10 06:43:52 <Guest32443> what version of musl is that?
2015-04-10 06:45:08 <Guest32443> I have ioctl returning  ENOTTY before the write
2015-04-10 06:49:14 <Guest32443> what version of musl did you use?
2015-04-10 07:08:19 <pcarrier> ncopa: any progress with jdk8?
2015-04-10 07:09:02 <pcarrier> ncopa: openconnect and jdk8 would be the 2 packages I'd truly miss now, so otherwise I might finally get started
2015-04-10 07:22:31 <ncopa> pcarrier: no progress with jdk8
2015-04-10 07:22:52 <pcarrier> jdk7 will go unmaintained this month
2015-04-10 07:22:57 <ncopa> ugh
2015-04-10 07:23:09 <pcarrier> I know :)
2015-04-10 07:23:27 <ncopa> we have a problem then
2015-04-10 07:23:45 <pcarrier> pretty scared by the need to patch bytecode generation for musl
2015-04-10 07:24:15 <pcarrier> jdk7 seemed very limited
2015-04-10 07:24:29 <ncopa> can you bootstrap jdk8?
2015-04-10 07:24:35 <ncopa> from gcc
2015-04-10 07:24:36 <pcarrier> I haven't even tried, tbh
2015-04-10 07:25:00 <pcarrier> every system I've built jdk8 on already had a jdk6
2015-04-10 07:25:44 <ncopa> fabled: we might need look at jdk8
2015-04-10 07:25:52 <ncopa> but really
2015-04-10 07:25:57 <ncopa> what i want is split the main repo
2015-04-10 07:26:24 <ncopa> one repo for "longterm" support
2015-04-10 07:26:46 <ncopa> and other that we can build and ship but after a year or so dont need to worry about sec fixes
2015-04-10 07:27:15 <ncopa> we put the packages that upstream dont provide longterm support
2015-04-10 07:27:19 <pcarrier> that sounds very reasonable
2015-04-10 07:27:26 <ncopa> qemu for example
2015-04-10 07:27:37 <ncopa> 1.5 year release does not get sec fixes from upstream
2015-04-10 07:27:59 <ncopa> while xen has excellent long term support
2015-04-10 07:28:06 <pcarrier> I moved to chromeos, it's time for me to reinvest time in alpine :)
2015-04-10 07:28:45 <ncopa> openjdk is other package that we can ship, but probably not be able to provide longterm support for
2015-04-10 07:29:00 <ncopa> need to think about that, hum
2015-04-10 07:29:18 <pcarrier> ncopa: if you're reasonably in sync with the release cycle, I think you could get 3 years pretty reasonably
2015-04-10 07:29:38 <pcarrier> I wonder.
2015-04-10 07:29:59 <pcarrier> it's quite reasonable to consider jdkN+1 as backwards-compatible with jdkN
2015-04-10 07:30:19 <pcarrier> so the LTS approach could be to bump "major" versions within a single alpine release
2015-04-10 07:30:31 <pcarrier> but not more than once
2015-04-10 07:31:16 <pcarrier> I don't know of an API that got dropped in less than 2 major versions after deprecation
2015-04-10 08:44:39 <vkrishn> hmm.. another iconv lib, apr-iconv
2015-04-10 13:05:09 <c0da> hi everyone
2015-04-10 13:05:39 <c0da> i've got some error while trying to install the ruby-nokogiri package
2015-04-10 13:05:43 <c0da> could anyone help?
2015-04-10 13:23:52 <ncopa> c0da: do you have an exact error message?
2015-04-10 13:26:19 <c0da> ncopa: just says
2015-04-10 13:26:19 <c0da> ERROR: unsatisfiable constraints:
2015-04-10 13:26:19 <c0da>   ruby-nokogiri (missing):
2015-04-10 13:26:20 <c0da>     required by: world[ruby-nokogiri]
2015-04-10 13:26:25 <c0da> that's in 3.1
2015-04-10 13:26:37 <c0da> just tested it in edge and it works fine
2015-04-10 13:26:40 <c0da> starnge
2015-04-10 13:26:44 <c0da> *strange
2015-04-10 13:27:23 <ncopa> is the package there?
2015-04-10 13:27:36 <ncopa> apk search ruby-nokogiri
2015-04-10 13:27:54 <c0da> # apk search ruby-nokogiri
2015-04-10 13:27:55 <c0da> WARNING: Ignoring APKINDEX.689bb31a.tar.gz: No such file or directory
2015-04-10 13:28:06 <c0da> is that the issue?
2015-04-10 13:28:46 <c0da> oh no wait
2015-04-10 13:28:53 <c0da> i restarted the container
2015-04-10 13:28:55 <c0da> give me a sec
2015-04-10 13:29:21 <c0da> apk search ruby-nokogiri doesn't return anything
2015-04-10 13:33:45 <ncopa> so it was introduced after v3.1 was branched
2015-04-10 13:34:14 <c0da> oh ok
2015-04-10 13:34:20 <c0da> thank you for the help
2015-04-10 13:34:27 <c0da> it's a bit confusing though ;)
2015-04-10 13:35:04 <c0da> would be great to add a field to the package listing to show which version of alpine you need for a specific package
2015-04-10 13:35:30 <ncopa> probably yes
2015-04-10 13:35:33 <c0da> and even when using edge I can't get it to work how I would want to
2015-04-10 13:35:43 <c0da> ruby bundler su***
2015-04-10 13:35:45 <c0da> ;)
2015-04-10 15:30:52 <__number5__> c0da: it should be `bundle exec su ***` XD
2015-04-10 16:07:06 <_mjones> uh. no nfs on alpine, correct? Because no onc rpc on musl.
2015-04-10 16:09:49 <jomat> i boot an alpine diskless client for music in my living room from an alpine server with nfs
2015-04-10 16:11:27 <rpu3uO8PEVZZ> there is libirpc
2015-04-10 16:11:33 <rpu3uO8PEVZZ> *libtirpc
2015-04-10 16:13:19 <Jean-Scotch> _mjones:  nfs works fine with AL both as client or server but not in LXC
2015-04-10 16:14:11 <_mjones> Jean-Scotch: thanks. I think you're using it and I should remember from your wiki user page that it works, no?
2015-04-10 16:15:27 <Jean-Scotch> _mjones: ;)
2015-04-10 16:17:34 <Jean-Scotch> I did upgrade my notes http://wiki.alpinelinux.org/wiki/User_talk:Jch#About_NFS
2015-04-10 16:19:15 <_mjones> Oh lots of new notes J-S.
2015-04-10 16:19:24 <_mjones> what's blocking xnbd?
2015-04-10 16:20:43 <_mjones> I need to share video over NFS, but since it will be wifi I need it to be client-authenticated and not based on IP, so that would normally mean nfsv4 with krb5.
2015-04-10 16:23:27 <Jean-Scotch> I did not tried krb5... I'have no use for it i my use case...
2015-04-10 16:24:16 <Jean-Scotch> abount xnbd, I donnu really. I was able to pass the configure stage with a little dirty hack but I not the compile stage
2015-04-10 16:24:30 <Jean-Scotch> it's to low level in code for my expertise level
2015-04-10 16:24:54 <_mjones> I don't need or want block for this purpose (nbd or iscsi) although I've used plenty of iscsi in the past. Hmm.
2015-04-10 16:25:23 <Jean-Scotch> I did use iscsi before discovering nbd...
2015-04-10 16:25:33 <Jean-Scotch> ans now with xnbd I'm happy ;)
2015-04-10 16:25:43 <Jean-Scotch> but on AL :/
2015-04-10 16:26:03 <_mjones> You mention live migration with xnbd, and I'm always interested in that for the enterprise. I do a lot of live vm migration (albeit with vmware; I should try with KVM now).
2015-04-10 16:26:12 <Jean-Scotch> still I mostly use basic features of NBD so xnbd is not required on all my nbd servers
2015-04-10 16:26:51 <Jean-Scotch> yes xNBD allos NBD proxies and NBD live migration with the help of the proxy deature
2015-04-10 16:26:53 <_mjones> I used to run decent sized clusters and found NFS flat filesystem to be much simpler and less troublesome than multipathed block devices.
2015-04-10 16:27:18 <Jean-Scotch> *allows *features
2015-04-10 16:27:51 <Jean-Scotch> I'm running just a small shop. We have 7 NBD servers for now.
2015-04-10 16:27:58 <_mjones> That makes me wonder if amazon aws and google cloud might be using xnbd. They have secret sauce but from what I know it resembles nbd maybe.
2015-04-10 16:28:52 <Jean-Scotch> the home page of xnbd https://bitbucket.org/hirofuchi/xnbd/wiki/Home shows nice use cases for live migration
2015-04-10 16:30:01 <Jean-Scotch> And I would like to migrate my NBD servers to RDB (ceph)  in the course of this year
2015-04-10 16:30:28 <_mjones> I haven't picked back up the ceph work until I submit a few much-simpler patches. ;)
2015-04-10 16:30:54 <_mjones> The big changes occuring in every ceph release also spook me.
2015-04-10 16:31:08 <Jean-Scotch> but still I will keep some (x)NBD servers around as I'm very accustomed to those and can use them while sleeping ;) usefull in case of disaster recovery situation for some customer
2015-04-10 16:32:14 <Jean-Scotch> I think, for ceph, we have to first focus on the annouced "stable" and LTS version...
2015-04-10 16:33:08 <Jean-Scotch> Giant (v0.87.x) is a major release series that receives ongoing critical bug-fixes and stability improvements.
2015-04-10 16:34:30 <Jean-Scotch> and not trt to follown the 0.9x.y serie
2015-04-10 16:34:30 <Jean-Scotch> ceph is a very hot topic
2015-04-10 16:34:44 <Jean-Scotch> not hot even for alpine ;)
2015-04-10 16:34:52 <_mjones> I was working with 0.87.x but ncopa found the H release 0.93. I think all but one of the patches are the same so far (work in progress of course, not working).
2015-04-10 16:35:00 <Jean-Scotch> too hot even for alpine (I mean) ;)
2015-04-10 16:35:19 <Jean-Scotch> If I can be of any help?
2015-04-10 16:35:34 <Jean-Scotch> I have some free time for now
2015-04-10 16:36:04 <_mjones> I'll send you what ncopa assembled from both our work
2015-04-10 16:36:15 <Jean-Scotch> (waiting for my new lab machine... but still have some servers, desktop and laptop to play with...)
2015-04-10 16:41:03 <_mjones> I'm afraid I copied the wrong url and I can't send the current work, sorry.
2015-04-10 16:45:10 <_mjones> No, I found it. from ncopa: "curl http://sprunge.us/IdRT | git apply"
2015-04-10 16:47:33 <_mjones> The res_nquery -> res_query patch for configure.ac is a little off and the code itself needs the change, but the statefulness has already been pruned. Who knows how much work it needs to finish.
2015-04-10 17:00:31 <Jean-Scotch> will give it a try...
2015-04-10 17:00:33 <Jean-Scotch> thanks
2015-04-10 17:01:27 <Jean-Scotch> As I have some time and still an alpine-dev machine for compiling apkbuilds ;)
2015-04-10 17:07:18 <Jean-Scotch> I will report progress on the b.a.o pages #3241 or #3418
2015-04-10 17:07:54 <Jean-Scotch> Thanks, agilebot, but you forgot #3418
2015-04-10 17:08:39 <Jean-Scotch> algitbot: sorry for the mispelling ;)
2015-04-10 17:17:50 <Jean-Scotch> most difficult part will be to get the patches right
2015-04-10 17:19:05 <Jean-Scotch> could I pull from your git directly? or receive a snapshot of it?
2015-04-10 17:19:28 <_mjones> All I have is what I pasted. That's directly from ncopa.
2015-04-10 17:19:33 <Jean-Scotch> ok
2015-04-10 17:19:46 <_mjones> I should set something up. I need a github account so badly.
2015-04-10 17:20:18 <Jean-Scotch> I only use my own git server...
2015-04-10 17:20:30 <Jean-Scotch> I'm a newbie git related
2015-04-10 17:20:39 <Orochimarufan> Github is less maintenance-heavy :P
2015-04-10 17:21:20 <_mjones> At my previous shop we only used svn for many years. I was using git-svn workflow at the end, but I'm still not fluent in git.
2015-04-10 17:21:42 <Jean-Scotch> I want to learn about maintenance for all IT services I use ;) It's the beauty of FLOSS: one can really understand what's going on and what's involved...
2015-04-10 17:27:42 <Jean-Scotch> next step is to integrate consul, consul-template, PXE and GIT to have a fully operationnal HA infrastructure with _only_ PXEboot machines, be it physical or virtual ;)
2015-04-10 17:27:47 <Jean-Scotch> only with AL
2015-04-10 17:29:15 <Jean-Scotch> I think I will have within two weeks after receiving my new lab machine
2015-04-10 17:29:58 <Orochimarufan> I actually had a private git server
2015-04-10 17:30:16 <Orochimarufan> But when I nuked my VPS, I was too lazy to set it up again...
2015-04-10 17:30:39 <Orochimarufan> Github has WAY better discoverability anyway
2015-04-10 17:31:14 <Jean-Scotch> On bare metal alpine boxes, I currently am running about 80 AL KVM and LXC ;)
2015-04-10 17:31:31 <Orochimarufan> o.O
2015-04-10 17:32:15 <Orochimarufan> I'm only running 5 AL LXCs :P
2015-04-10 17:32:43 <_mjones> If you need code in your CV, Github has a lot of benefits. I've run my own svn servers so that isnt a big deal.
2015-04-10 17:33:05 <Orochimarufan> Plus one Arch container
2015-04-10 17:35:41 <Jean-Scotch> I do not want to appear on github...
2015-04-10 17:36:01 <Orochimarufan> Well that's your choice
2015-04-10 17:37:33 <Jean-Scotch> I need to practise because every so often I'm asked by some startups to setup private development environements ;)
2015-04-10 17:37:59 <Orochimarufan> Good for you :P
2015-04-10 17:39:25 <Jean-Scotch> I have still a few debian boxes around also because of xNBD and legacy mostly
2015-04-10 17:39:43 <Orochimarufan> xNBD?
2015-04-10 17:40:00 <Jean-Scotch> yep. an enhanced NBD
2015-04-10 17:40:10 <Orochimarufan> I had most of this stuff running on debian before too
2015-04-10 17:40:15 <Jean-Scotch> with proxy and live-migration features
2015-04-10 17:40:26 <Orochimarufan> And it doesn't work on AL?
2015-04-10 17:40:30 <Jean-Scotch> not yet
2015-04-10 17:40:48 <Jean-Scotch> NBD is in edge/testing
2015-04-10 17:41:14 <Orochimarufan> but not xNBD I guess?
2015-04-10 17:41:21 <Jean-Scotch> not xNBD
2015-04-10 17:41:43 <Orochimarufan> You could try putting together an APKBUILD
2015-04-10 17:42:07 <Jean-Scotch> I'm stuck trying to compile it see #3703
2015-04-10 17:44:09 <Orochimarufan> Argh... I'm doing this thing again.. Opening tons of tabs and not closing them after D:
2015-04-10 17:46:08 <Orochimarufan> looks like syntax errors to me
2015-04-10 17:48:08 <_mjones> Orochimarufan: I'm a tab-aholic in treatment. Working down from 200 right now. Worst was >580 (ff, Ubuntu, flashblock, noscript, adblock).
2015-04-10 17:48:26 <Orochimarufan> ouch
2015-04-10 17:48:58 <Orochimarufan> I usually roll multiple windows and still get up to like 30-60 tabs each >.<
2015-04-10 17:49:11 <_mjones> It's not new, though, tabs just make it worse. Twenty years ago you just had to use many browser windows.
2015-04-10 17:49:19 <Orochimarufan> yeah
2015-04-10 17:49:50 <_mjones> Chromium refuses to make your tabs wider so you can distinguish and manipulate them. They had a dev plugin to do it but seem to have exorcised it.
2015-04-10 17:50:03 <Orochimarufan> I mean, talking about tabs and windows, same goes for Terminals, not just browsers >.>
2015-04-10 17:50:57 <Orochimarufan> I have 3 terminals open with a few tabs each right now, but that's like the lower end of the spectrum -.-
2015-04-10 17:51:42 <Orochimarufan> And the fun really starts when you start splitting windows
2015-04-10 17:53:01 <_mjones> apparently on xfce-terminal all of your windows are one process?!?
2015-04-10 17:53:39 <_mjones> 3842 mjones    20   0  538m  27m 7364 S   0.0  0.2  81:07.86 xfce4-terminal
2015-04-10 17:54:03 <Orochimarufan> no clue how Konsole works
2015-04-10 17:54:58 <Orochimarufan> it seems to have 1 "host" process and 1 process per window, plus the stuff running in them ofc
2015-04-10 17:55:08 <Orochimarufan> Don't quote me on it tho
2015-04-10 17:56:28 <Orochimarufan> actually, scratch that, it also does 1 process only
2015-04-10 17:56:33 <Orochimarufan> those are threads :P
2015-04-10 19:41:21 <shodan45> is there a way to search for a package by file?
2015-04-10 19:43:41 <_mjones> shodan45: can you be more specific?
2015-04-10 19:44:30 <_mjones> shodan45: usually searching by package name is sufficient. But I bet you mean find outfrom which  package a given file came.
2015-04-10 19:45:08 <shodan45> _mjones: I'm looking for the package that has the real "dd"
2015-04-10 19:46:45 <_mjones> 'Real'?
2015-04-10 19:46:46 <_mjones> ;)
2015-04-10 19:47:03 <_mjones> I thought it was util-linux but apparently that's not it.
2015-04-10 19:48:42 <Diftraku> ian't it provided by busybox?
2015-04-10 19:48:54 <Diftraku> as a multi-call
2015-04-10 19:48:57 <shodan45> it is
2015-04-10 19:49:30 <shodan45> but I want the "real" one... I'm getting an error with the busybox one
2015-04-10 19:49:59 <_mjones> yes, you're right. I was just teasing.
2015-04-10 19:50:11 <_mjones> You mean the 'gnu' one, and I was casting aspersions at gnu.
2015-04-10 19:50:26 <ttilley> where do i find the definitions for the functions used in APKBUILD files? like update_config_sub?
2015-04-10 19:50:35 <_mjones> The 'real' one is BSD or maybe at&t, but I'm being obtuse.
2015-04-10 19:51:06 <_mjones> ttilley: good question; I'm having trouble with update_config_sub too. I was reading the source code.
2015-04-10 19:51:16 <_mjones> abuild is just a /bin/sh script.
2015-04-10 19:51:44 <_mjones> shodan45: I'm still looking, but "apk info --help" is generally useful
2015-04-10 19:54:53 <ttilley> ahhh... it just copies /usr/share/abuild/config.sub over any local config.sub
2015-04-10 19:55:13 <_mjones> shodan45: there's "dc3dd" in the unmaintained tree, but unmaintained stuff isn't built for repos.
2015-04-10 19:55:35 <shodan45> _mjones: found it: coreutils
2015-04-10 19:55:54 <_mjones> ttilley: I was having problems with the if ! ./$f bit
2015-04-10 19:56:28 <shodan45> oddly, I just asked my local opensuse desktop install what package owned /usr/bin/dd & it said "coreutils"
2015-04-10 19:56:38 <shodan45> same package name in alpine :)
2015-04-10 19:57:34 <_mjones> shodan45: indeed, you found it.
2015-04-10 19:57:52 <shodan45> and sure enough, the real dd worked
2015-04-10 19:58:02 <_mjones> Odd I didn't have that installed on my dev vm, which has most non-X things.
2015-04-10 19:58:15 <_mjones> shodan45: for future reference, what didn't work with busybox?
2015-04-10 19:58:20 <shodan45> holy crap... 812MB/s
2015-04-10 19:58:28 <_mjones> cache.
2015-04-10 19:58:33 <shodan45> dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync
2015-04-10 19:58:40 <shodan45> _mjones: that ^
2015-04-10 19:58:48 <shodan45> didn't like the fdatasync
2015-04-10 19:58:52 <_mjones> the conv didn't work?
2015-04-10 19:58:59 <shodan45> yeah
2015-04-10 19:59:22 <shodan45> _mjones: this is on my ramnode VM running alpine
2015-04-10 20:00:05 <_mjones> that seems like a nonstandard extension.
2015-04-10 20:00:11 <_mjones> I'm guessing it just fsyncs
2015-04-10 20:00:32 <_mjones> someone told you to run that to find the I/O speed eh?
2015-04-10 20:00:44 <shodan45> _mjones: yeah, more or less
2015-04-10 20:00:49 <shodan45> not "told"
2015-04-10 20:01:06 <shodan45> https://news.ycombinator.com/item?id=9353185
2015-04-10 20:01:21 <shodan45> I was just curious what my VM's disk I/O was like
2015-04-10 20:02:58 <_mjones> 1073741824 bytes (1.1 GB) copied, 16.288 s, 65.9 MB (Thinkpad T430, single Samsung 840P). More importantly, running KVM with QCOW2 image!
2015-04-10 20:04:16 <_mjones> They can still run most of that out of cache regardless of the memory bandwidth.
2015-04-10 20:04:21 <_mjones> Try it with 1TB ;)
2015-04-10 20:04:57 <shodan45> _mjones: yeah, repeating that test shows the speed varies a LOT
2015-04-10 20:05:02 <_mjones> fsync flushes to storage, unless it doesn't.
2015-04-10 20:05:16 <shodan45> 481MB/s, 1.5GB/s...
2015-04-10 20:05:24 <_mjones> i.e., the kernel or filesystem or hardware can lie to you.
2015-04-10 20:05:41 <shodan45> yeah, especially with virtualization
2015-04-10 20:05:51 <_mjones> "fsync? Totally written out to disk, dude!" ;)
2015-04-10 20:05:56 <shodan45> more layers = more people that LIE ;)
2015-04-10 20:06:05 <_mjones> Pretty much.
2015-04-10 20:06:10 <darkfader> more linux based layers :>
2015-04-10 20:06:14 <_mjones> More people who can lie, anyway.
2015-04-10 20:06:34 <_mjones> I'd look at bcache or dm-cache.
2015-04-10 20:06:53 <_mjones> Until recently I spent a lot of time doing heavy storage work (both block and nas).
2015-04-10 20:06:58 <darkfader> your data is "kinda reasonably safe as per some dev who defined safe based on his laptop experience"
2015-04-10 20:07:23 <_mjones> darkfader: let's not even go down that path.
2015-04-10 20:07:26 <darkfader> ok
2015-04-10 20:07:38 <darkfader> at least alpine is really clean on that end
2015-04-10 20:07:50 <darkfader> no need for conv=fdatasync in dd
2015-04-10 20:07:50 <_mjones> For starters, there's probably no COTS laptop that uses ECC ram.
2015-04-10 20:08:45 <_mjones> what fsync gets you is that storage is no longer your problem.
2015-04-10 20:08:59 <_mjones> If the stack lies to you after fsync(), it's all on them.
2015-04-10 20:09:35 <_mjones> And apps that want it flushed should call fsync and not assume. Calling fsync is best practise and avoids arguments with filesystem devs.
2015-04-10 20:13:09 <_mjones> So we've learned that ramnode probably has fat motherboards with big lanes to sata, or maybe 40gbit, or ideally infiniband, but really just that they have a huge amount of room to do the usual lying.
2015-04-10 20:13:30 <nwmcsween> hmm does alpine patch openrc for echoing /sys entries?
2015-04-10 20:13:47 <nwmcsween> seems gentoo openrc is broken
2015-04-10 20:14:19 <shodan45> I'd still like some way to ask apk "what packages have a file named $foo"
2015-04-10 20:14:21 <ttilley> so... how do you guys handle things like circular dependencies? i want to create a package for libcxx and libcxxabi but libcxxabi depends on libcxx and libcxx depends on either libcxxabi OR parts of libstdc++... and they both depend on clang. so you end up compiling each package several times in several configurations before you can get a clang that actually uses libcxx and libcxxabi
2015-04-10 20:14:29 <_mjones> how so nwmcsween?
2015-04-10 20:14:58 <_mjones> ttilley: compilers have bootstrap procedures; it shouldn't be an issue with most anything else.
2015-04-10 20:14:59 <nwmcsween> see #musl it tries to echo to sysfs entries and fails
2015-04-10 20:15:13 <nwmcsween> when it needs to echo -n
2015-04-10 20:15:20 <_mjones> I know I was just talking to you there.
2015-04-10 20:15:27 <_mjones> openrc uses regular echo?
2015-04-10 20:15:32 <nwmcsween> yep
2015-04-10 20:15:36 <_mjones> ...where?
2015-04-10 20:16:09 <ttilley> _mjones: so i'd end up having to make libcxx and libcxxabi sub-packages of llvm just like clang is i guess? and... fuck, i dunno, detect what you're compiling it with?
2015-04-10 20:16:22 <nwmcsween> lib/rc/sh/rc-cgroup.sh
2015-04-10 20:17:03 <ttilley> _mjones: then again i'm not done porting fixes to llvm/clang 3.6 or even started on compiler-rt so... one step at a time i guess
2015-04-10 20:17:58 <ttilley> does the alpine llvm package maintainer hang out here? William Pitcock aka nenolod?
2015-04-10 20:20:50 <_mjones> ttilley: not to my recollection.
2015-04-10 20:46:17 <_mjones> nwmcsween: if this is a bug between musl and openrc scripts, why hasn't it been noticed?
2015-04-10 20:48:16 <_mjones> about the ca-certificates: there is a fetch-crl in unmaintained that I was thinking about reviving. Seems there is current revocation mechanism than an apk update (and the debian upstream package is months old)?
2015-04-10 20:57:45 <_mjones> nwmcsween: I'm not getting an error using regular echo to cgroups specials in /sys...
2015-04-10 21:41:25 <BitL0G1c> _mjones - I've got copy & pasting working now for KVM guests
2015-04-10 21:56:16 <_mjones> BitL0G1c: how?
2015-04-10 21:57:08 <BitL0G1c> by fixing spice-vdagent
2015-04-10 21:57:38 <BitL0G1c> I've got it working for a single user at the moment
2015-04-10 21:58:23 <_mjones> nwmcsween: viz. # echo  "18959" > /sys/fs/cgroup/openrc/syslog/tasks && echo $? 0
2015-04-10 21:59:24 <BitL0G1c> need to find out why GetSessionForUnixProcess() fails so multiple sessions are possible
2015-04-10 22:07:03 <darkfader> ttilley: he used to
2015-04-10 22:07:10 <darkfader> (nenolog)
2015-04-10 22:07:19 <darkfader> there's also some alpine-dev channel.
2015-04-10 22:08:55 <_mjones> #alpine-devel. I should have mentioned that.
2015-04-10 22:10:44 <darkfader> 
ACTION collects his price for installing new stuff while moving /usr

2015-04-10 22:10:52 <darkfader> looking forward to more sleep :)
2015-04-11 05:04:37 <systmkor> ScrumpyJack: hey have you made any progress with libseccomp?
2015-04-11 05:05:51 <systmkor> I was successful compiling it on my machine, however I've not had a chance to try and test it.
2015-04-11 06:28:39 <maatunix> Hi there
2015-04-11 06:29:34 <maatunix> Do alpine support intel 7260 wireless card ?
2015-04-11 06:29:44 <maatunix> I can't find anything about that
2015-04-11 06:35:39 <maatunix> no packages here: http://pkgs.alpinelinux.org/packages?package=%25iwl%25&repo=all&arch=x86_64
2015-04-11 06:44:47 <maatunix> grmbl every body's sleeping X)
2015-04-11 06:45:59 <maatunix> Oh I find myself
2015-04-11 06:46:00 <maatunix> http://pkgs.alpinelinux.org/contents?filename=%25wifi%25&path=&pkgname=&arch=x86
2015-04-11 06:46:05 <maatunix> in da linux-firmware
2015-04-11 06:46:09 <maatunix> sounds good
2015-04-11 14:19:56 <Orochimarufan> Anyone tried running an AD member samba on AL? For some reason, I'm having authentication issues
2015-04-11 14:35:12 <Orochimarufan> hrm... It almost seems like the DC has issues
2015-04-11 14:38:27 <Orochimarufan> On another note: Is there any servlet container available on AL?
2015-04-11 15:11:05 <yunfan> hi, how to show/modify the repo address after login the system?
2015-04-11 15:13:18 <yunfan> ok got it
2015-04-11 16:22:59 <Orochimarufan> Anyone interested in a Jetty APKBUILD?
2015-04-11 16:33:01 <jyf> hi i have met some segment faul error when running w3m
2015-04-11 22:16:31 <owen1> i get: Error relocating /usr/lib/libstdc++.so.6: get_nprocs: symbol not found when running node - https://github.com/oren/alpine-iojs/blob/master/Dockerfile  any ideas?
2015-04-11 22:16:38 <mosez> andyshinn: your consule-* containers are not documented yet?
2015-04-11 22:22:35 <nsz> owen1: musl version?
2015-04-11 22:22:53 <nsz> run /lib/ld-musl*
2015-04-11 22:23:33 <nsz> if it's < 1.1.7 then there was no get_nprocs api in musl yet
2015-04-11 22:27:08 <owen1> nsz:
2015-04-11 22:27:14 <owen1> Version 1.1.6
2015-04-11 22:27:28 <andyshinn> mosez: not sure what you are referring to
2015-04-11 22:27:53 <owen1> nsz: how do i fix that?
2015-04-11 22:28:49 <nsz> update musl
2015-04-11 22:29:42 <owen1> nsz: RUN apk add --update iojs musl && rm -rf /var/cache/apk/*    ?
2015-04-11 22:30:23 <nsz> i think just updating musl should be enough
2015-04-11 22:30:35 <owen1> apk update musl ?
2015-04-11 22:30:52 <nsz> apk update is something different
2015-04-11 22:31:06 <owen1> nsz: oh. how do i update musl?
2015-04-11 22:31:44 <nsz> i dont know :P
2015-04-11 22:32:41 <owen1> aahha
2015-04-11 22:32:43 <owen1> np thanks
2015-04-11 22:33:12 <owen1> nsz: it used to work before. not sure why things changed
2015-04-11 22:33:41 <owen1> maybe this line is not safe? RUN echo "http://dl-4.alpinelinux.org/alpine/edge/testing" >> /etc/apk/repositories
2015-04-11 22:33:45 <nsz> apk update
2015-04-11 22:33:48 <nsz> apk upgrade
2015-04-11 22:34:01 <nsz> if you run these then you should get latest musl
2015-04-11 22:34:14 <owen1> nsz: yup! thank you.
2015-04-11 22:34:20 <nsz> but upgrade will do a lot of other things..
2015-04-11 22:34:32 <owen1> i guess i can put 'apk upgrade' in my Dockerfile
2015-04-11 22:38:22 <owen1> nsz: ERROR: Failed to set xattrs on usr/bin/iojs.apk-new: Not supported
2015-04-11 22:38:55 <owen1> i added 1 line to my Dockerfile (apk upgrade) and that's the error i see
2015-04-11 22:39:28 <noahmorrison> startx is causing my system to become unusable until reboot, screen goes black and Ctrl+Alt+F1 doesn't change my tty
2015-04-11 22:41:06 <nsz> noahmorrison: look at logs etc.. after login your env might not be set up the way you expect, so your .xinitrc script can fail
2015-04-11 22:41:38 <nsz> owen1: xattrs is some filesystem thing i guess.. i dont know why that is needed or why would it fail
2015-04-11 22:43:06 <owen1> ERROR: iojs-1.6.4-r0: Not supported
2015-04-11 22:43:22 <owen1> this is the other error i see
2015-04-11 22:44:38 <owen1> here is the dockerfile, in case anyone here have more insights - https://github.com/oren/alpine-iojs/blob/master/Dockerfile
2015-04-11 22:44:58 <nsz> strace -f command 2>&1 |grep ENOTSUP
2015-04-11 22:45:16 <owen1> nsz: great idea
2015-04-11 22:47:44 <owen1> nsz: grep ENOTSUP shows nothing. here is the output without the grep - http://paste.ubuntu.com/10804338/
2015-04-11 22:48:48 <owen1> 1847  write(1, "ERROR: Failed to set xattrs on u"..., 67) = 67
2015-04-11 22:49:09 <owen1> 1847  write(1, "ERROR: iojs-1.6.4-r0: Not suppor"..., 36) = 36
2015-04-11 22:52:15 <nsz> you are running the strace outside docker?
2015-04-11 22:53:02 <nsz> you get the error through a socket.. over http..
2015-04-11 22:53:24 <nsz> you need to strace the commnd that fails not the container thing
2015-04-11 22:53:43 <owen1> i did this: strace -f docker-compose -o output.txt build 2>&1
2015-04-11 22:53:56 <owen1> sorry,
2015-04-11 22:54:05 <owen1> strace -o output.txt -f docker-compose build 2>&1
2015-04-11 22:54:54 <nsz> actually i dont know what docker is
2015-04-11 22:55:00 <owen1> ahahha
2015-04-11 22:55:07 <nsz> so you are on your own
2015-04-11 22:55:12 <owen1> yup ):
2015-04-11 22:56:02 <owen1> docker is a way to isolate your app. docker-compose build is a command that look at your Dockerfile and execute whatever in that file.
2015-04-11 22:56:26 <owen1> it creates an image that u can later turn into a running container.
2015-04-11 22:56:35 <nsz> i know that much
2015-04-11 22:56:43 <nsz> but details matter
2015-04-11 23:02:41 <owen1> nsz: i fixed it by adding another 'apk update'.. looks hacky but now the build works and i can run node - https://github.com/oren/alpine-iojs/blob/master/Dockerfile#L5
2015-04-11 23:06:59 <owen1> interesting to note that the image is 29 MB (it used to be 22)
2015-04-11 23:45:28 <ovf> owen1: you might want to clear the cache last
2015-04-11 23:56:26 <owen1> ovf: run rm -rf /var/cache/apk/* as the last command?
2015-04-11 23:57:33 <ovf> owen1: yes, just a wild guess seeing as you added more apk business after cleaning the cache and your image sized increased
2015-04-11 23:58:27 <owen1> ovf: oh. from some reason i can't build the image
2015-04-11 23:58:35 <owen1> so i guess my hack didn't work
2015-04-11 23:59:07 <owen1> i get the same errors: Failed to set xattrs on usr/bin/iojs.apk-new: Not supported   iojs-1.6.4-r0: Not supported
2015-04-12 00:06:19 <ovf> quick google shows there's no xattr support in docker: https://github.com/docker/docker/issues/1070
2015-04-12 00:06:38 <ovf> so just ignore this, i guess? if it's for some reason fatal it is likely has to be made non-fatal
2015-04-12 00:41:00 <owen1> ovf: i am so confused. i remvoed all the images on my laptop and now 'docker build' works..
2015-04-12 00:47:45 <ovf> i'm sorry.
2015-04-12 00:57:57 <owen1> ahah
2015-04-12 00:58:54 <owen1> is there a way to make sure my docker images are IDENTICAL? i feel that this line might install different things on my image each time i run it - RUN echo "http://dl-4.alpinelinux.org/alpine/edge/testing" >> /etc/apk/repositories
2015-04-12 00:59:21 <owen1> and my guess is that's the reason that my image suddenly broke.
2015-04-12 03:46:43 <turvamies> Hello, Alpine world!
2015-04-12 03:47:08 <turvamies> Got some very persistent installation troubles
2015-04-12 03:48:44 <turvamies> In a nutshell, if I do a straight-forward setup-alpine script based installation, it works fine. However, when trying to go through "LVM on LUKS" (which is a *very* familiar setup for me) according to the instructions at http://wiki.alpinelinux.org/wiki/LVM_on_LUKS, I'm just getting "mounting root failed".
2015-04-12 03:49:34 <turvamies> The problem clearly has to do with SCSI drivers that do get loaded if I install according to the book, but obviously are not included (in the initramfs?) when going through the LVM on LUKS process.
2015-04-12 03:50:42 <turvamies> I've been working with similar setups using different Linux distros for over a decade and must admit that I'm just about to get frustrated. I've been trying to solve this problem for months already. I really do want to switch to Alpine but this single driver issue is preventing the transition.
2015-04-12 03:51:22 <turvamies> Also, I can't register on the project forum since I just can't figure out what the f*ck I'm supposed to answer to the "what is our project name" question. =D
2015-04-12 03:52:58 <turvamies> ...so any assistance would be most welcome
2015-04-12 03:58:57 <turvamies> These servers I'm working with are HP-Compaq gear that have Smart Array 5i and 6i controllers. The drive(s) is/are visible (in /dev/cciss) when booting from installation media, but when Alpine is trying to boot after the custom LVM on LUKS setup, there's no /dev/cciss at all. I've tried adding scsi-mod and cciss to the kernel modules line but I assume that does not help much if the necessary components are not compiled into t
2015-04-12 19:10:10 <Evil_Bob> alpinelinux.org seems to be down (sorry if its been said before)
2015-04-12 19:38:40 <ncopa> Evil_Bob: hi
2015-04-12 19:38:53 <ncopa> yes, i got sms and are trying to troubleshoot
2015-04-12 19:43:26 <tdtrask_> DNS resolution fails
2015-04-12 19:47:18 <ncopa> i think we found the cause
2015-04-12 19:48:04 <Evil_Bob> ah good :)
2015-04-12 19:48:19 <_ikke_> dns works here
2015-04-12 19:48:31 <_ikke_> but not ping
2015-04-12 19:48:34 <ncopa> should be fixed 1 min ago
2015-04-12 19:48:37 <_ikke_> ah ok
2015-04-12 19:48:40 <ncopa> ping is blocked i think
2015-04-12 19:48:43 <_ikke_> ok
2015-04-12 19:48:47 <_ikke_> too bad
2015-04-12 19:49:45 <_ikke_> dns service failed?
2015-04-12 19:49:58 <_mjones> nobody should block ping. It's a stupid practise with no advantage.
2015-04-12 19:51:02 <ncopa> re ping. its clandmeter's box, clandmeter's net so its up to him to block ping or not
2015-04-12 19:51:37 <_mjones> Agreed. I'll make a note to try to talk him out of doing so when he's next on.
2015-04-12 19:52:18 <ncopa> i would prefer ping enabled, but respect clandmeter's choice
2015-04-12 19:52:47 <ncopa> might be he enables it on request
2015-04-12 19:52:58 <ncopa> i have used ping enabled with rate limit
2015-04-12 19:53:16 <ncopa> max 6 pings per 10 sec or so from same ip
2015-04-12 19:53:34 <ncopa> ok dns works again
2015-04-12 19:53:38 <ncopa> i think go home again
2015-04-12 19:53:42 <ncopa> see u tm
2015-04-12 19:54:04 <_mjones> @ncopa: I have firm opinions on the matter in general, and would prefer to change his decision in general.
2015-04-12 19:54:45 <_mjones> rate limit is fine, great even. Blocking echo reply, time exceeded are super unhelpful, and blocking frag needed hurts you and other people too.
2015-04-12 20:39:52 <bOSKE> im trying to install cherokee but i get this error http://www.dodaj.rs/f/15/QW/4OUNoXxi/snapshot2.png
2015-04-12 21:31:55 <_mjones> BoSKE: it can't reach repos.
2015-04-12 21:32:51 <_mjones> b0SKE: paste your /etc/apk/repositories file after making sure something sensible is there and uncommented.
2015-04-12 21:37:40 <_mjones> b0SKE: as soon as you fix the repos everything will work (assuming you can reach the internet).
2015-04-12 21:49:28 <bOSKE> i cant copy paste in vb
2015-04-12 21:50:11 <bOSKE> _mjones, but i uncomented testing repo and do apk update then i tried to install cherokee
2015-04-12 21:50:33 <bOSKE> i have access to internet i downloaded some stuff
2015-04-12 21:51:10 <_mjones> Does apk update work?
2015-04-12 21:51:24 <_mjones> if so, and a URL is uncommented, it should just work.
2015-04-12 21:52:03 <_mjones> also do 'setup-apkrepos'
2015-04-12 21:52:14 <_mjones> It will do it for you automatically.
2015-04-12 21:59:18 <BitL0G1c> also pin the testing / edge repos
2015-04-12 22:00:30 <BitL0G1c> for a non dev box you will want http://hastebin.com/zoqefowipu.rb
2015-04-12 22:00:58 <BitL0G1c> if a package you want is in testing - ' apk add pkg@testing'
2015-04-12 22:01:22 <BitL0G1c> or @edge
2015-04-12 22:03:55 <BitL0G1c> is there a trick to making docker work ? - 'ApplyLayer exit status 1 chmod /bin/mount: permission denied '
2015-04-12 22:09:23 <BitL0G1c> nm - need to temporarily loosen grsecurity
2015-04-12 22:13:44 <bOSKE> _mjones, yes update work
2015-04-12 22:16:08 <bOSKE> _mjones, i made it i uncomented edge/main  and comented v3.1/main
2015-04-12 22:16:14 <bOSKE> it is installed now
2015-04-12 22:31:29 <mosez> my docker containers are running on edge and got a tag for testing :D
2015-04-13 02:02:53 <_mjones> ah, what was that mount line again? I think this may be working in edge and not in 3.1.3.
2015-04-13 02:04:37 <BitL0G1c> # mkdir /dev/mqueue
2015-04-13 02:04:37 <BitL0G1c> # mount -t mqueue none /dev/mqueue
2015-04-13 02:05:43 <_mjones> failed, no such device on 3.1.3. Didn't do any sysctls.
2015-04-13 02:05:59 <_mjones> Already in place on Edge, to be specific.
2015-04-13 02:06:24 <BitL0G1c> ok - many thx
2015-04-13 02:07:08 <BitL0G1c> root@alpine-build64 [/srv/lxc]# apk add docker@edge
2015-04-13 02:07:08 <BitL0G1c> (1/1) Upgrading docker@edge (1.4.1-r0 -> 1.5.0-r0)
2015-04-13 02:26:57 <owen1> is there a way to make sure my docker images are IDENTICAL? i feel that this line (might) install different things on my image each time i run it - RUN echo "http://dl-4.alpinelinux.org/alpine/edge/testing" >> /etc/apk/repositories
2015-04-13 02:27:35 <_mjones> so, testing repos get updated.
2015-04-13 02:28:10 <_mjones> So do regular. You have probably already seen io.js 1.6.4 in edge replacing 1.6.2, for instance.
2015-04-13 02:28:48 <_mjones> I suppose one can install by version.
2015-04-13 02:29:36 <owen1> _mjones: i didn't realize 1.6.4 is there! is there a website i can see all the packages in testing?
2015-04-13 02:33:00 <_mjones> owen1: http://git.alpinelinux.org/cgit/aports/tree/testing shows all the packages, and their versions are at the top of each respective package's APKBUILD file.
2015-04-13 02:33:22 <_mjones> This shows most recent activity, but old stuff scrolls off fast: http://git.alpinelinux.org/cgit/aports/
2015-04-13 02:34:56 <_mjones> I think this is what you want, once you know the package name: http://git.alpinelinux.org/cgit/aports/log/testing/iojs
2015-04-13 02:36:50 <_mjones> So I would go to aports/tree/, find the package in which I was interested in main/ or testing/ dirs, click on it, then click on 'log' at the top to see the change history of that package.
2015-04-13 02:37:30 <owen1> _mjones: thanks. looking at those links now
2015-04-13 02:38:11 <_mjones> looks like dropbear is one version out. Anyway...
2015-04-13 02:39:00 <_mjones> Reproductible builds. I think maybe you want to handle that at the Docker or image level. I'm not seeing a way to pin package versions or pull certain versions.
2015-04-13 02:45:18 <owen1> _mjones: when does testing become main?
2015-04-13 02:51:32 <_mjones> so you can run a stable version of the OS and pull things in from testing -- I just found out the syntax for the repositories file to do that, today.
2015-04-13 02:51:57 <_mjones> There's not necessarily a set schedule for moving to main. It happens continually, not everything at once.
2015-04-13 02:52:54 <_mjones> iojs is pretty well behaved. I don't know that there's any particular criterion for moving to main, or any criteria for not.
2015-04-13 02:54:33 <_mjones> nodejs is already in main. I don't know if ncopa waits for a certain time or what.
2015-04-13 03:01:33 <owen1> _mjones: cool. what i need is confidence that when someone pull my docker image from dockerhub, she gets the same exact image.
2015-04-13 03:01:41 <owen1> or as close as possible
2015-04-13 03:01:56 <owen1> right now it's not the case. my iojs was broken 2 days ago.
2015-04-13 03:02:23 <owen1> so i had to add apk update and apk upgrade lines to my Dockerfile to fix the issue.
2015-04-13 03:02:27 <_mjones> oh. So just download all of the things into your image and then save it, instead of having the script RUN and pull in the current versions of the things.
2015-04-13 03:02:55 <_mjones> I remember last time we spoke I didn't have my Docker machine, but I picked it up and upgraded it to Docker 1.5.0.
2015-04-13 03:03:36 <_mjones> I could have put Docker in a KVM guest but whatever. Nicer this way.
2015-04-13 03:10:11 <__number5__> docker-machine from docker is quite handy
2015-04-13 04:11:38 <owen1> _mjones: nice! can i see your dockerfile and how you use machine to run it? i never used machine before so i think that might force me to give it a spin.
2015-04-13 04:17:02 <_mjones> Clarification: I'm talking about an actual separate Mint-running laptop (machine) where I have Docker installed; I don't use Docker on my Debian machine because the prerequisites are so old.
2015-04-13 04:18:08 <_mjones> You do more with Docker than do I. I was planning on running some up soon but I see I'm rather near the bandwidth cap for this connection (LTE).
2015-04-13 04:32:42 <owen1> 
ACTION smells a blog post (:

2015-04-13 06:01:09 <ncopa> morning
2015-04-13 06:01:25 <ncopa> bOSKE: looks like cherokee needs rebuild
2015-04-13 08:30:22 <sillysausage> anyone having trouble getting to http://nl.alpinelinux.org/alpine/v3.1/main
2015-04-13 08:31:08 <sillysausage> im sure it worked for me the other day but i can't seem to get it now
2015-04-13 08:31:20 <_mjones> sillysausage: yes. it's timing out for me.
2015-04-13 08:32:29 <fabled> ISP issues in nl.a.o
2015-04-13 08:33:44 <sillysausage> ;ulz
2015-04-13 08:33:53 <sillysausage> dl-2.alpinelinux.org is hosted by Watchtower Bible and Tract Society of New York, Inc.
2015-04-13 08:33:58 <sillysausage> i wonder if they pay tax ;)
2015-04-13 08:36:48 <sillysausage> fabled: probably everyone trying to get game of thrones
2015-04-13 08:45:32 <owen1> I am trying to refresh my browser whenever a file was changed in my app. I tried adding -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix --privileged to the 'run' command but it didn't help. here is the repo, in case anyone is curious: https://github.com/oren/budo-template
2015-04-13 08:45:52 <owen1> not sure if it's alpine related or docker issue.
2015-04-13 08:52:08 <ScrumpyJack> what package provides vfat?
2015-04-13 08:53:21 <ScrumpyJack> ignore :)
2015-04-13 08:53:23 <_mjones> # apk info -W `which mkfs.vfat` /sbin/mkfs.vfat symlink target is owned by busybox-1.23.2-r0
2015-04-13 08:55:14 <ScrumpyJack> yup
2015-04-13 08:56:38 <ScrumpyJack> apk upgrade breaks the modules location symlink
2015-04-13 08:56:50 <ScrumpyJack> i forget every time
2015-04-13 08:57:43 <_mjones> apk upgrade has never done that for me.
2015-04-13 09:04:26 <bOSKE> ncopa, i made to install cherokee but now when i want to start cherokee-admin it reports some error on python http://www.dodaj.rs/f/M/Po/xRiVWqg/snapshot3.png
2015-04-13 09:04:34 <bOSKE> but cherokee web server works
2015-04-13 09:04:38 <bOSKE> as u can see
2015-04-13 09:09:17 <clandmeter> bOSKE: is this on 3.1?
2015-04-13 09:12:17 <bOSKE> clandmeter, no v3.1 repository is comented apk could not install cherokee when that reposityory was only uncomented there was some errors
2015-04-13 09:12:47 <bOSKE> www.dodaj.rs/f/15/QW/4OUNoXxi/snapshot2.png  that is the  error when i tried to install it with v3.1 reposititory
2015-04-13 09:13:41 <clandmeter> which repo do you use?
2015-04-13 09:13:44 <clandmeter> edge?
2015-04-13 09:15:23 <clandmeter> i just admin on edge which works
2015-04-13 09:15:32 <clandmeter> make sure you apk update && apk upgrade
2015-04-13 09:24:14 <bOSKE> clandmeter, im a noob i didnt upgrade :)
2015-04-13 09:24:26 <bOSKE> i apk upgrade and now it works :)
2015-04-13 09:42:25 <clandmeter> bOSKE: alpine linux is not very noob friendly :)
2015-04-13 09:44:32 <bOSKE> well yeah i like systems like alpine (i have arch on my  home server) but i thought how i didn't figure out alone to do upgrade :)
2015-04-13 09:53:57 <clandmeter> bOSKE: make sure to check out our wiki, it has lots of usefull information about our dist.
2015-04-13 09:54:02 <ncopa> someon messaged me this privately: http://pastebin.com/raw.php?i=fDneETha
2015-04-13 09:54:09 <ncopa> its how to get HDMI sound in alpine linux
2015-04-13 09:54:13 <ncopa> i think its pretty cool
2015-04-13 09:54:23 <ncopa> anyone want to test that and add it to the wiki?
2015-04-13 09:54:57 <clandmeter> i added something to aports recently
2015-04-13 09:55:10 <clandmeter> to switch between multiple sound cards
2015-04-13 09:55:48 <ncopa> linux-tools?
2015-04-13 09:56:14 <clandmeter> nah, thats a piece of crap :)
2015-04-13 09:56:25 <clandmeter> i need to remove that or fix it
2015-04-13 09:57:02 <clandmeter> http://git.alpinelinux.org/cgit/aports/commit/?id=536cedc4fa1d7118f43603a7cc34e06c0269bb74
2015-04-13 09:57:08 <ncopa> it was not too long ago i would have had use of know how to get sound out of hdmi
2015-04-13 10:00:13 <ncopa> asoundconf could also be added to a hdmi wiki article
2015-04-13 10:00:17 <ncopa> nice
2015-04-13 12:15:14 <kahiru> hey
2015-04-13 12:44:57 <Orochimarufan> Any idea why samba's interactive domain provisioning doesn't work?
2015-04-13 12:49:27 <clandmeter> Orochimarufan: probably a missing dep? what is the error msg?
2015-04-13 12:49:36 <Orochimarufan> Nothing
2015-04-13 12:49:39 <Orochimarufan> that's my problem
2015-04-13 12:49:50 <Orochimarufan> it just hangs or something
2015-04-13 12:50:04 <clandmeter> can you strace it?
2015-04-13 12:50:11 <Orochimarufan> using the cmdline switches and no --interactive seems to work
2015-04-13 12:51:17 <Orochimarufan> hmm
2015-04-13 12:52:00 <Orochimarufan> it seems it's waiting for input (readv) but it doesn't seem to be printing anything first...
2015-04-13 12:59:39 <Orochimarufan> strange thing is, it's using stock python2 print...
2015-04-13 13:05:25 <Orochimarufan> wouldn't a print call do a write() syscall?
2015-04-13 13:17:44 <ncopa> Orochimarufan: is it inside docker?
2015-04-13 13:17:56 <ncopa> or in a container? or vm or on the iron
2015-04-13 13:18:05 <ncopa> what kernel
2015-04-13 13:22:00 <Orochimarufan> LXC
2015-04-13 13:22:20 <Orochimarufan> on top of linux 3.19.3 iirc
2015-04-13 13:25:59 <Orochimarufan> Using samba-dc 4.2.0-r1 from edge
2015-04-13 13:34:36 <ncopa> do you have a way to reproduce it?
2015-04-13 13:34:40 <ncopa> simple way
2015-04-13 13:34:46 <ncopa> from clean envronment
2015-04-13 13:42:13 <Orochimarufan> clean environment being?
2015-04-13 14:09:44 <ncopa> apk add samba ...
2015-04-13 14:11:20 <Orochimarufan> which is what I have :O
2015-04-13 14:16:04 <ncopa> Orochimarufan: what command do i run to trigger the issue?
2015-04-13 14:16:14 <ncopa> do i need configure samba in any way?
2015-04-13 14:17:57 <Orochimarufan> I just ran "samba-tool domain provision"
2015-04-13 14:18:13 <Orochimarufan> it defaults to --interactive
2015-04-13 14:19:24 <ScrumpyJack> linux kernel 4
2015-04-13 14:20:24 <ScrumpyJack> so what's new?
2015-04-13 14:22:22 <clandmeter> the version number :)
2015-04-13 14:22:37 <Orochimarufan> Basically, yeah :P
2015-04-13 14:23:10 <Orochimarufan> It was literally a G+ poll that decided 4.0 over 3.10 -.-
2015-04-13 14:24:21 <Orochimarufan> ncopa: Could you reproduce it?
2015-04-13 14:27:39 <ncopa> seems so yes
2015-04-13 14:27:47 <ncopa> it is supposed to ask lots of questions i suppose
2015-04-13 14:28:00 <Orochimarufan> not quite "lots", but yeah
2015-04-13 14:28:36 <Orochimarufan> https://wiki.samba.org/index.php/Samba_AD_DC_HOWTO#Provisioning_The_Samba_Active_Directory
2015-04-13 14:29:20 <Orochimarufan> (It's quite doable without, but it's still a bug in my book^^)
2015-04-13 14:29:39 <ncopa> yes
2015-04-13 14:29:44 <ncopa> i can reproduce
2015-04-13 14:29:51 <ncopa> i dont know why it happens
2015-04-13 14:29:59 <Orochimarufan> Me neither >.<
2015-04-13 14:30:03 <ncopa> could you please file a bug for it?
2015-04-13 14:30:14 <ncopa> i will try get a chance to look at it later
2015-04-13 14:30:24 <Orochimarufan> right
2015-04-13 14:30:31 <ncopa> thanks
2015-04-13 14:30:48 <Orochimarufan> I might take a look at it myself, not sure if I'll spend too much time on it though
2015-04-13 14:33:40 <Orochimarufan> ncopa: What should I file against? the toplevel "Alpine Linux" project?
2015-04-13 14:34:35 <ncopa> yes
2015-04-13 14:34:41 <ncopa> category aports
2015-04-13 14:34:46 <Orochimarufan> right
2015-04-13 14:52:57 <Orochimarufan> Does this work for you? https://bugs.alpinelinux.org/issues/4061
2015-04-13 15:03:44 <ncopa> perfect
2015-04-13 15:03:45 <ncopa> thanks!
2015-04-13 15:04:15 <Orochimarufan> np. Glad to be of service ;)
2015-04-13 15:04:37 <ncopa>                  default = socket.getfqdn().split(".", 1)[1].upper()
2015-04-13 15:04:42 <ncopa> i guess it hangs there...
2015-04-13 15:04:58 <Orochimarufan> uuuh I might have caught it..
2015-04-13 15:05:07 <Orochimarufan> I say it's missing a flush
2015-04-13 15:05:12 <ncopa> ha!
2015-04-13 15:05:16 <ncopa> öfcourse
2015-04-13 15:05:27 <ncopa> i bet thats true
2015-04-13 15:05:37 <Diftraku> a flush and a rinse
2015-04-13 15:05:42 <Orochimarufan> Cause all the prompts end up after the thing is done
2015-04-13 15:05:44 <ncopa> we had similar issue with lvm
2015-04-13 15:06:55 <ncopa> http://stackoverflow.com/questions/230751/how-to-flush-output-of-python-print
2015-04-13 15:07:06 <ncopa> try add sys.stdout.flush() after print
2015-04-13 15:07:15 <ncopa> gotta go
2015-04-13 15:07:16 <ncopa> see u
2015-04-13 15:07:36 <Diftraku> always remeber to flush after you go
2015-04-13 15:07:42 <Diftraku> *remember
2015-04-13 15:08:30 <Orochimarufan> have a look at this :P https://drive.google.com/drive/folders/0B5I8UGgSqvvuYkZZTUpRcnVMNHM
2015-04-13 15:09:08 <Orochimarufan> (is there any way to capture stuff from a VT that doesn't involve a camera? lol)
2015-04-13 15:10:40 <Diftraku> a capture card/dev and another pc?
2015-04-13 15:11:22 <Orochimarufan> uhm... I think the camera is the less complex solution :O
2015-04-13 15:21:37 <Orochimarufan> ALso, actually posting the right link might help >.<
2015-04-13 15:48:30 <royger> unetbootin doesn't seem to be able to create a bootable USB stick from the Alpine Xen images, is there any workaround?
2015-04-13 16:24:14 <Orochimarufan> Uuuhm... why is edge bind9 built with --enable-threads and --disable-linux-caps?
2015-04-13 16:24:34 <Orochimarufan> And on top of that, it ships with an init script that includes -u
2015-04-14 08:50:43 <ScrumpyJack> morning
2015-04-14 09:20:54 <ncopa> hi
2015-04-14 10:12:05 <kocka> ohai DrWhax_
2015-04-14 10:12:12 <kocka> ahoi ncopa
2015-04-14 12:48:04 <Tekki> uh, so, is this an ok place to ask newbie questions? I'm having some trouble installing
2015-04-14 12:53:25 <stateless> Tekki: don't ask to ask, just ask
2015-04-14 12:55:37 <Tekki> right, I've got no clue about freenode customs. so, I'm trying to setup alpine as sys install on my laptop with a boot partition and a dm-crypted root partition, filesystem being btrfs and no lvm
2015-04-14 12:56:44 <Tekki> first I got a kernel panic due to broken init, then somehow fiddled with it yet init couldn't open my crypted root (mounting root failed, initramfs shell), decided to restart the whole process and now I'm getting the kernel panic again
2015-04-14 12:57:22 <Tekki> so I've been mostly figuring stuff out with arch wiki and this http://wiki.alpinelinux.org/wiki/LVM_on_LUKS
2015-04-14 12:58:57 <Tekki> is that wiki-info sensible and up-to-date? I've got three questions about that. 1) if I skip the lvm stuff, does that "# *** note mounts" row apply? I've been using /dev/mapper/<name>
2015-04-14 13:03:15 <Tekki> 2) does that init failing have to do something with mkinitfs errors about something /lib/modules/<kernel version> with the version being something-7 and "uname -r" giving something-6? basically, what is that "fix with a symlink" suggestion there, I haven't yet figured out what to use symbolic links for
2015-04-14 13:05:00 <Tekki> 3) haven't figured out why those dollar sign "$MNT/" commands haven't worked when just "/mnt/" has
2015-04-14 13:05:35 <Tekki> it's been a learning experience for sure, but after this many hours I just had to ask somewhere whether that info on the wiki is valid or should I do something in another way
2015-04-14 13:24:35 <ncopa> its very possible that the wiki is a bit outdated
2015-04-14 13:33:26 <darkfader> err, this is kinda funny to ask but
2015-04-14 13:33:27 <darkfader> ===>   openssl-1.0.2_1 depends on file: /usr/local/sbin/pkg - found
2015-04-14 13:33:27 <darkfader> => 0009-crypto-hmac-support-EVP_MD_CTX_FLAG_ONESHOT-and-set-.patch doesn't seem to exist in /usr/ports/distfiles//openssl-1.0.2a.
2015-04-14 13:33:30 <darkfader> => Attempting to fetch http://git.alpinelinux.org/cgit/aports/plain/main/openssl/0009-crypto-hmac-support-EVP_MD_CTX_FLAG_ONESHOT-and-set-.patch
2015-04-14 13:33:33 <darkfader> fetch: http://git.alpinelinux.org/cgit/aports/plain/main/openssl/0009-crypto-hmac-support-EVP_MD_CTX_FLAG_ONESHOT-and-set-.patch: Not Found
2015-04-14 13:33:39 <darkfader> freebsd seems now to include an alpine patch
2015-04-14 13:34:25 <darkfader> ah its now called '1001' :)
2015-04-14 13:43:46 <ncopa> what build system uses it?
2015-04-14 13:43:49 <ncopa> just curious
2015-04-14 14:01:49 <fabled> darkfader, yes, i renumbered and reordered the patches when openssl changed indentation
2015-04-14 14:02:00 <fabled> those while are also subject to change
2015-04-14 14:02:13 <fabled> use url with the git commit id to get permanant url
2015-04-14 14:03:47 <vkrishn> noticed, https://github.com/Microsoft/openssl ;)
2015-04-14 14:04:37 <vkrishn> and was watching Elysium, the other day, and realized can we make it proprietory on Mars !!
2015-04-14 14:08:58 <darkfader> fabled: i opened a pr that they should adjust the port
2015-04-14 14:27:56 <vkrishn> https://bugs.debian.org/cgi-bin/bugreport.cgi?archive=no&bug=734813 , :-))
2015-04-14 14:30:22 <darkfader> so much top-grade software
2015-04-14 15:02:02 <vkrishn> systemd: chomp..chomp.. Linux beware here we come to devour you!!
2015-04-14 15:02:02 <vkrishn> lxc: have mercy, spare us, we are just little guys and do no mischief outside our homes.
2015-04-14 15:12:00 <darkfader> docker: we have a page that says you should be security-aware and please click yes and then rock the world without worrying! yay!
2015-04-14 15:27:15 <BitL0G1c> lxc - we just need a bit of work to make unprivileged containers work - they are great when started by root (but run as user) - no cgmanager needed
2015-04-14 15:28:04 <BitL0G1c> I need to fix the segfault on shadow
2015-04-14 16:38:11 <vkrishn> systemd: and docker you ?
2015-04-14 16:38:11 <vkrishn> docker: (kneels in full compliance), I am just a puppeteer like you, big brother ! Spare me, and I promise to bind and serve lxc to behemoth you.
2015-04-14 16:43:35 <vkrishn> systemd: (scratches its dependency params in deep thought), serve well and thou shall live !
2015-04-14 17:54:54 <systmkor> has anyone worked with lxd?
2015-04-14 17:55:38 <systmkor> if so, what do think about it. I ran into some runtime errors most likely related to gettext library.
2015-04-14 18:29:14 <mosez> lol, got my alpine docker container started locally in my vagrant vm. curl: (6) Couldn't resolve host 's3.amazonaws.com'. from the vm and my host it works properly oO
2015-04-14 20:54:22 <systmkor> can you ping out?
2015-04-14 20:55:31 <systmkor> 1st local vagrant or w/e internal router then ping 8.8.8.8 (google dns server)
2015-04-15 00:33:24 <gad> Hi
2015-04-15 00:58:26 <systmkor> gad: hello
2015-04-15 01:00:32 <lrvick> Hey all, so I am trying to do 'apk --update add duplicity' and I am only getting 0.6.25 when http://pkgs.alpinelinux.org/package/main/x86/duplicity shows the latest in the registry is 0.7.02-r0
2015-04-15 01:00:45 <lrvick> how do i force alpine to pull the latest version?
2015-04-15 01:04:23 <BitL0G1c> apk update && apk add duplicity
2015-04-15 01:08:28 <systmkor> BitL0G1c: if that doesn't work would he need to also run apk upgrade if one of the dependencies of duplicity is locally already installed but not to the version needed by duplicity
2015-04-15 01:09:26 <BitL0G1c> yes true - you may want to pin repos http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning as occasionally you may want or need a package from edge
2015-04-15 01:09:57 <gad> I've been playing around with this distro and like it so far. Has anyone tried using Docker with it? (as a host)
2015-04-15 01:10:36 <BitL0G1c> yes - you need docker from edge
2015-04-15 01:11:06 <BitL0G1c> see repo pinning above
2015-04-15 01:11:30 <lrvick> this is a docker image, so any one-shot method is fine reaally
2015-04-15 01:11:30 <gad> I tried to build but got a permission denied error. I don't have the error details with me at the moment. I tried it with edge.
2015-04-15 01:11:49 <gad> I tried to run a hello world image.
2015-04-15 01:12:08 <BitL0G1c> there are also some settings needed in /etc/sysctl.conf
2015-04-15 01:12:32 <BitL0G1c> add:
2015-04-15 01:12:36 <BitL0G1c> kernel.grsecurity.chroot_deny_chmod=0
2015-04-15 01:12:36 <BitL0G1c> kernel.grsecurity.chroot_deny_mknod=0
2015-04-15 01:12:48 <BitL0G1c> & 'sysctl -p'
2015-04-15 01:13:18 <gad> Thanks! I'll give it a try the first chance I get.
2015-04-15 01:13:40 <BitL0G1c> also you need the linux-grsec in edge
2015-04-15 01:14:46 <BitL0G1c> if you see http://bugs.alpinelinux.org/issues/4010 - it's because you are not using the linux-grsec in edge
2015-04-15 01:19:06 <gad> Will do. Also just saw the repo pinning. Sweet! :)
2015-04-15 01:42:19 <ttilley> if anyone wants an update to llvm/clang, have at it - http://sprunge.us/QSBA
2015-04-15 02:02:12 <lrvick> FROM gliderlabs/alpine:3.1
2015-04-15 02:02:22 <lrvick> RUN apk update && apk add duplicity
2015-04-15 02:02:41 <lrvick> Still only get  (0.6.25-r0)
2015-04-15 02:02:46 <lrvick> When 7.x is available
2015-04-15 02:06:42 <lrvick> when In try :edge I do get latest duplicity
2015-04-15 02:06:56 <lrvick> but everything crashes with: Error relocating /usr/lib/libpython2.7.so.1.0: __flt_rounds: symbol not found
2015-04-15 02:08:08 <Diftraku> at least nl is not up to date
2015-04-15 02:08:19 <Diftraku> http://nl.alpinelinux.org/alpine/v3.1/main/x86/
2015-04-15 02:11:10 <Diftraku> same for dl-3, seems it is not yet pushed to main, despite being on package browser
2015-04-15 02:40:52 <systmkor> BitL0G1c: my b if I cam across pedantic, it's just I've had those occasional experiences which were definitely confusing to me earlier in my linux days
2015-04-15 02:45:15 <BitL0G1c> no problem ;-)
2015-04-15 02:45:56 <BitL0G1c> duplicity only really depends on python so just an apk update was needed
2015-04-15 02:46:42 <systmkor> has there been any suggestions on how to handle python2/python3 on the same system
2015-04-15 02:47:08 <systmkor> I'm guessing python2 as default then python3 as needed for other applications
2015-04-15 02:47:57 <Diftraku> seems most sensible for older stuff
2015-04-15 02:48:06 <systmkor> BitL0G1c: do you know if anyone has an automated build environment, to quickly test APKBUILDs on a clean install
2015-04-15 02:48:34 <Diftraku> but if you don't need stuff running python2, go with python3 from the start :3
2015-04-15 02:49:01 <systmkor> yup, sadly there are a fair amount of core modules that are still in the transition phase to python3
2015-04-15 02:51:14 <BitL0G1c> don't know about any automated testing
2015-04-15 02:52:12 <systmkor> m'kay, well if that's the cause I was thinking of building something
2015-04-15 02:53:08 <BitL0G1c> docker would probably be a good choice - I'm just about to start using it
2015-04-15 02:53:20 <systmkor> BitL0G1c: docker is definitely a good option
2015-04-15 02:53:35 <systmkor> was thinking of sticking with lxc at moment
2015-04-15 02:53:38 <systmkor> or work with both
2015-04-15 02:53:41 <systmkor> idk
2015-04-15 02:54:20 <BitL0G1c> i use LXC normally for alpine - going to be using docker for building debian packages
2015-04-15 02:55:48 <systmkor> I started trying to port LXD to Alpine however ran into some linking then runtime errors
2015-04-15 02:56:11 <BitL0G1c> debian in lxc is 240meg - a reusable docker debian is 90meg
2015-04-15 02:57:12 <BitL0G1c> http://container42.com/2014/11/18/data-only-container-madness/ - to store persistent data
2015-04-15 03:55:46 <systmkor> BitL0G1c: will check it out
2015-04-15 03:55:54 <systmkor> woo woo, got free Indian food for dinner
2015-04-15 03:57:26 <BitL0G1c> to run docker you need docker + linux-grsec in edge
2015-04-15 03:57:56 <BitL0G1c> & in /etc/sysctl.conf set:
2015-04-15 03:57:59 <BitL0G1c> kernel.grsecurity.chroot_deny_chmod=0
2015-04-15 03:57:59 <BitL0G1c> kernel.grsecurity.chroot_deny_mknod=0
2015-04-15 04:07:56 <systmkor> yarps, annoying about the chroot_deny
2015-04-15 04:07:58 <systmkor> parts
2015-04-15 04:08:49 <systmkor> BitL0G1c: I remember hearing about SELinux policies for Docker containers. Has anyon worked on grsecurity policies for Docker containers?
2015-04-15 04:40:59 <BitL0G1c> don't think so - don't know
2015-04-15 09:39:24 <ScrumpyJack> systmkor: hi
2015-04-15 12:20:35 <ScrumpyJack> I'm looking for a server host. does this look good? http://lowendbox.com/blog/joes-datacenter-555/
2015-04-15 12:21:13 <ScrumpyJack> i'm hoping some folks here would have more experience than me on prices etc
2015-04-15 18:56:55 <shodan45> I'm learning/toying with Ansible... its "user" module requires "useradd" & doesn't seem to work with busybox's "adduser"
2015-04-15 18:57:35 <shodan45> is there a package for useradd? or is adduser it?
2015-04-15 18:59:04 <shodan45> heh, actually... is there an easy way to get rid of busybox completely & use all the standard tools?
2015-04-15 19:01:43 <rsal> shodan45: apk add coreutils
2015-04-15 19:04:44 <shodan45> rsal: yeah, I know about that one, but it doesn't have useradd
2015-04-15 19:11:25 <shodan45> in opensuse, "shadow" owns useradd... ubuntu is just "passwd"
2015-04-15 19:11:43 <shodan45> neither of those seem to be in alpine :(
2015-04-15 19:13:17 <shodan45> btw, alpinelinux.org devs: the packages tool seems to break when going to page 2
2015-04-15 19:14:56 <clandmeter> shodan45: which uri?
2015-04-15 19:17:50 <shodan45> clandmeter: http://pkgs.alpinelinux.org/packages?page=2&arch=x86&repo=all&package=%tools
2015-04-15 19:28:28 <clandmeter> shodan45: thx. ill take a look at it
2015-04-15 19:38:35 <clandmeter> shodan45: try now
2015-04-15 19:57:30 <shodan45> clandmeter: works now :)
2015-04-15 19:57:45 <shodan45> clandmeter: was the %tools the problem?
2015-04-15 20:14:27 <BitL0G1c> shodan45 - shadow is in testing
2015-04-15 20:14:37 <BitL0G1c> it has useradd
2015-04-15 20:16:43 <BitL0G1c> http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning & 'apk add shadow@testing'
2015-04-15 20:17:50 <BitL0G1c> there is a problem with subuid / gid's in musl
2015-04-15 21:09:54 <shodan45> BitL0G1c: oh, duh... forgot about testing... thanks
2015-04-15 21:11:22 <clandmeter> shodan45: yes it was % problem, which wasnt correctly encoded
2015-04-15 23:21:40 <ttilley> ...is there an APK command to see everything in the entire repository that depends on a package/library?
2015-04-15 23:25:11 <darkfader> ttilley: not that i know of, for that kind of thing i would make a git clone of the aports repository and then grep through the dependencies (yes that is going to take longer than running a single apk command, but all i have to offer :)
2015-04-15 23:25:26 <ttilley> heh
2015-04-15 23:25:56 <ttilley> fair enough
2015-04-15 23:29:13 <clandmeter> ttilley: did you try pkgs.a.o?
2015-04-15 23:31:36 <ttilley> clandmeter: try what? (take my confusion to mean a "no")
2015-04-15 23:31:42 <ttilley> heh
2015-04-15 23:31:53 <clandmeter> which pkgs do you want to check?
2015-04-15 23:32:32 <clandmeter> http://pkgs.alpinelinux.org/packages
2015-04-15 23:33:09 <clandmeter> it should show required by for a pkg
2015-04-15 23:33:17 <ttilley> i want to see what depends on llvm-libs so i can locally rebuild them all against a newer llvm. i ported 3.6 to musl/alpine. the patch is here: http://bugs.alpinelinux.org/issues/4068
2015-04-15 23:33:33 <ttilley> neat
2015-04-15 23:34:21 <ttilley> it says required by 0, which is false.
2015-04-15 23:35:09 <ttilley> various mesa bits link against libLLVM-3.5.so
2015-04-15 23:35:16 <ttilley> at least
2015-04-15 23:35:17 <clandmeter> http://pkgs.alpinelinux.org/package/main/x86_64/llvm-libs
2015-04-15 23:35:28 <clandmeter> shows 7 matches for me
2015-04-15 23:35:41 <ttilley> ahh
2015-04-15 23:35:47 <ttilley> i was at llvm, not llvm-libs
2015-04-15 23:35:55 <ttilley> nevermind, i'm retarded :)
2015-04-15 23:35:57 <ttilley> thanks
2015-04-15 23:35:58 <ttilley> lol
2015-04-15 23:36:12 <clandmeter> im not saying its perfect though
2015-04-15 23:36:23 <ttilley> clandmeter: i wonder why this is a website and not a locally usable tool. it'd be very useful on the commandline
2015-04-15 23:36:32 <clandmeter> apk-tools has much more logic
2015-04-15 23:37:03 <ttilley> clandmeter: thanks for the link btw. i'm new to alpine. very new. like within the past few days new. :)
2015-04-15 23:37:14 <clandmeter> you can
2015-04-15 23:37:17 <clandmeter> apk info --help
2015-04-15 23:37:53 <clandmeter> atleast i think you can. never tried it.
2015-04-15 23:38:09 <ttilley> sweet.
2015-04-15 23:40:25 <clandmeter> i guess -r only works on packages which are installed.
2015-04-15 23:41:27 <clandmeter> i wonder why busybox has no "required by"
2015-04-15 23:42:01 <ttilley> to add my locally built packages to apk... do i just edit /etc/apk/repositories? any way to give local builds priority?
2015-04-15 23:42:21 <clandmeter> must be my "perfect" programing skills
2015-04-15 23:42:30 <clandmeter> yes and i think no
2015-04-15 23:42:50 <clandmeter> apk looks at version
2015-04-15 23:43:02 <clandmeter> not sure what happends when versions match
2015-04-15 23:45:40 <clandmeter> going to bed. gnite.
2015-04-15 23:45:45 <ttilley> cya
2015-04-16 00:48:18 <jmonreal> how can I install apache? i have tried apk add apache2-proxy but I got the folowing error: http://pastebin.com/NW1r4pT8
2015-04-16 00:48:24 <jmonreal> can anyone help me please ?
2015-04-16 00:57:38 <jmonreal> fixed :) i needed to do an update first
2015-04-16 03:07:58 <systmkor> ScrumpyJack: hey, I've been out today for finding applicants at a career fair. I don't know if I can help out tonight but message me what's up and I can email or message you
2015-04-16 07:19:05 <ScrumpyJack> systmkor: ok dude. what's your TZ?
2015-04-16 07:22:54 <ScrumpyJack> systmkor: is this about libseccomp?
2015-04-16 08:00:10 <ScrumpyJack> Bulgaria UTC +2?
2015-04-16 13:28:33 <crow> i do like to install alpinelinux vanille to an mSata SSD, but would like to make a less as posible write to that ssd what would be best to do? (no tmpfs install recomendation please, i already lost two times some stuff).
2015-04-16 13:30:31 <clandmeter> crow: why write less to msata ssd?
2015-04-16 13:31:14 <clandmeter> by default alpine will not write to the sd, except when installing and probably syslog.
2015-04-16 13:33:24 <crow> well to make lifetime of that ssd longer
2015-04-16 13:33:52 <clandmeter> its not like an usb or sd card
2015-04-16 13:34:05 <clandmeter> ssd's are made to write to them
2015-04-16 13:34:48 <clandmeter> it probably all depends on usage case
2015-04-16 13:35:38 <crow> well i have 4gb ram and some stuff to /tmp
2015-04-16 13:36:01 <clandmeter> 4g, you can probably go without swap
2015-04-16 13:36:51 <crow> its apu4c
2015-04-16 13:37:00 <clandmeter> i thought so
2015-04-16 13:37:01 <clandmeter> :)
2015-04-16 13:37:13 <clandmeter> i also have ssd in it
2015-04-16 13:37:44 <clandmeter> i expect it to last untill its too old to use anymore :)
2015-04-16 13:38:01 <crow> i bouht 30gb one
2015-04-16 13:38:27 <crow> Kingston 30GB SSDNow mSata 6Gbps
2015-04-16 13:39:02 <clandmeter> i think thats the same one as mine :)
2015-04-16 13:39:15 <clandmeter> it was cheap
2015-04-16 13:39:36 <crow> 34,19eur
2015-04-16 13:39:47 <crow> it is still laying
2015-04-16 13:39:56 <clandmeter> laying?
2015-04-16 13:41:42 <crow> well i need to poweroff device to setup everything thus downtime, need to do it at weekend. thus ssd is just near apu4c
2015-04-16 13:51:43 <crow> clandmeter lying arround :)
2015-04-16 13:51:52 <clandmeter> :)
2015-04-16 13:52:33 <clandmeter> downtime should be short
2015-04-16 13:52:45 <clandmeter> just reboot with the sdd insidie and trigger a hd install
2015-04-16 13:52:57 <clandmeter> then boot from the sdd
2015-04-16 13:53:13 <crow> wonder if i can trigger a sys install from SD card once booted?
2015-04-16 13:53:56 <clandmeter> it should install your current sd install to ssd
2015-04-16 13:56:29 <crow> i see
2015-04-16 14:03:01 <crow> to bad there is no edge vanille iso :/
2015-04-16 14:03:31 <crow> (it is not the firewall here)
2015-04-16 15:17:08 <clandmeter> crow: http://dev.alpinelinux.org/~clandmeter has vanilla iso from today
2015-04-16 15:50:22 <crow> thanks will try it :)
2015-04-16 16:12:22 <Jean-Scotch> Hi
2015-04-16 16:12:41 <Jean-Scotch> is there some secure-delete package of some sort?
2015-04-16 16:16:26 <fabled> Jean-Scotch, there's --purge to remove config files always
2015-04-16 16:16:45 <fabled> but there's no option to say "wipe the files from disk with military grade rewriting patterns"
2015-04-16 16:17:13 <Jean-Scotch> what I mean is a way to delete files on hdd with no way to recover deleted files
2015-04-16 16:17:26 <Jean-Scotch> ok
2015-04-16 16:20:48 <vok4o6atlSdZ> there is wipe in main and nwipe in testing
2015-04-16 16:27:29 <Jean-Scotch> thanks
2015-04-16 16:50:05 <BitL0G1c> Jean-Scotch: - I'm sending a patch for "secure-delete" https://www.thc.org/releases.php
2015-04-16 16:50:20 <Jean-Scotch> :)
2015-04-16 16:53:30 <BitL0G1c> do you use a desktop ?
2015-04-16 17:18:09 <Jean-Scotch> BitL0G1c: not yet. only servers. but my new desktop will be AL. (in a few days I hope)
2015-04-16 17:20:46 <BitL0G1c> there's a plugin for nautilus that uses the binaries in 'secure-delete' - am trying to build the plugin - it's used in TAILS
2015-04-16 17:22:35 <BitL0G1c> I also have  some patches to send to give copy / paste desktop functionality for KVM guests. It may also work with XEN LXC guests started with libvirt
2015-04-16 18:46:33 <ekristen> greetings
2015-04-16 18:46:50 <ekristen> I’d like to get the docker package for alpine updated, what is the best way to do that?
2015-04-16 18:48:44 <_ikke_> ekristen: update the APKBUILD, test it, and send a patch
2015-04-16 18:49:09 <ekristen> _ikke_: is there good docs someplace I can read on how to do those steps?
2015-04-16 18:49:51 <ekristen> _ikke_: where would I find the APKBUILD file?
2015-04-16 18:50:11 <_ikke_> This has some starting points: http://wiki.alpinelinux.org/wiki/Developer_Documentation
2015-04-16 18:50:32 <_ikke_> http://git.alpinelinux.org/cgit/aports/tree/main/docker
2015-04-16 18:51:18 <ekristen> ah that is what I was looking for
2015-04-16 18:52:11 <_ikke_> If it's a simple update, it's usually enough to bump the version and to regenerate the hashes
2015-04-16 18:52:21 <_ikke_> But test if the resulting package works properly
2015-04-16 18:52:23 <ekristen> _ikke_: it looks like the package is already 1.5.0 any reason why when I run apk add it is only installing version 1.4.1?
2015-04-16 18:52:36 <_ikke_> ekristen: cat /etc/apk_release
2015-04-16 18:52:41 <_ikke_> sorry
2015-04-16 18:52:45 <_ikke_> ekristen: cat /etc/alpine_release
2015-04-16 18:52:58 <ekristen> 3.1.3
2015-04-16 18:53:06 <ekristen> I just downloaded the iso from the site
2015-04-16 18:53:20 <ekristen> I’m obviousl new to alpine, love its concept though
2015-04-16 18:53:21 <_ikke_> right, this is master, which is the latest, not released version
2015-04-16 18:54:02 <ekristen> _ikke_: how do I update to that package?
2015-04-16 18:54:06 <ekristen> or can I?
2015-04-16 18:55:05 <_ikke_> You can get it from edge
2015-04-16 18:55:20 <ekristen> 
ACTION goes to google that

2015-04-16 18:55:57 <_ikke_> what is in your /etc/apk/repositories file?
2015-04-16 18:57:00 <ekristen> just v3.1 looks like edge is commented out, assume I can just uncomment, update and upgrade?
2015-04-16 18:57:22 <_ikke_> ekristen: I would do it differently
2015-04-16 18:57:29 <_ikke_> You can use repository pinning
2015-04-16 18:57:52 <ekristen> ok?
2015-04-16 18:57:55 <_ikke_> uncomment that line, but add "@edge " in front of it
2015-04-16 18:58:07 <_ikke_> http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning
2015-04-16 18:58:50 <ekristen> cool
2015-04-16 18:58:54 <ekristen> thanks for the link too
2015-04-16 18:59:11 <_ikke_> so you can do apk add docker@edge
2015-04-16 18:59:33 <_ikke_> that way, only docker will come from the edge repository, but other packages still come from the stable repository
2015-04-16 19:01:29 <BitL0G1c> if you want to run docker you need linux-grsec@edge too
2015-04-16 19:01:54 <BitL0G1c> otherwise there is no /dev/mqueue device
2015-04-16 19:02:31 <BitL0G1c> & add to /etc/sysctl.conf:
2015-04-16 19:02:36 <BitL0G1c> kernel.grsecurity.chroot_deny_chmod=0
2015-04-16 19:02:36 <BitL0G1c> kernel.grsecurity.chroot_deny_mknod=0
2015-04-16 19:02:51 <BitL0G1c> & then run 'sysctl -p'
2015-04-16 19:04:40 <ekristen> BitL0G1c: thanks!
2015-04-16 19:04:45 <ekristen> _ikke_: thanks
2015-04-16 19:05:24 <ekristen> I am playing around having alpine linux as my base host OS in my datacenter
2015-04-16 19:05:28 <ekristen> we run 99% in docker
2015-04-16 19:05:50 <ekristen> so it would be really nice to have lightweight secure host os
2015-04-16 19:09:57 <ekristen> hrm, unable to start docker container :/
2015-04-16 19:10:07 <ekristen> operation not supported
2015-04-16 19:10:07 <BitL0G1c> reboot
2015-04-16 19:10:31 <BitL0G1c> you should have installed the kernel from edge
2015-04-16 19:10:57 <ekristen> +1
2015-04-16 19:11:09 <ekristen> BitL0G1c: working nicely, thank you
2015-04-16 19:11:22 <_ikke_> BitL0G1c: thanks for chiming in :-)
2015-04-16 19:11:32 <BitL0G1c> ;-) np
2015-04-16 19:16:23 <ekristen> just out of curiousity
2015-04-16 19:16:47 <ekristen> why not use something like github to host the repos for packages so people can do simple Pull Requests vs sending patch emails?
2015-04-16 19:20:55 <BitL0G1c> LXC containers make a good development environment - it;s pretty simple to send patches by email - I just add an alias in ~/.profile for 'gitmail'
2015-04-16 19:24:31 <gra> Hi. I was asking a bit about docker here before. I got things set up thanks to BitL0G1c's help. Still a bit new to Alpine and Linux, but I assume given that I want to expose some of my containers's ports I would need to expose them not only between my machine and the network but between my machine and the docker containers as well?
2015-04-16 19:25:34 <_ikke_> gra: You need to publish the port from the container
2015-04-16 19:28:49 <gra> I assume when you mean publish you mean docker's -p option?
2015-04-16 19:29:15 <_ikke_> gra: yes
2015-04-16 19:30:47 <BitL0G1c> openvswitch for bridging may be helpful too - it makes networking simpler with LXC
2015-04-16 19:30:56 <gra> I am. I'm trying to reproduce this from another Linux distro but for some reason when I try to hit the endpoint it appears that it doesn't exist.
2015-04-16 19:31:46 <gra> Bookmarking openvswitch to take a look. Thanks.
2015-04-16 19:34:51 <BitL0G1c> install tcpdump in the container & listen on it's interface & the same on the container host - setup logging on iptables for both too
2015-04-16 19:35:12 <_ikke_> netstat to see if something is listening
2015-04-16 19:35:30 <BitL0G1c> netstat -lptun
2015-04-16 19:36:42 <gra> Sounds good. Will try that out asap.
2015-04-16 20:11:45 <crow> clandmeter this alpine-vanilla-150416-x86_64.iso from you is edge iso right?
2015-04-16 20:20:25 <ekristen> BitL0G1c: or _ikke_ I have a new package I want to add, I’ve done abuild -r and seems to pass, how do I do a test install?
2015-04-16 20:21:45 <BitL0G1c> the package will be ~/packages
2015-04-16 20:23:07 <BitL0G1c> the 'apk add ~/packages/x/xx/xxxx.apk'
2015-04-16 20:27:46 <ekristen> do I need to sign my commit?
2015-04-16 20:27:53 <ekristen> before I make the patch or anything?
2015-04-16 20:28:24 <BitL0G1c> if you followed the wiki instructions you would have generated a signing key
2015-04-16 20:28:42 <BitL0G1c> which will be in ~/.abuild
2015-04-16 20:28:50 <_ikke_> BitL0G1c: I think he means signing commits / patches
2015-04-16 20:29:00 <BitL0G1c> it's done automatically
2015-04-16 20:29:17 <ekristen> yeah I meant signed commits and patches, yes
2015-04-16 20:29:24 <ekristen> I do have the generated abuild key
2015-04-16 20:29:29 <BitL0G1c> you only really need the *.pub if you are going to install your own packages
2015-04-16 20:30:03 <BitL0G1c> put your ~/.abuild/*.pub in /etc/apk/keys
2015-04-16 20:30:39 <BitL0G1c> otherwise the apk will not install
2015-04-16 20:31:21 <BitL0G1c> the packages you generate are signed - the patch is just an email
2015-04-16 20:33:06 <ekristen> the commit message http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Commit_your_work am I suppose to use those variables or am I suppose to replace those variables with real info?
2015-04-16 20:34:35 <ekristen> BitL0G1c: ^
2015-04-16 20:34:48 <BitL0G1c> replace with real info not use $pkgurl $pkgdesc
2015-04-16 20:35:05 <BitL0G1c> normally i don't put those in as they are already in the email
2015-04-16 20:35:25 <BitL0G1c> in my ~/.profile I add:
2015-04-16 20:35:30 <BitL0G1c> alias gitmail='git send-email --to alpine-aports@lists.alpinelinux.org HEAD^'
2015-04-16 20:36:29 <BitL0G1c> http://wiki.alpinelinux.org/wiki/Creating_patches#Only_the_last_commit_with_.27git_send-email.27
2015-04-16 20:37:30 <ekristen> lol “git: 'send-email' is not a git command. See 'git --help’.”
2015-04-16 20:37:41 <BitL0G1c> to add the files to git normally I:     rm -rf pkg src      in the new aports directory - & 'git add .' (git add dot)
2015-04-16 20:38:07 <BitL0G1c> apk add git-email
2015-04-16 20:42:31 <ekristen> can I just format the patch and attach it to an email outside, I really don’t want to go through setting up email on the alpine system at the moment
2015-04-16 20:43:45 <BitL0G1c> yes
2015-04-16 20:46:19 <ekristen> FWIW as a new person to the project, but a very long time git user, I think that this project would get more visibility and more people contributing if you were to move things like the aports to github, to make forking, pull requesting and such much easier.
2015-04-16 20:47:48 <ekristen> with continuous integration like Travis, you could automatically build and test new PRs and vet them for proper compliance and provide near instant feedback to the person who submits the pull request, etc
2015-04-16 20:48:08 <ekristen> I just sent a patch for iojs, hopefully I did it right
2015-04-16 20:49:29 <BitL0G1c> to get up email add to ~/.gitconfig http://hastebin.com/exuzafowab.xml
2015-04-16 20:50:39 <BitL0G1c> I'm not the person to speak to about git - the other guys are around in CET - ncopa / clandmeter / stateless
2015-04-16 20:50:53 <ekristen> BitL0G1c: appreciate it, I’ll set it up for the next patch for that
2015-04-16 20:51:24 <ekristen> ncopa: clandmeter stateless ^^ see my comment above as new person to the project — just my perception and two cents,
2015-04-16 20:51:27 <ekristen> thanks BitL0G1c
2015-04-16 20:51:32 <BitL0G1c> i was going to write a script to make setting up a build environment a bit less hassle
2015-04-16 20:51:34 <BitL0G1c> np
2015-04-16 20:56:33 <ekristen> BitL0G1c: that would be handy for sure
2015-04-16 20:58:57 <BitL0G1c> I've written some scripts to make encrypted installs / custom installs a bit easier
2015-04-16 21:11:26 <BitL0G1c> http://it-offshore.co.uk/linux/21-linux/alpine-linux/25-alpine-linux-luks-encrypted-installations - they also support GPT disks
2015-04-16 21:20:06 <ekristen> nice
2015-04-16 21:20:30 <ekristen> I’m looking at using alpine as a virtual appliance for something I’m working on
2015-04-16 21:20:51 <ekristen> 100% of everything I need is docker containers, so if I can make a nice little base system using alpine, that would be awesome
2015-04-16 21:23:59 <BitL0G1c> docker runs with linux-grsec@edge & docker@edge
2015-04-16 21:24:37 <BitL0G1c> & in etc/sysctl.conf:
2015-04-16 21:24:41 <BitL0G1c> kernel.grsecurity.chroot_deny_chmod=0
2015-04-16 21:24:41 <BitL0G1c> kernel.grsecurity.chroot_deny_mknod=0
2015-04-16 21:25:43 <BitL0G1c> I'm hoping to also use it soon as a desktop (no systemd)
2015-04-16 21:28:05 <Risk64> When i do: 'apk add bash-4.3.30-r0' it says: ERROR: unsatisfiable constraints: \ bash-4.3.30-r0 (missing): \ required by: world[bash-4.3.30-r0]
2015-04-16 21:28:06 <Mo0O> hi there
2015-04-16 21:28:23 <Mo0O> Risk64: apk add bash
2015-04-16 21:28:27 <Mo0O> ;)
2015-04-16 21:28:32 <Risk64> ok.. :)
2015-04-16 21:28:36 <Risk64> thx
2015-04-16 21:28:38 <Mo0O> np
2015-04-16 21:29:02 <Risk64> but that was the string it gave me by 'apk search'.
2015-04-16 21:29:33 <Risk64> how should i know 'bash'?
2015-04-16 21:30:52 <Mo0O> the format is <pkg_name>-<pkg_version>-<aport_version>
2015-04-16 21:31:22 <Risk64> ok, so its always 1st segment then...
2015-04-16 21:31:28 <Mo0O> yes
2015-04-16 21:33:37 <BitL0G1c> you only need the full xxxxx-1234.apk if you are manually installing a package
2015-04-16 21:34:00 <BitL0G1c> that you built or downloaded
2015-04-16 21:42:49 <jomat> you can also tag your ~/packages/ to the repositories so i can install them as usual
2015-04-16 21:42:57 <jomat> s/tag/pin/
2015-04-16 21:45:06 <Risk64> i'm not so far yet.. :)
2015-04-16 22:05:44 <Mo0O> any idea why /etc/profile is not sourced?
2015-04-17 00:08:40 <jake__> On raspberry pi, filesystem changes do not persist through reboots
2015-04-17 02:10:46 <BitL0G1c> Mo0O: check 'env' - perhaps you need to log off & on again - or temporarily export VAR="xxxx" to add to the environment
2015-04-17 06:36:56 <uridium> moin
2015-04-17 07:45:43 <Psi|4ward> Anyone has NodeJS 0.12 running?
2015-04-17 07:49:51 <mosez> psi|4ward: :)
2015-04-17 07:50:02 <Psi|4ward> :O
2015-04-17 08:02:03 <Psi|4ward> official nodejs image VIRTUAL SIZE 700.6 MB
2015-04-17 08:02:06 <Psi|4ward> (facepalm)
2015-04-17 08:02:19 <mosez> ubuntu... welcome to the damn ubuntu docker world :D
2015-04-17 09:05:56 <kocka> moin
2015-04-17 14:39:01 <ekristen> morning everyone
2015-04-17 14:39:30 <ekristen> I submitted a patch yesterday to the dev list, is there any type of feedback as to whether or not it’ll be accepted?
2015-04-17 14:39:45 <clandmeter> hi
2015-04-17 14:39:48 <ekristen> hi clandmeter
2015-04-17 14:39:59 <clandmeter> you mention github yesterday
2015-04-17 14:40:12 <clandmeter> we do have a mirror on github if you like to send pr's
2015-04-17 14:40:19 <clandmeter> but most ppl just send it to the ml
2015-04-17 14:41:03 <ekristen> clandmeter: interesting, I’d have to image mainly because that is how the developer instructions are setup. Just give my perception as a new user for the project
2015-04-17 14:41:25 <clandmeter> ekristen: which patch did you send?
2015-04-17 14:41:29 <ekristen> iojs
2015-04-17 14:41:39 <clandmeter> somebody already replied to it
2015-04-17 14:41:43 <ekristen> :/
2015-04-17 14:41:47 <ekristen> 
ACTION goes to check email

2015-04-17 14:42:01 <clandmeter> its in testing repo
2015-04-17 14:43:20 <ekristen> odd I have no reply email
2015-04-17 14:44:19 <ekristen> good to know it was accepted into the testing repo, thanks! how does the process go to vet something from testing into main?
2015-04-17 14:44:40 <clandmeter> http://lists.alpinelinux.org/alpine-aports/0349.html
2015-04-17 14:51:00 <ekristen> odd clandmeter, thanks, I didn’t even think or know to check testing
2015-04-17 14:51:02 <ekristen> so I’ll do that now
2015-04-17 20:09:45 <darkfader> one of you know how i can pass parameters for the bonding module in alpine?
2015-04-17 20:09:50 <darkfader> i don't really remember
2015-04-17 20:10:01 <darkfader> i manually did echo 'layer3+4' >/sys/class/net/bond0/bonding/xmit_hash_policy now
2015-04-17 22:02:06 <knite> I'm having trouble running python. I've installed python 2.7.9-r5, and it refuses to start with error: Error relocating /usr/lib/libpython2.7.so.1.0: __flt_rounds: symbol not found
2015-04-17 22:02:40 <nsz> you have to update musl to 1.1.8
2015-04-17 22:04:01 <knite> ah, looks like I forgot an apk upgrade on edge/main before using that python. :-)
2015-04-17 22:05:35 <knite> nsz: but, there's another issue. I bumped up to this python because pip is supposed to be packaged by default with Python 2.7.9. this isn't the case for r0 in the 3.1 repo, and is still not the case for this r5, either.
2015-04-17 22:07:18 <nsz> i dont know about python packaging issues
2015-04-17 22:08:20 <knite> I'll drop a note in #alpine-dev
2015-04-18 18:44:10 <BitL0G1c> for a bootable usb - where should the apkovl live to be automatically detected ?
2015-04-18 18:48:20 <crow> BitL0G1c i think something like /media/usb/cache
2015-04-18 18:49:19 <BitL0G1c> ok - I think maybe i need to run 'setup-lbu' as well as 'setup-apkcache'
2015-04-18 18:52:51 <crow> BitL0G1c http://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Local_Cache
2015-04-19 13:40:19 <SVKroderik> on wiki: Hi can someone please do something with 2 YEARS OLD!!!!! CUPS article (its almost impossible to make it work right), half of article is obsolete (for example command lppasswd no longer working anymore).... thanks ...
2015-04-20 06:58:29 <ScrumpyJack> morning. quiet here
2015-04-20 07:12:31 <kocka> moin
2015-04-20 07:12:39 <kocka> yea, everything working fine :P
2015-04-20 09:15:32 <pierre76> hello
2015-04-20 09:16:48 <pierre76> I upgraded from alpine 2.7 to alpine 3.0 and then 3.1 but it seems the network driver alx is missing in the modules. something missing in the build ? how can i fix this ?
2015-04-20 09:45:36 <clandmeter> pierre76: which kernel and which arch?
2015-04-20 09:46:44 <clandmeter> Qualcomm Atheros seems like ARM?
2015-04-20 09:47:02 <clandmeter> oh atheros
2015-04-20 09:48:20 <clandmeter> pierre76: seems to be disabled in kernel config. please add a bug report so ncopa can add it.
2015-04-20 09:49:15 <pierre76> crap… thank for the info. Looket in GIT but seemed ok from what i found
2015-04-20 09:49:42 <pierre76> i’m 3.1.3 so kernel is 3.14.37 and it x86-64 i think
2015-04-20 09:49:47 <pierre76> thanks i’ll file the bug
2015-04-20 09:49:51 <clandmeter> dont know which kernel you use
2015-04-20 09:50:01 <clandmeter> but in edge grsec its not enabled
2015-04-20 09:50:12 <pierre76> ah
2015-04-20 09:50:18 <pierre76> i use a grsec with xen
2015-04-20 09:50:53 <clandmeter> http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec/config-grsec.x86_64#n2186
2015-04-20 09:53:30 <pierre76> thanks, this is what i had found : http://git.alpinelinux.org/cgit/aports/commit/?id=a9d4503b6ebe0ea683115cf9259d26ebda2d03c1
2015-04-20 09:55:04 <clandmeter> when you use 3.1, you should change branch to 3.1
2015-04-20 09:55:15 <clandmeter> http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec/kernelconfig.x86_64?h=3.1-stable#n2119
2015-04-20 09:58:57 <pierre76> thanks
2015-04-20 09:59:39 <pierre76> if i follow this page http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2015-04-20 09:59:55 <pierre76> should i be able to recompile the kernel myself ?
2015-04-20 10:00:12 <clandmeter> yes
2015-04-20 10:01:49 <pierre76> maybe i’ll try in the meantime, thanks for the help
2015-04-20 10:05:44 <ncopa> pierre76: did you file a bug?
2015-04-20 10:05:53 <pierre76> not yet, i will
2015-04-20 10:05:58 <ncopa> we are planning to update the v3.1 kernel
2015-04-20 10:06:02 <ncopa> soonish
2015-04-20 10:06:08 <ncopa> looks like a trivial thing to fix
2015-04-20 10:06:20 <ncopa> so I'd like to do it while there
2015-04-20 10:06:51 <pierre76> indeed , i will file in the hour
2015-04-20 10:32:02 <pierre76> http://bugs.alpinelinux.org/issues/4074
2015-04-20 11:04:33 <ncopa> thanks
2015-04-21 14:21:07 <ginjachris> hello all!
2015-04-21 14:24:46 <ginjachris> in their search for knowledge, some people might ask seemingly dumb questions
2015-04-21 14:24:52 <ginjachris> like this one:
2015-04-21 14:25:06 <ginjachris> how come I have two versions  of openssl installed, is this normal?
2015-04-21 14:25:08 <ginjachris> # apk info openssl openssl-1.0.2a-r1 description: Toolkit for SSL v2/v3 and TLS v1  openssl-1.0.2a-r1 webpage: http://openssl.org  openssl-1.0.2a-r1 installed size: 610304  openssl-1.0.1m-r1 description: Toolkit for SSL v2/v3 and TLS v1  openssl-1.0.1m-r1 webpage: http://openssl.org  openssl-1.0.1m-r1 installed size: 581632
2015-04-21 14:25:20 <ginjachris> 1.02 and 1.0.1m
2015-04-21 14:26:03 <ginjachris> would it be because I have both 'main' and 'edge' repo's enabled?
2015-04-21 14:26:16 <ginjachris> I guess I should only have edge enabled, right?
2015-04-21 14:32:27 <ncopa> that is probably due to both are enabled yes
2015-04-21 14:32:30 <ncopa> v3.1 and edge
2015-04-21 14:32:33 <ncopa> or similar
2015-04-21 14:32:41 <fabled> ginjachris, use --installed
2015-04-21 14:32:41 <ginjachris> I am dumb
2015-04-21 14:32:48 <ginjachris> just tested it to confirm it is the case
2015-04-21 14:32:49 <fabled> otherwise you get list of *available* ones
2015-04-21 14:35:06 <ginjachris> @fabled,  apk info --installed openssl just tells me "openssl" but, thanks to your guidance,  "apk info --installed --description openssl" gives "openssl-1.0.2a-r1 description: Toolkit for SSL v2/v3 and TLS v1"
2015-04-21 14:35:12 <ginjachris> thank you both :)
2015-04-21 14:35:45 <uridium> http://devopsreactions.tumblr.com/post/116986728409/coming-home-after-work
2015-04-21 15:08:11 <ginjachris> bye all, go to dash!  Keep up the excellent work!
2015-04-21 16:10:42 <kahiru> hey, is there a way to run puppet master on alpine?
2015-04-21 23:50:17 <ericflo> Anyone happen to have installed phash on alpine?
2015-04-21 23:50:24 <ericflo> I’m running into issues with https://github.com/gH0StArthour/pHash-0.9.6/blob/master/src/pHash.h#L60
2015-04-21 23:50:38 <ericflo> pHash.h:60:24: fatal error: sys/sysctl.h: No such file or directory
2015-04-21 23:51:34 <ericflo> Sounds like it’s a musl thing
2015-04-22 00:25:15 <BitL0G1c> ericflo - install linux-headers & patch the sources to include linux/sysctl.h
2015-04-22 00:37:13 <ericflo> ok will take a look at that, not too sure what all that entails
2015-04-22 01:04:21 <BitL0G1c> ericflo - find #include <sys/sysctl.h> in the header giving you the error - make a patch to replace it with #include <linux/sysctl.h>
2015-04-22 01:04:33 <yumbox> hello, is this the official alpine channel?
2015-04-22 01:12:33 <BitL0G1c> yes but most the guys are on European time
2015-04-22 01:14:11 <yumbox> i am actually in the netherlands
2015-04-22 01:14:23 <yumbox> but it's a bit late, haha
2015-04-22 01:17:00 <ericflo> BitL0G1c: thanks!
2015-04-22 01:17:58 <BitL0G1c> ;-)
2015-04-22 01:41:40 <uridium> 
ACTION is australian thank you very much!

2015-04-22 01:41:47 <uridium> 
ACTION types upside down for emphesis.

2015-04-22 01:47:47 <Diftraku> so, is small endian really big endian in the aussie?
2015-04-22 01:49:09 <uridium> Diftraku: Everything is big endian here.. except for immigrants.. often they're bi-endian but have aclimatised after a while.
2015-04-22 01:49:15 <uridium> you're right.
2015-04-22 01:49:25 <Diftraku> :3
2015-04-22 01:58:00 <uridium> :>
2015-04-22 02:11:23 <Mo0O> hi there
2015-04-22 02:14:51 <uridium> gruntings.
2015-04-22 02:16:23 <Mo0O> 
ACTION have found the solution when thinking about the question… rubber duck effect

2015-04-22 02:23:20 <uridium> 
ACTION wonders if the solution quacks appropriately

2015-04-22 02:27:51 <egrep> It should.
2015-04-22 02:31:24 <uridium> egrep: at least it won't echo.
2015-04-22 02:33:03 <egrep> uridium: Depends on where said rubber duck is. In certain theater-based bathrooms, the echo would be quite dramatic. In a squalid, curtain-covered, mattress stuffed bathroom, the echo would be verging on nonexistent.
2015-04-22 02:38:55 <uridium> egrep: perhaps .. but one should take care to have a duck-friendly environment and those with so much gruff in their flat shouldn't be blessed with the responsibility and care of a quacking duck in their bathroom. They are simply not worthy and Duck Abuse Services ought be called.
2015-04-22 02:43:36 <egrep> uridium: Places with echos are technically not a good, duck-friendly environment. As the duck quacks, the quack reverberates throughout and is fed back into the duck's ears, and the duck thinks it has friends. As the quacking duck searches for its quacking friends, it finds none, and obtains depression as a result of their achievement in finding nobody. Therefore, I say we unleash the D.A.S. upon all
2015-04-22 02:43:38 <egrep> occurrences of T.B.B.s: theaters, banks, and bathrooms.
2015-04-22 02:49:10 <uridium> You raise a valid point. General duck mental health ought be targeted for the populace. I suspect we need a global campaigne and an attempt to have "National Duck Health Day" .. the norwegians have a "national flightless duck day" .. so perhaps this can be rolled into one, single, global duck awareness and care day. Much like earth hour etc?
2015-04-22 02:56:40 <egrep> You make a well made argument for duck support, however, I fear for the well-being of breakfast, the most important meal of the day. My main quiet quandary with your questionable quest is that upon a surge of support for the safety and security of ducks, you have forgotten that omlettes are made from a lack of ducks. In a wild, hopeless outrage against those against ducks, omlettes will start to be
2015-04-22 02:56:43 <egrep> critisized, and a conundrum will occur. Shall we savour our breakfasts, or save our ducks? My greatest concern is that a breakfast food will be diminished, demolished, and in certain cases defenestrated out of diners across the world. I fear the omlette will cease to be, but to be fair, I am but an advocate for the devil himself.
2015-04-22 02:58:46 <uridium> Possibly the humble breakfast duck egg may be deminished but I see further complications from other ethnic minorities such as chickens who may feel they have a superior claim to the humble bum-nut than the ducks by sheer virtue of market penetration and global eating traditions. I suspect you're opening a can of worms there.
2015-04-22 03:05:43 <egrep> While worms may worm themselves into this matter, what matters is that ducks themselves prefer to think themselves as sole heirs to humble bum-nuts throughout, seeing as how they are a by-product of days upon days of such bum-nut heating. As a product, they feel they own the titles to that from which they came, and all of their possessions whence said bum-nut growers cease to be. But that is yet another
2015-04-22 03:05:46 <egrep> manifold of this conundrum, I fear that whichever one is chosen, either to save ducks, or to squander them, numerous consequences will ensue the likes of which have been seen before. I therefore would like to state that personally I feel the D.A.S. should be called upon itself to implode in a simple, quiet press-release. No matter which side the D.A.S. sides with, to abuse or to avoid abuse, chaos will
2015-04-22 03:05:48 <egrep> ensue, worms will crawl, make worm-holes throughout the dirt, and make the soil fertile with thousands upon thousands of arguments, and such. That is my opinionated, biased, brazenly decided stance.
2015-04-22 03:10:29 <uridium> I suspect this will all end in tears and D.A.S will be required to arbitrate between a number of powerful union bodies and hopefully not run afoul of either foul union's fixed agenda. They will no doubt have quite a fight on their hands from the Chickens by sheer size of membership .. and perhaps heavy tactics fromt he emu-laying union, though they mostly cater to the tourist and native carved bum-nut trade but some are eaten. 
2015-04-22 03:15:28 <egrep> I would like to add in that other foul fowl unions will try to intrude upon this volatile marketplace, to try to theatrically upset the balance in their favour, unsuccessfully. I fear this will all cause a surge in the bum-nut trade which will increase demand and yet supply will falter. I fear the days that chickens cease to be because of what we did this very day. Ducks should not be allowed to quack,
2015-04-22 03:15:31 <egrep> for quacking ducks make quarrels spout throughout the lands. The real issue behind this all is the quack. Silence the quack, I say, silence the quack.
2015-04-22 03:20:42 <uridium> Silence the quack? Are you daft man.. think of the civil liberties you attempt to impinge upon! Further, foreign trade will be an issue as prices fluctuate for the humble bum-nut. The price diffences in low socio-economic duck ponds will be far from chicken feed I can assure you. Even the poulty effects of a de-valuing of the humble duck bum-nut could seriously undermine global exchange rates!
2015-04-22 03:21:22 <uridium> 
ACTION ponders the exchange rate of duck to chickent to emu.

2015-04-22 03:26:35 <egrep> I never said to de-value the humble duck bum-nut, I only said to silence the quack! Indeed I may impinge upon some civil liberties, but then again so does censorship and yet that happens weekly if not daily! By sitting upon their bum-nuts, chickens manage to impinge upon the civil liberties of eggs. Merely to state that it impinges upon civil liberties has no meaning in this fluctuating bum-nut price
2015-04-22 03:26:38 <egrep> world, where those seeking a steady price for duck bum-nut are silenced by the facts. I never tried to de-value any bum-nut of any sort, I merely pondered to de-value the value in the duck produced from the humble duck bum-nut. The duck plays but a minor role in this cycle of events, and even then its value is diminished by very little. While you may try to save the ducks, the ducks need not to be
2015-04-22 03:26:40 <egrep> saved. They must be silenced. Silence the quack, I say repetetively and annoyingly. As for the exchange rate, it seems the market is a bit too unstable to see at this particular moment, as the very essence of worth of duck is being questioned. Perhaps all duck-related activities should cease to be for a while, to settle things down. I suggest we silence it all.
2015-04-22 03:37:29 <egrep> (as one egrep's word-play and knowledge decreases, the effects of accidental sleep deprivation begin to show)
2015-04-22 05:57:29 <bitshark_> hey
2015-04-22 08:59:00 <pierre76> hello everyone, just wanted to report back that the alx driver works like a charm on 3.1.3, thanks ncopa
2015-04-22 09:01:31 <pierre76> Also, FYI, i had trouble with XEN Pygrub, because it copies kernel and intrd from domU into /var/run/xen in order to boot the domU. This directory is now on tmpfs (with a size of 13MB on my system) with the implementation on /run on tmpfs.
2015-04-22 09:02:16 <pierre76> I notified xen-users about this (after they helped me sort it out)
2015-04-22 09:06:01 <ncopa> oh
2015-04-22 09:06:20 <ncopa> pierre76: do you have any suggestion how to properly fix that?
2015-04-22 10:30:09 <_ikke_> ncopa: Was wondering about the redmine docker container. How do you want to verify everything is working?
2015-04-22 10:42:42 <ncopa> _ikke_: i was thinking just open it in a web browser
2015-04-22 10:45:58 <_ikke_> ncopa: okay
2015-04-22 10:49:48 <ncopa> will anybody be sad if i purge mplayer?
2015-04-22 10:49:51 <ncopa> we have mpv now
2015-04-22 10:49:53 <vok4o6atlSdZ> !
2015-04-22 10:49:57 <vok4o6atlSdZ> i would be
2015-04-22 10:50:00 <ncopa> ok
2015-04-22 10:50:15 <ncopa> you like mplayer better than mpv?
2015-04-22 10:50:38 <vok4o6atlSdZ> no, rather i'm clueless about mpv.
2015-04-22 10:50:59 <ncopa> its a fork of mplayer i think
2015-04-22 10:51:03 <ncopa> but its better supported
2015-04-22 10:51:19 <vok4o6atlSdZ> letmegooglethat
2015-04-22 10:51:29 <ncopa> apk add mpv
2015-04-22 10:51:58 <ncopa> i think it does most (all?) things that mplayer does
2015-04-22 10:52:05 <_ikke_> ncopa: For the redmine package, why do you symlink routes.rb to /etc/redmine?
2015-04-22 10:52:14 <vok4o6atlSdZ> puh, i don't like the backwardscompatibility sound of this: "MPlayer's options parser was improved to behave more like other CLI programs, and many option names and semantics were reworked to make them more intuitive and memorable. "
2015-04-22 10:52:34 <ncopa> _ikke_: because its considered to be a config thing, which you might want modify
2015-04-22 10:52:49 <ncopa> _ikke_: at least that was what i thought then, but im not sure thats the case
2015-04-22 10:53:24 <_ikke_> ncopa: but makes it difficult to upgrade I think
2015-04-22 10:53:52 <vok4o6atlSdZ> does this mpv support framebuffer ops?
2015-04-22 10:54:28 <ncopa> vok4o6atlSdZ: i'd expect that, but i havent tested it
2015-04-22 10:55:11 <vok4o6atlSdZ> well, i guess alpine is a good reason to say goodbye to a lot of stuff i used before. farewell mplayer....
2015-04-22 10:57:00 <vok4o6atlSdZ> unrelated, but amusing rant on container security: http://www.vitavonni.de/blog/201503/2015031201-the-sad-state-of-sysadmin-in-the-age-of-containers.html
2015-04-22 11:00:13 <ncopa> lol
2015-04-22 11:00:20 <ncopa> Stack is the new term for "I have no idea what I'm actually using".
2015-04-22 11:00:25 <ncopa> lol
2015-04-22 11:03:12 <ncopa> well that kinda shows that we are not completely braindead who still try to package things with apk
2015-04-22 11:03:39 <ncopa> _ikke_: how does it make things difficult to upgrade?
2015-04-22 11:03:49 <ncopa> if its unchanged it will just be replaced
2015-04-22 11:04:08 <ncopa> if its changed you can use update-config to run a diff
2015-04-22 11:23:08 <Kruge_> hi
2015-04-22 11:56:42 <ScrumpyJack> hi
2015-04-22 12:06:49 <kocka> hi
2015-04-22 12:38:59 <ScrumpyJack> ncopa: thanks for mpv. 32bit mplayer was missing. 32 bit mpv is super :)
2015-04-22 12:58:55 <Kruge_> Hi.  I'd like to get a minimal graphical desktop running under alpine - just enough to run a web browser.  I tried installing the xfce-desktop, but it pulls down a world of packages.  Is there a good way to get just the bare necessities installed, please?
2015-04-22 13:12:08 <ScrumpyJack> Kruge_: you could try using a minimal window manager, like twm or i3
2015-04-22 13:14:31 <ScrumpyJack> try setup-xorg-base first
2015-04-22 13:14:38 <ScrumpyJack> than add a window manager
2015-04-22 13:39:57 <Kruge_> ScrumpyJack: Thanks for that.  I did the setup-xorg-base and then install xfce4 (without the desktop).  X seems to start up, but sits at a black screen.  I'll revisit this particular problem at a later date, I think.
2015-04-22 13:40:30 <Kruge_> In the meantime, is there a route to getting vmware tools or Virtualbox guest additions to install in an alpine vm?
2015-04-22 13:56:45 <Kruge_> nm, I appear to be an idiot today
2015-04-22 14:02:35 <BitL0G1c> Kruge_ - if you are using KVM don't use the Cirrus driver (it's broken & gives a black screen) - use the spice (QXL) driver in libvirt / qemu
2015-04-22 14:03:39 <BitL0G1c> soon spice will have the copy / paste functionality you get from vmware / vbox - I sent some patches on Monday
2015-04-22 14:16:28 <stef204> hi, exploring new linux distros (not using systemd) and came upon alpine. quick question: packages downloaded via apk are binaries, correct? if i want to compile on my own box, i can sync the aports tree and use/modify the scripts there, correct?
2015-04-22 14:16:47 <fabled> stef204, correct
2015-04-22 14:16:58 <stef204> fabled: ok, thanks
2015-04-22 14:39:00 <stef204> sorry, in the install process, after the ntp client "chrony" i am being asked which disk I would like to use; and the default preslected is [none]. This is for the disk formatting step, for a hard install correct? So choose the vbox hard disk, correct?
2015-04-22 14:42:18 <BitL0G1c> yes
2015-04-22 14:42:22 <stef204> it's fine; i chose sda and 'sys' and installing now
2015-04-22 14:42:41 <stef204> i missed the manual formatting of drive tho....
2015-04-22 14:42:45 <stef204> BitL0G1c: tx
2015-04-22 14:43:03 <BitL0G1c> yes - or 'setup-disk -m sys'
2015-04-22 14:43:47 <BitL0G1c> with the automatic configuration you've chosen the disks are formatted
2015-04-22 14:46:24 <stef204> BitL0G1c: I missed the opportunity to do setup-disk -m sys ... I guess I didn't really understand at what step I woud insert that; I guess after the ntp client?
2015-04-22 14:47:23 <BitL0G1c> not necessary now - you just chose "sys"
2015-04-22 14:48:20 <BitL0G1c> you should end up with a 100mb ext2 boot partition - a swap partition + rest of disk as ext4
2015-04-22 14:48:48 <stef204> I know but I'd like a /home partition, etc.
2015-04-22 14:50:12 <BitL0G1c> to do a custom partition scheme mount your partions somewhere & 'setup-disk -m sys $MNT'
2015-04-22 14:50:42 <BitL0G1c> or I've written some scripts for custom schemes http://it-offshore.co.uk/linux/21-linux/alpine-linux/25-alpine-linux-luks-encrypted-installations
2015-04-22 14:52:13 <BitL0G1c> or to do it manually http://wiki.alpinelinux.org/wiki/LVM_on_LUKS
2015-04-22 14:52:36 <stef204> BitL0G1c: ok, checking, thanks.
2015-04-22 16:52:38 <pierre76> @ncopa : the discussion went that mybe according to FHS they should put it in /tmp (but apparently FHS plans to put /tmp also in tmpfs)
2015-04-22 16:54:37 <pierre76> so…. other possibility, override their default location as I was told it is a compile time option …
2015-04-22 17:15:45 <_ikke_> ncopa: ok, I've got a docker file that builds the image for redmine
2015-04-22 22:52:19 <Risk64> Do i need to install 'gradm' to see logs or are there logs already somewhere if a socket got blocked?
2015-04-22 22:59:32 <BitL0G1c> if you are just trying to fix a problem - look at 'dmesg' or /var/log/messages
2015-04-22 23:08:54 <Risk64> BitL0G1c: Nothing mentioned there.. but my program claims 'Could not start media socket0'. I suspect grsec blocking sockets.
2015-04-22 23:09:55 <BitL0G1c> an easy way to confirm is install 'linux-vanilla' & boot the non grsec kernel
2015-04-22 23:11:00 <BitL0G1c> with VNC or Spice as it is not the default kernel - or change the default kernel to boot
2015-04-22 23:14:25 <Risk64> BitL0G1c: That wouldn't give me auditing logs. If there are no standard grsec deny logs i will install gradm.
2015-04-22 23:15:21 <BitL0G1c> grsec should show up in the logs
2015-04-22 23:15:54 <Risk64> ok, then its probably something different...
2015-04-23 00:03:04 <Norman__> How can I fix "mkdir: can't create directory '/usr/share/ntopng': Permission denied" when building a package with abuild
2015-04-23 00:17:29 <Norman__> How can I fix "mkdir: can't create directory '/usr/share/ntopng': Permission denied" when building a package with abuild
2015-04-23 00:24:30 <Risk64> Norman__: Can you manually create that folder?
2015-04-23 00:25:21 <Risk64> or in other words.. you have corresponding permissions..
2015-04-23 00:25:23 <Norman__> sure, why shouldn´t I
2015-04-23 00:26:03 <Norman__> abuild run as user ... so no access to that folders. anyway  - question is why does abuild not create the folder in pkg-dir
2015-04-23 00:26:46 <Norman__> did I just miss something or am I totally wrong?
2015-04-23 00:27:14 <Norman__> make install tries to install manpages etc. but failed becaus of permissions
2015-04-23 00:30:11 <Risk64> I can't help with that, sry. Didn't work with it yet..
2015-04-23 00:30:39 <Norman__> thx anyway :)
2015-04-23 00:31:27 <jomat> try sth. like: make DESTDIR="$pkgdir" install
2015-04-23 00:36:03 <Norman__> jomat: I already did
2015-04-23 00:43:12 <Norman__> jomat: it looks like abuild is ignoring the pkddir ... because, it also tries to copy the bin to /usr/bin/ instead of destfolder
2015-04-23 03:34:43 <jalateras> hey all. Just starting to use alpine-linux to deploy apps in docker containers.
2015-04-23 03:35:10 <jalateras> I am new to the package manager and was wondering whether there was a recipe for inatalling mongodb
2015-04-23 05:54:23 <jalateras>  /msg nickserv identify Plucker1
2015-04-23 07:54:10 <_ikke_> auch
2015-04-23 07:54:44 <_ikke_> jalateras: i'd change my password if I was you ;-)
2015-04-23 07:58:32 <jalateras> _ikke_: yeah already done :-)
2015-04-23 11:15:52 <ginjachris> greetings!
2015-04-23 12:10:41 <ScrumpyJack> is there a framebuffer image viewer in the repo?
2015-04-23 12:11:04 <ScrumpyJack> apk search --description doesn't work for me
2015-04-23 12:11:41 <vok4o6atlSdZ> if jfbview is testing, then yes.
2015-04-23 12:12:00 <vok4o6atlSdZ> dunno my packages seem to repell from alpine
2015-04-23 12:14:09 <ScrumpyJack> could find anything in the git repo APKBUILD description: field. I'll look for jfbv
2015-04-23 12:14:54 <ScrumpyJack> nothing. do you have a APKBUILD file?
2015-04-23 12:19:07 <vok4o6atlSdZ> sure
2015-04-23 12:20:42 <vok4o6atlSdZ> available in the next 60min: https://www.ctrlc.hu/~stef/jfbview.txz
2015-04-23 21:52:37 <rhapsodhy> what's the best way to set up a cross-compile env to armhf? the rpi2 is still not quite fast at compiling a kernel
2015-04-24 07:15:25 <looop> hello, i'm trying to install alpine-linux, i installed xorg, xfce, xf86-video-vesa, synaptics, evdev, when i do startx, a black screen comes up not xfce
2015-04-24 07:16:50 <ScrumpyJack> vok4o6atlSdZ: that build file doesn't work for me :(
2015-04-24 07:18:39 <ScrumpyJack> looop: have you tried setup-xorg-base?
2015-04-24 07:18:58 <ScrumpyJack> have a look at http://wiki.alpinelinux.org/wiki/Tutorials_and_Howtos#Desktop_Environment
2015-04-24 07:20:38 <looop> ScrumpyJack: yes i did
2015-04-24 07:20:45 <looop> that's how i did it
2015-04-24 07:22:48 <ScrumpyJack> vok4o6atlSdZ: missing mupdf-dev. do you have that too?
2015-04-24 07:23:48 <ScrumpyJack> you have the correct video driver?
2015-04-24 07:24:40 <looop> ScrumpyJack: yes, vesa should work
2015-04-24 07:25:41 <ScrumpyJack> which wiki article did you follow?
2015-04-24 07:26:08 <looop> the one you post it
2015-04-24 07:26:44 <ScrumpyJack> what's in your xinitrc file?
2015-04-24 07:27:28 <looop> exec startxfce4
2015-04-24 07:27:40 <ScrumpyJack> perhaps let a display manager do all the work. did you install one?
2015-04-24 07:28:38 <looop> i did the rc-service lxdm
2015-04-24 07:28:48 <looop> a gray screen came up
2015-04-24 07:29:06 <ScrumpyJack> grey is good :) that's X
2015-04-24 07:29:30 <ScrumpyJack> can you start any x client?
2015-04-24 07:30:09 <ScrumpyJack> maybe start a terminal then start your window manager to see if it works
2015-04-24 07:30:20 <ScrumpyJack> (in X)
2015-04-24 07:31:00 <looop> ScrumpyJack: sorry i have to go, i'll try that later on today, thanks for your help
2015-04-24 07:31:07 <ScrumpyJack> good luck
2015-04-24 07:31:12 <looop> thank you
2015-04-24 09:03:59 <vok4o6atlSdZ> ScrumpyJack: you need to add `subpackages="$pkgname-doc $pkgname-dev"` to testing/mupdf and rebuild it.
2015-04-24 11:26:01 <newuser474> Hello. i got Pentium 4, 216RAM, old Compqaq computer. Can i use Alpine in it ??????
2015-04-24 11:54:58 <BitL0G1c> newuser474 - I run a reverse proxy on a vps with 256m ram & alpine only uses about half of the ram
2015-04-24 11:58:48 <BitL0G1c> an alpine host with 4 alpine lxc containers uses 52 meg
2015-04-24 15:13:54 <Kruge_> You'll probably spend more on electricity running that thing than it would cost to buy a new Atom-based system
2015-04-24 15:33:11 <dalias> :)
2015-04-24 16:10:05 <_ikke_> 1
2015-04-24 17:46:38 <looop> ScrumpyJack: i'm back :)
2015-04-24 17:51:22 <anand123> Hello. I read that Alpine Linux is used for router, firewall and application gateway. Can it be used for personal daily use like Browsing, Music etc needs also???
2015-04-24 17:56:30 <looop> anand123: i would say yes you can use kde,xfce...
2015-04-24 17:56:32 <buckley310> anand123: it should work. it supports xorg, firefox, and vlc.
2015-04-24 17:56:58 <anand123> looop: buckley310 Thanks.
2015-04-24 17:57:20 <looop> anand123: yw, i'm stuck on xfce though lol
2015-04-24 17:57:54 <anand123> looop: how much RAM does Alpine consume when idle, in xfce ?
2015-04-24 17:58:51 <looop> anand123: i dont know, i installed xfce,xorg,xf86 drivers...but i cant start xfce
2015-04-24 17:59:29 <looop> i do startx, and a black screen comes up...not xfce
2015-04-24 17:59:43 <anand123> looop: does xorg, xfce needs install? Will it not work outta box?? Am a puppy  user.
2015-04-24 18:00:14 <looop> you have to look at the wiki, it's not difficult
2015-04-24 18:00:28 <looop> anand123: you are doing a real install?
2015-04-24 18:00:59 <anand123> looop: of course, by real i mean a USB install.
2015-04-24 18:01:34 <looop> oh ok, you better look at the wiki, i'm doing it in vbox
2015-04-24 18:02:34 <buckley310> xdm might work a little better than startx
2015-04-24 18:02:42 <anand123> looop: yes virtual box is surely a safe and good try
2015-04-24 18:03:25 <looop> buckley310: i did rc-service lxdm start, a grey screen comes up
2015-04-24 18:04:24 <buckley310> not sure about lxdm.
2015-04-24 18:08:05 <looop> buckley310: i type apk add xfce4.....it tells me 0K: 629Mib in 264 packages...does that mean i have it
2015-04-24 18:08:45 <BitL0G1c> lxdm - may be broken on 64 bit - is fine on 32 bit
2015-04-24 18:08:54 <BitL0G1c> on 64bit use alim
2015-04-24 18:09:00 <BitL0G1c> slim
2015-04-24 18:10:20 <looop> i never used slim, how i start it?
2015-04-24 18:16:00 <buckley310> does it not add an init script?
2015-04-24 18:17:55 <looop> i have no ida
2015-04-24 18:17:57 <looop> idea*
2015-04-24 18:18:39 <buckley310> find /etc/init.d | grep slim
2015-04-24 18:19:31 <buckley310> i can't actually test stuff myself at the moment :|
2015-04-24 18:20:31 <looop> buckley310: ty, no worry
2015-04-24 18:21:13 <looop> i did slim start
2015-04-24 18:21:32 <looop> i see alpinelinux in blue & white, it's blinking
2015-04-24 18:22:31 <looop> i dont have the mice though
2015-04-24 18:23:16 <buckley310> yeah i dont think slim has mouse support
2015-04-24 18:23:36 <buckley310> the login is all keyboard-driven, but the mouse will work after your desktop starts
2015-04-24 18:24:53 <buckley310> iirc
2015-04-24 18:25:55 <looop> what is alpine-dekstop?
2015-04-24 18:27:49 <buckley310> looks like a meta-package for xfce+lxde and other desktop-ish things
2015-04-24 18:29:02 <looop> ok, ty
2015-04-24 18:29:10 <buckley310> oh, something to consider is that i dont know if alpine packages adobe flashplayer
2015-04-24 18:29:41 <buckley310> oh anand left :\
2015-04-24 18:30:10 <looop> can anyone live without flash? lol
2015-04-24 18:30:15 <ahills> yes
2015-04-24 18:30:17 <buckley310> i almost can
2015-04-24 18:30:19 <ahills> I haven't used flash in two years
2015-04-24 18:30:29 <ahills> at work I literally refuse to use it
2015-04-24 18:30:30 <looop> youtube-dl ?
2015-04-24 18:30:47 <looop> html5 works fine with youtube
2015-04-24 18:30:50 <buckley310> imo youtube with html5 is acceptible quality at this point
2015-04-24 18:30:54 <buckley310> also vlc can open youtube links
2015-04-24 18:31:44 <buckley310> but google-music and spotify both depend on flash and shumway isnt really good at... well doing stuff yet
2015-04-24 18:33:09 <ahills> more reasons not to use those services...
2015-04-24 18:41:23 <buckley310> at work i actually need flash lol :\
2015-04-24 18:42:02 <ahills> my work thought I did
2015-04-24 18:42:06 <ahills> but I told them "no"
2015-04-24 18:42:08 <ahills> and now I don't
2015-04-24 18:42:10 <buckley310> lol
2015-04-24 18:42:33 <looop> hehe ahills, good for you
2015-04-24 18:42:37 <buckley310> i use windows at work so i figure whats one more
2015-04-24 18:42:49 <ahills> lol
2015-04-24 18:42:55 <looop> it does take a new way to use your pc not using flash
2015-04-24 18:42:56 <ahills> I also immediately wiped my windows install
2015-04-24 18:43:18 <ahills> I guess, I never really integrated flash into my "PC life"
2015-04-24 18:43:20 <ahills> so I have it easy
2015-04-24 18:43:59 <looop> i enjoy watching videos on youtube, i guess i dont have to use it, i can use youtube-dl which i do for long videos
2015-04-24 18:44:18 <ahills> yeah, youtube-dl is handy there
2015-04-24 18:44:20 <buckley310> if i switched back to pandora or actually started building a real music library i would uninstall flash at home. i use click-to-play so i know exactly where i need it
2015-04-24 18:44:52 <ahills> most of the music I listen to is available on bandcamp or similar services
2015-04-24 18:44:55 <ahills> so I just download it
2015-04-24 18:45:12 <looop> for music, i love clementine
2015-04-24 18:45:24 <ahills> on the mac?
2015-04-24 18:46:19 <looop> linxu
2015-04-24 18:46:22 <looop> linux*
2015-04-24 18:46:51 <ahills> oh, that's what I used for the few months I used a macbook
2015-04-24 18:46:58 <ahills> I just figured it was a mac-only thing
2015-04-24 18:47:02 <BitL0G1c> you don't really need a display manager with xfce - just 'startxfce4'
2015-04-24 18:48:04 <looop> i know, bit it's not working...dont know why, i can see slim but not xfce
2015-04-24 18:48:22 <buckley310> try using jwm instead of xfce
2015-04-24 18:48:47 <buckley310> jwm pretty much always works if X is working
2015-04-24 18:48:59 <BitL0G1c> also the cirrus (kvm) video driver isn't working
2015-04-24 18:49:34 <BitL0G1c> xfce works
2015-04-24 18:49:56 <looop> i tried xfce4, openbox, i'll try jwm :)
2015-04-24 18:49:59 <BitL0G1c> use the QXL (spice) driver
2015-04-24 18:50:07 <ahills> dwm!
2015-04-24 18:50:10 <ahills> dwm always works
2015-04-24 18:50:11 <buckley310> ive never seen openbox not work either...
2015-04-24 18:50:20 <buckley310> or dwm actually lol
2015-04-24 18:50:23 <buckley310> is that packaged?
2015-04-24 18:50:26 <ahills> is jwm the twm of 2015?
2015-04-24 18:50:39 <ahills> dwm packaged would be kind of silly...
2015-04-24 18:51:26 <HarryHaaren> but it happens :) Arch AUR for example. Awesome for a DWM with easier config
2015-04-24 18:51:49 <BitL0G1c> the black screen is a broken or missing video driver
2015-04-24 18:51:51 <buckley310> jwm is at least 4 years old. its basically the windows 2000 taskbar/WM
2015-04-24 18:52:34 <BitL0G1c> install the needed xf86-video-xxxx
2015-04-24 18:52:34 <looop> i didn't see dwm in it
2015-04-24 18:52:40 <looop> i did
2015-04-24 18:53:17 <ahills> yeah, something like AUR it makes sense, but prebuilt binary is only good as a stopgap
2015-04-24 18:54:07 <looop> i like dwm but i prefer i3
2015-04-24 18:54:13 <buckley310> Awesome is my favorite tiling window manager. i used it on my main desktop for about a year
2015-04-24 18:55:05 <ahills> is i3 the new wmii or am I thinking of something else?
2015-04-24 18:55:14 <ahills> I've been using dwm too long to switch, probably
2015-04-24 18:55:53 <looop> did you customised dwm?
2015-04-24 18:55:56 <buckley310> if you've got it set up already and never change it its probabbly as good as Awesome
2015-04-24 18:56:25 <ahills> yeah I have a config.h I ship around with me
2015-04-24 18:56:31 <looop> i3 only have 5 tabs, dwm has 9
2015-04-24 18:56:40 <ahills> you can change how many dwm has
2015-04-24 18:56:48 <looop> but i had more options in i3..for me at least
2015-04-24 18:57:01 <ahills> the tags are the reason I use dwm, I'm constantly recombining sets of windows
2015-04-24 18:57:04 <ahills> for my dev process
2015-04-24 18:57:41 <looop> it's fast and once you are use to it...2 click and you go
2015-04-24 18:58:15 <ahills> can you show multiple tags at once in i3?
2015-04-24 18:58:37 <looop> not that i know
2015-04-24 18:58:42 <ahills> ok
2015-04-24 18:59:37 <looop> i remember the first time i used dwm, i couldn't see firefox lol
2015-04-24 18:59:58 <looop> firefox open in tab 9 only
2015-04-24 19:00:38 <ahills> yeah, that's one of the things I configure away
2015-04-24 19:01:18 <looop> hey, i heard some people dont even use X in their desktop lol
2015-04-24 19:01:39 <buckley310> hardcore O_o
2015-04-24 19:02:59 <looop> i would do that only in a very old pc
2015-04-24 19:03:08 <looop> a dinosaur one lol
2015-04-24 19:03:54 <ahills> sometimes I don't feel like starting x
2015-04-24 19:04:01 <ahills> dvtm makes that easy
2015-04-24 19:04:04 <looop> tiling window managers will make use more terminals, it can be a good thing
2015-04-24 19:04:18 <ahills> on my laptop
2015-04-24 19:04:20 <looop> like i had my mail in terminal, i had cmus, i had irssi
2015-04-24 19:04:32 <looop> all in one tab, it was nice
2015-04-24 19:06:05 <looop> that's probably the closest to without X, ahills
2015-04-24 19:06:12 <buckley310> im probabbly the only one here who primarily uses gnome :$
2015-04-24 19:06:12 <ahills> x is nice though
2015-04-24 19:06:21 <ahills> tabbed + st + dvtm
2015-04-24 19:06:42 <ahills> are you on gnome 3?
2015-04-24 19:06:45 <ahills> the os x edition of gnome?
2015-04-24 19:06:50 <buckley310> yes
2015-04-24 19:06:57 <ahills> it's fine I guess
2015-04-24 19:07:04 <ahills> the keyboard shortcuts for window management are good
2015-04-24 19:07:07 <ahills> better than os x imho
2015-04-24 19:07:17 <ahills> but there's always a lot of shit I don't need on the screen
2015-04-24 19:07:28 <ahills> maybe on a big monitor I'd use it
2015-04-24 19:07:54 <looop> if your pc can handle it, it's ok
2015-04-24 19:08:11 <ahills> I put FC on my work computer which ships with gnome 3 by default
2015-04-24 19:08:19 <looop> it's a nice looking desktop
2015-04-24 19:08:21 <ahills> if I work for redhat I'll probably have to get used to it
2015-04-24 19:09:18 <looop> do you know about minix ?
2015-04-24 19:09:51 <ahills> yeah, looks pretty cool
2015-04-24 19:09:55 <looop> minix 3 actually
2015-04-24 19:10:05 <ahills> I thought about hacking on it and then I (re)discovered openbsd
2015-04-24 19:10:11 <ahills> which also has an elegant kernel
2015-04-24 19:10:18 <ahills> and works on more hardware
2015-04-24 19:10:25 <ahills> are you writing an OS?
2015-04-24 19:10:32 <looop> i discovered it about 2 months ago, it's a system that fix itself
2015-04-24 19:10:46 <buckley310> O_o
2015-04-24 19:10:47 <looop> it never stop
2015-04-24 19:10:54 <ahills> it's a research OS
2015-04-24 19:11:11 <ahills> a lot of really cool stuff that ends up in mainstream OSes starts in minix
2015-04-24 19:11:37 <looop> well, the european union gave him money to make it an OS
2015-04-24 19:12:00 <buckley310> what exactly does "fix itself"
2015-04-24 19:12:04 <buckley310> mean?
2015-04-24 19:12:04 <looop> like $3 M i believe
2015-04-24 19:13:06 <looop> fixing itself like, no bug, if some driver stop, it will communicate to another one that will wake him up
2015-04-24 19:13:34 <looop> a virus attack
2015-04-24 19:14:23 <looop> like for the nasa...they need an OS like this
2015-04-24 19:14:24 <HarryHaaren> yeah Minix runs "drivers" as "servers", not inside the actual kernel, but as a seperate "program" instead. This also has downsides though - so its not often used in industry/production environments
2015-04-24 19:14:29 <looop> 150% stable
2015-04-24 19:15:04 <looop> HarryHaaren: nicely said
2015-04-24 19:15:53 <looop> everyone would buy a system that is 150% stable
2015-04-24 19:16:01 <looop> i have no doubt about it
2015-04-24 19:16:16 <looop> even if it is slower
2015-04-24 19:16:36 <HarryHaaren> 
ACTION notes Awesome can show multiple tags just like DWM with rclick

2015-04-24 19:17:51 <looop> i never tried awesome nor ratpoison
2015-04-24 19:58:30 <ahills> looop: I don't think that's true; most of the systems I've been involved in, we've been happy with a watchdog that hard resets a server
2015-04-24 19:59:50 <alacerda> 
ACTION 0 x 10 roundcube

2015-04-24 20:01:50 <looop> ahills: but that would be the OS itself being able to do that
2015-04-24 20:02:30 <ahills> Right, but at the cost of performance
2015-04-24 20:02:47 <ahills> when you're already pushing hundreds of parallel systems to their limit, you can't afford to lose I/O
2015-04-24 20:03:14 <ahills> especially when having one of them go down every now and then doesn't dramatically affect your overall system
2015-04-24 20:04:16 <looop> it would be a system for one application...something about security...for now
2015-04-24 20:04:40 <looop> but could you imagine a system that stable
2015-04-24 20:10:06 <looop> a non breakable system :)
2015-04-24 20:10:25 <ahills> or a hard-to-break system, anyways
2015-04-24 20:10:37 <looop> right
2015-04-24 22:38:05 <bitshark_> hey
2015-04-24 22:38:37 <bitshark_> anyone think it's possible to get uclibc and glibc working side by side
2015-04-24 22:38:48 <bitshark_> but with dynamic symbol resolution for the linking
2015-04-24 22:39:25 <systmkor> it's possible but I don't know the technical solution off hand
2015-04-24 22:39:42 <bitshark_> ie. something linked against libc gethostbyname() in /lib64/lib.so gets redirected to uclibc or musl
2015-04-24 22:39:47 <bitshark_> yeah
2015-04-24 22:39:49 <bitshark_> me neither
2015-04-24 22:40:01 <systmkor> oh you mean has already been linked?
2015-04-24 22:40:08 <bitshark_> the only thing i can think is like a custom linker for source
2015-04-24 22:40:17 <bitshark_> or a kernel module for binaries
2015-04-24 22:40:20 <bitshark_> either one
2015-04-24 22:40:48 <bitshark_> but the general idea would be to support both recompilaton from source of stuff that's normally linked against glibc
2015-04-24 22:40:59 <bitshark_> but without having the user to edit makefiles etc
2015-04-24 22:41:27 <bitshark_> and on a more speculative level, have some way that ELF binaries linked against glibc are transparently resolved to musl / uclibc
2015-04-24 22:41:59 <bitshark_> anyway just a thought.... there's lots of posts in the alpine linux forums where ppl are asking about how to do this
2015-04-24 22:42:43 <dalias> what do you mean by "side by side"?
2015-04-24 22:42:53 <bitshark_> i see on  the musl web site there's a musl-gcc wrapper
2015-04-24 22:43:18 <dalias> the gcc wrapper is not used/useful on alpine
2015-04-24 22:43:23 <bitshark_> oh okay
2015-04-24 22:43:44 <dalias> it's just a way to compile/link against musl installed as a non-system libc using the existing gcc toolchain on your host
2015-04-24 22:43:53 <dalias> rather than having to have a new musl-targeted toolchain
2015-04-24 22:43:57 <bitshark_> i mean side by side in the sense like... i adapt alpine linux to be a bit more desktop friendly so i get the benefits of ditching glibc (smaller distro, faster , less security holes)...
2015-04-24 22:44:23 <dalias> go on
2015-04-24 22:45:57 <bitshark__> i guess i mean by 'side by side' is this:  (1) libc.so.6 is somehow 'universally compatible'   so no matter what something is linked against, as long as the symbols resolve it will compile and/or run
2015-04-24 22:46:17 <dalias> that doesn't make sense to me
2015-04-24 22:46:27 <bitshark__> okay maybe i'm misunderstanding
2015-04-24 22:46:31 <dalias> can you explain a specific example of something you'd like to see working
2015-04-24 22:46:33 <bitshark__> yeah oaky
2015-04-24 22:47:21 <bitshark__> so let's say i have a .deb package for like nmap or something
2015-04-24 22:47:26 <bitshark__> it's a binary compiled on debian or ubuntu
2015-04-24 22:47:41 <bitshark__> so i extract the ELF binary or whatever on alpine linux
2015-04-24 22:47:45 <bitshark__> and run ldd ./nmap
2015-04-24 22:47:48 <bitshark__> where nmap is the binary
2015-04-24 22:48:02 <bitshark__> pretend ldd returns the following
2015-04-24 22:48:22 <bitshark__> $ ldd ./nmap  .... libc.so.6 => /lib64/libc.so.6 (0x00007f42a9bc8000)
2015-04-24 22:48:57 <bitshark__> and it's calling one external symbol...  let's say nmap calls libc gethostbyname()
2015-04-24 22:49:37 <bitshark__> that's it..   im wondering if there's some clever way to get dynamic libraries for glibc to transparently redirect to alpine's musl / uclibc
2015-04-24 22:49:51 <bitshark__> maybe that's impossible if the external symbols are stripped
2015-04-24 22:50:02 <bitshark__> i don't know
2015-04-24 22:50:37 <bitshark__> so that's the first question of side by side  (meaning this function either points nmap at glibc or it somehow remaps it to gethostbyname() in uclibc/musl )
2015-04-24 22:51:13 <bitshark__> the second is the same problem, but pretending now we are using nmap source package  .. again where it calls glibc gethostbyname()  as a dynamic library
2015-04-24 22:51:33 <bitshark__> maybe in this latter case we can call musl-gcc wrapper to compile it and that solves our problem
2015-04-24 22:52:17 <bitshark__> i don't know but that's the two scenarios i'm curious about...   getting stuff normally linked against glibc to  work on alpine automatically.. the first case is the binary elf case, the second case is the source code case
2015-04-24 22:53:01 <dalias> i think you have a serious misunderstanding of a lot of things
2015-04-24 22:53:04 <bitshark__> okay
2015-04-24 22:53:11 <dalias> have you _tried_ any of this?
2015-04-24 22:53:30 <dalias> nmap is not going to have onf external symbol reference but tens if not hundreds
2015-04-24 22:53:31 <bitshark__> yeah i tried compiling something linked against glibc against musl and it didn't work
2015-04-24 22:53:35 <bitshark__> oh i know
2015-04-24 22:53:49 <dalias> "compiling something linked against glibc against musl" that phrase does not even make sense
2015-04-24 22:53:51 <bitshark__> i just picked one for a thought eperiment
2015-04-24 22:53:55 <dalias> you don't compile something that's already linked
2015-04-24 22:53:58 <bitshark__> okay
2015-04-24 22:54:00 <dalias> compiling happens before linking
2015-04-24 22:54:13 <bitshark__> right
2015-04-24 22:54:47 <dalias> if what you're asking is if you can take the nmap binary from a .deb and run it on alpine using musl, the answer is "maybe"
2015-04-24 22:55:11 <dalias> you can either ask the dynamic linker to load and try to run it as:
2015-04-24 22:55:27 <dalias> /lib/ld-musl-x86_64.so.1 ./nmap ...
2015-04-24 22:56:19 <bitshark___> question: "given that alpine linux has ditched glibc for security reasons and bloat reasons ,what disadvantages does this present in terms of interoperability with software that is normally built against traditional gnu libc  ?  are there any problems that might arise in terms of compiling source code build and tested against gnu libc , or is this a non issue entirely when compiling from source ? "
2015-04-24 22:56:42 <nsz> abi compat
2015-04-24 22:56:43 <dalias> if you're compiling from source then glibc is irrelevant
2015-04-24 22:56:55 <bitshark___> okay .. interesting
2015-04-24 22:57:21 <dalias> there may be buggy or non-portable source that's dependent on properties of glibc
2015-04-24 22:57:39 <bitshark___> well i ran into some problems compiling linux from scratch (lfs) stuff on alpine's musl libc
2015-04-24 22:57:41 <dalias> that's unlikely to find in portable cross-platform apps that are also used on bsds, etc.
2015-04-24 22:57:46 <bitshark___> but that might be an edge case
2015-04-24 22:57:48 <nsz> are you running binaries or compile source? i'm confused..
2015-04-24 22:57:50 <bitshark___> not a normal case
2015-04-24 22:58:06 <nsz> alpine itself compiles on alpine
2015-04-24 22:58:08 <bitshark___> no , problems with linux from scratch ... so actually compiling a toolchain
2015-04-24 22:58:09 <dalias> but it happens more often with "linux only" software
2015-04-24 22:58:15 <nsz> and there are a lot of packages there..
2015-04-24 22:58:16 <bitshark___> yeah
2015-04-24 22:58:57 <dalias> for toolchain stuff you want to get the patches from the musl-cross repo. nsz is working on upstreaming them so they'll probably be upstream in the next gcc or so
2015-04-24 22:59:32 <bitshark___> so essentially i think the answer to what i'm asking is like the following (A) normally stuff built from source will work fine against musl , but there might be some edge cases that fail.   most will compile okay .   and (B) the idea of having dynamically linked binaries against glibc run natively on alpine is so difficult as to be impossible... just recompile from src
2015-04-24 23:00:04 <bitshark___> because like u said
2015-04-24 23:00:05 <nsz> not so difficult
2015-04-24 23:00:11 <bitshark___> nmap will have hundreds of external symbols
2015-04-24 23:00:49 <bitshark___> and any  nmap  ELF binary with dynamic libraries (specifically glibc) ... pointing to /lib/libc.co.6 will be problematic
2015-04-24 23:00:56 <bitshark___> on alpine
2015-04-24 23:01:10 <nsz> you can symlink all glibc libraries to libc.so
2015-04-24 23:01:14 <bitshark___> really?
2015-04-24 23:01:18 <nsz> and the glibc loader to the musl loader
2015-04-24 23:01:19 <bitshark___> okay cool
2015-04-24 23:01:25 <nsz> it is not guaranteed to work
2015-04-24 23:01:27 <bitshark___> yeah
2015-04-24 23:01:38 <bitshark___> i didn't even realize u could do that
2015-04-24 23:01:43 <nsz> because musl is not completely abi compatible with glibc
2015-04-24 23:01:55 <bitshark___> okay yeah ABI is the part where i get lost
2015-04-24 23:02:11 <bitshark___> that was where it was failing on compiling linux from scratch toolchain....
2015-04-24 23:02:25 <bitshark___> something to do with ABI something or other that was missing... so that makes sense
2015-04-24 23:02:26 <nsz> compiling has nothing to do with this
2015-04-24 23:02:43 <nsz> linux from scratch can fail for all sorts of reasons
2015-04-24 23:02:49 <bitshark___> i mean compiling as the process
2015-04-24 23:02:59 <bitshark___> well in this case it was specificly to do with alpine
2015-04-24 23:03:19 <dalias> <nsz> you can symlink all glibc libraries to libc.so
2015-04-24 23:03:20 <dalias> no need
2015-04-24 23:03:22 <nsz> (eg alpine has busybox sed which cannot run some gnu sed scripts used in buildsystems)
2015-04-24 23:03:24 <bitshark___> alpine and something to do with libc... but that's fine i just compiled lfs on a different box
2015-04-24 23:03:26 <dalias> they're automatically handled
2015-04-24 23:03:44 <dalias> and making the symlinks can be problematic when linking new programs, so it's a bad idea
2015-04-24 23:04:21 <dalias> linux from scratch process is HIGHLY over-complicated
2015-04-24 23:04:25 <bitshark___> okay , what u guys are saying makes sense .
2015-04-24 23:04:29 <bitshark___> oh no worries about lfs
2015-04-24 23:04:34 <bitshark___> i got it working fine
2015-04-24 23:04:42 <dalias> in fact i would say one of the big motivations for switching away from glibc is NOT having to jump thru all the hoops LFS does to get results
2015-04-24 23:04:47 <dalias> like compiling your toolchain 3 or 4 times
2015-04-24 23:05:54 <nsz> recently learnt that the debian bootstrap process involves compiling x11 and gtk libraries.. so lfs is not the worst :)
2015-04-24 23:06:03 <bitshark__> i was just trying to understand if the problems i saw linking lfs stuff on alpine would extend to normal day-to-day issues
2015-04-24 23:06:38 <nsz> we still dont know what happened with lfs
2015-04-24 23:06:47 <bitshark__> okay
2015-04-24 23:06:51 <nsz> (i thought you wanted to get nmap work)
2015-04-24 23:06:55 <bitshark__> no no
2015-04-24 23:07:02 <bitshark__> damn i wished i hadn't picked that example haha
2015-04-24 23:07:29 <dalias> nsz, :)
2015-04-24 23:07:43 <dalias> heh
2015-04-24 23:07:43 <bitshark__> that would just a made-up example for expediency  .. a fake 'thought experiment' nmap with only one dynamically linked library and one external symbol ...
2015-04-24 23:08:01 <dalias> it helps to have real examples
2015-04-24 23:08:16 <bitshark__> let me boot up the alpine virtual machine again and try some other stuff
2015-04-24 23:09:09 <bitshark__> but i think the core issue or question i'm asking was like  what does glibc provide that's not in musl
2015-04-24 23:09:38 <bitshark__> if anything
2015-04-24 23:10:19 <nsz> (bugs)
2015-04-24 23:11:43 <bitshark___> i think actually the problems with lfs had to do with the fact that some of the packages need g++... "Be aware that, "out of the box", the wrapper only supports C applications, not C++. This is because the C++ libraries and headers are missing from the musl include/library path"
2015-04-24 23:11:49 <bitshark___> http://www.musl-libc.org/faq.html
2015-04-24 23:14:26 <dalias> bitshark___, yes some things
2015-04-24 23:14:43 <dalias> many are things you probably don't want
2015-04-24 23:15:02 <bitshark___> yeah
2015-04-24 23:15:10 <dalias> but some of those might be needed by certain apps or setups anyway even if they're not good; most can be provided by separate libs instead
2015-04-24 23:15:15 <dalias> like nfs related stuff needs sunrpc
2015-04-24 23:15:27 <dalias> that's provided by the external tirpc package on alpine
2015-04-24 23:15:52 <dalias> (glibc has also been trying to get rid of rpc for a long time but they can't because stuff depends on it :/)
2015-04-24 23:16:48 <bitshark_> geez i need to reboot this router heh
2015-04-24 23:17:08 <bitshark_> but thanks delias, that makes sense... i dislike the gnu stuff generally
2015-04-24 23:17:39 <bitshark_> i'm just trying to understand how i can make a sick dev workstation out of alpine linux
2015-04-24 23:18:24 <dalias> well i don't think you need to LFS anything to do that
2015-04-24 23:18:26 <bitshark_> but it seems like i just picked a bad example (building a cross compiler) to test compatibility
2015-04-24 23:18:35 <bitshark_> probably anything else would have worked out of the box
2015-04-24 23:18:48 <bitshark_> oh the LFS stuff is a different project
2015-04-24 23:18:51 <dalias> ah ok
2015-04-24 23:20:02 <bitshark_> yeah the LFS stuff i was just making sure that there's ways to keep systemd out of the toolchain
2015-04-24 23:20:23 <bitshark_> ensure i can boot a minimal linux system with the latest kernel with sysvinit and not systemd
2015-04-24 23:20:53 <bitshark_> so that's seperate from alpine (since alpine linux already does this, heheh)
2015-04-24 23:47:15 <dalias> anyone know how to get adb (android util) on alpine?
2015-04-24 23:54:42 <dalias> haha awesome debian has a flat makefile for adb :)
2015-04-24 23:54:58 <dalias> to work around the idiotic Android.mk stuff
2015-04-24 23:55:16 <dalias> i bet there's a programmatic way Android.mk could be converted to flat makefiles
2015-04-25 05:45:17 <bitshark_> okay awesome
2015-04-25 05:45:32 <bitshark_> got gtk and all that crap built against mucl
2015-04-25 05:46:07 <bitshark_> alpine could definitely be a drop in replacement for ubuntu 15 without systemd
2015-04-25 05:46:26 <bitshark_> got mate compiled with gtk
2015-04-25 16:52:46 <dalias> alpine might be interested in this: https://twitter.com/canadianbryan/status/591997995459051520
2015-04-25 17:22:19 <S_K> Hello.
2015-04-25 20:04:00 <rivarun> hello. i'm trying alpine linux out in virtualbox. i installed awesome, xinit, xorg, xf86-video-vesa, xf86-input-{mouse,keyboard}. when i run `startx`, i get the awesome logo and a mouse cursor (but no status bar at the top). i can't click or do anything until i reboot the vm. any ideas about this?
2015-04-25 20:05:10 <vok4o6atlSdZ> did you run setup-xorg-base?
2015-04-25 20:05:16 <vok4o6atlSdZ> it replaces your udev
2015-04-25 20:05:27 <rivarun> vok4o6atlSdZ: no i didn't, since X does come up
2015-04-25 20:05:49 <rivarun> oh. awesome doesn't work with mdev?
2015-04-25 20:06:05 <vok4o6atlSdZ> no, it's xorg which needs this for the devices
2015-04-25 20:06:14 <vok4o6atlSdZ> input devices
2015-04-25 20:06:50 <rivarun> well, i just ran it but no luck. it seems to me that awesome gets stuck since it doesn't render the top bar
2015-04-25 20:07:16 <BitL0G1c> kvm / libvirt & spice (QXL) video driver works well
2015-04-25 20:11:22 <rivarun> BitL0G1c: which package covers these?
2015-04-25 20:12:13 <BitL0G1c> spice-server xf86-video-qxl
2015-04-25 20:12:22 <BitL0G1c> libvirt
2015-04-25 20:12:49 <BitL0G1c> you have a linux host ?
2015-04-25 20:13:20 <rivarun> yes
2015-04-25 20:13:51 <BitL0G1c> debian ?
2015-04-25 20:14:38 <rivarun> gentoo
2015-04-25 20:14:54 <BitL0G1c> install libvirt & kvm
2015-04-25 20:15:11 <BitL0G1c> let me get my settings for libvirt
2015-04-25 20:15:25 <rivarun> oh. i thought that somehow that driver worked on virtualbox. i have no interest in either, but thanks
2015-04-25 20:16:47 <BitL0G1c> ok
2015-04-25 20:17:07 <BitL0G1c> i never had any luck with virtualbox - kvm works great
2015-04-25 20:19:13 <BitL0G1c> the QXL driver is kvm / xen or lxc with libvirt
2015-04-25 20:20:18 <BitL0G1c> i also have copy & paste working in kvm between the guest & host
2015-04-25 20:23:40 <BitL0G1c> you tried xf86-video-vmware ? for virtualbox instead of vga
2015-04-25 20:50:38 <ryuo> how viable is alpine linux for a server on real hardware? it appears to lack a lot of utility software i would want like samba.
2015-04-25 20:53:47 <ryuo> or i was looking in the wrong place...
2015-04-25 21:00:09 <_ikke_> ryuo: alpine is used a lot on real hardware with all kinds of usages
2015-04-25 21:00:44 <_ikke_> ryuo: There is a samba package: http://git.alpinelinux.org/cgit/aports/tree/main/samba?h=3.1-stable
2015-04-25 21:01:03 <ryuo> _ikke_, so how does the design differ from other distros? it appears you don't offer binary packages for everything you have a build file for.
2015-04-25 21:01:36 <_ikke_> ryuo: Different packages are in different stages. Some are in testing for example
2015-04-25 21:02:00 <ryuo> Partly due to the different libc?
2015-04-25 21:02:24 <_ikke_> Might be (I'm not too involved with Alpine)
2015-04-25 21:02:51 <_ikke_> ryuo: Note that alpine does tend to split things up in several sub packages
2015-04-25 21:02:58 <_ikke_> http://dl-3.alpinelinux.org/alpine/v3.1/main/x86_64/
2015-04-25 21:03:08 <ryuo> i noticed that it uses a format that resembles ARCH.
2015-04-25 21:03:13 <_ikke_> Yes
2015-04-25 21:03:19 <ryuo> Is there a relation?
2015-04-25 21:03:36 <_ikke_> I think mostly inspiration
2015-04-25 21:03:43 <ryuo> oh, so no direct source reuse?
2015-04-25 21:04:00 <_ikke_> I think most of it is self developed
2015-04-25 21:04:46 <ryuo> i assume it is required to build anything from aports by hand?
2015-04-25 21:04:58 <_ikke_> What do you mean?
2015-04-25 21:05:19 <ryuo> if i wanted to install samba, is it required to compile it first on the local machine?
2015-04-25 21:05:23 <_ikke_> No
2015-04-25 21:05:27 <_ikke_> just apk add samba
2015-04-25 21:05:34 <_ikke_> assuming you have the correct repository
2015-04-25 21:05:39 <ryuo> eh. i thought there would be a list of packages.
2015-04-25 21:06:06 <_ikke_> I don't know what each of those subpackages do, so I don't know which one you require
2015-04-25 21:06:51 <_ikke_> But alpine doesn't require you to compile each package yourself, it does come with pre-built packages
2015-04-25 21:06:58 <ryuo> k.
2015-04-25 21:07:06 <ryuo> i thought it did because i couldn't find a list.
2015-04-25 21:07:17 <ryuo> http://pkgs.alpinelinux.org/packages
2015-04-25 21:07:20 <ryuo> is pretty limited.
2015-04-25 21:07:35 <ryuo> oh.
2015-04-25 21:07:37 <ryuo> ...
2015-04-25 21:07:42 <ryuo> i thought there was only one page.
2015-04-25 21:08:08 <_ikke_> Yeah, pagination seems a bit confusing
2015-04-25 21:08:25 <_ikke_> though, if you search for samba, it does return the package name
2015-04-25 21:10:11 <ryuo> lol. the fact that it uses SQL shows.
2015-04-25 21:10:18 <ryuo> uses SQL wildcards. :)
2015-04-25 21:10:35 <_ikke_> kiss :P
2015-04-25 21:11:23 <Diftraku> LIKE whatever %
2015-04-25 21:11:31 <ryuo> indeed.
2015-04-25 21:12:21 <ryuo> huh. mono was removed.
2015-04-25 21:12:57 <ryuo> oh. because it is new.
2015-04-25 21:14:20 <ahills> Huh, I can't find the package "samba; DROP TABLE packages"
2015-04-25 21:14:45 <ryuo> properly designed system. :)
2015-04-25 21:15:13 <_ikke_> I would hope that the maintainers of a distro focused on security would think of that
2015-04-25 21:19:09 <ahills> haha
2015-04-25 21:30:05 <Diftraku> little bobby tables we call him
2015-04-25 21:31:15 <_ikke_> http://xkcd.com/327
2015-04-25 21:31:36 <_ikke_> How bad is it that I know the exact ID?
2015-04-25 21:31:52 <BitL0G1c> ryuo - run 'apk search samba' - everything is there
2015-04-25 21:32:02 <Diftraku> lol
2015-04-25 21:32:31 <BitL0G1c> hahahahahaha
2015-04-25 21:32:33 <ryuo> BitL0G1c, even the kitchen sink!?
2015-04-25 21:32:57 <Diftraku> why not: apk search samba'; DROP TABLE users;--
2015-04-25 21:33:35 <ryuo> delete from users where username = 'Diftraku';
2015-04-25 21:33:55 <ryuo> :)
2015-04-25 21:34:04 <_ikke_> Diftraku: That would leave your shell hanging waiting for the closing ' ;-)
2015-04-25 21:35:01 <Diftraku> DELETE FROM users WHERE username = 'Diftraku'; DROP TABLE users;--
2015-04-25 21:35:29 <_ikke_> Diftraku: Why first delete a single user, and then drop the whole table?
2015-04-25 21:35:44 <Diftraku> _ikke_: because sql injection
2015-04-25 21:36:21 <Diftraku> proabaly should've added an extra ' and ; after the comment lines
2015-04-25 21:36:43 <_ikke_> Diftraku: Doesn't matter, after the comment line, nothing gets interpreted
2015-04-25 21:36:58 <Diftraku> yup
2015-04-25 21:37:44 <ahills> is the apk index a sqlite database?
2015-04-25 21:37:58 <Diftraku> not sure, tbh
2015-04-25 21:38:10 <clandmeter> it is
2015-04-25 21:38:20 <clandmeter> regenerated every 15 min
2015-04-25 21:38:29 <Diftraku> inb4 sql injection in apk xD
2015-04-25 21:40:54 <ahills> of course everyone in this channel would report such a vulnerability responsibly
2015-04-25 21:40:58 <ahills> 
ACTION looks at Diftraku

2015-04-25 21:41:59 <ryuo> Hm.
2015-04-25 21:42:18 <ryuo> The easier solution would be to shoot the messenger. :P
2015-04-25 21:42:36 <ryuo> lol
2015-04-25 21:42:48 <ahills> well, I can't easily get to Finland, but it looks like you're in the US!
2015-04-26 09:57:44 <Jean-Scotch> To clone a run_from_ram instance is just to start a new one with the apkovl of the source one. Isn't it? It"s a PITA to have some packages config depending setting up the hostname in an explicit manner  (like collectd for instance) :(
2015-04-26 10:43:09 <Diftraku> define the hostname during install in format localhost.localDomain
2015-04-26 12:28:54 <sillysausage> hi, currently im using the x86_64 version of alpine linux on my router, an old desktop machine. I'm thinking of moving it all to a raspberry pi 2, because that would be more efficient on power.
2015-04-26 12:29:33 <sillysausage> the router I'm using at the moment, is pretty stock alpinelinux except that I've set up unbound dns (forwards to dnscrypt), openvpn, dns, dhcp the usual stuff you'd find on a router
2015-04-26 12:30:02 <sillysausage> can i copy the apkprov file from my x86_64 ramdisk to a armhf usb stick
2015-04-26 12:30:11 <sillysausage> ie one created with alpine-rpi-3.1.3-armhf.rpi.tar.gz
2015-04-26 12:30:44 <sillysausage> basically i want to convert my x86_64 router to an arm raspberry pi 2, what sorts of things should i look out for?
2015-04-26 12:40:03 <crow> does someone here use LVM on SSD? is that recomended at all? preparing for new alpinelinux install
2015-04-26 12:40:28 <sillysausage> i do that on my archlinux  machine, but that's because i'm using dmcrypt
2015-04-26 12:42:14 <crow> sillysausage you use it with grub or uefi? (because of that EF02 BIOS boot partition)
2015-04-26 12:45:11 <sillysausage> i use bios, but i do use gpt
2015-04-26 12:45:42 <sillysausage> eg:
2015-04-26 12:45:45 <sillysausage> 1      1049kB  2098kB  1049kB               BIOS boot partition  bios_grub
2015-04-26 12:45:54 <sillysausage> 2      3146kB  210MB   207MB   ext2         Linux filesystem
2015-04-26 12:46:01 <sillysausage> 3      211MB   64.0GB  63.8GB               Linux LVM            lvm
2015-04-26 12:47:19 <crow> sillysausage so for LVM there should be also that SSD aligement which is the important?
2015-04-26 12:47:44 <crow> for some time i wrote an update wiki http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM so actually i could use it on this ssd disk to
2015-04-26 12:47:55 <sillysausage> i used this
2015-04-26 12:47:57 <sillysausage> http://pteradigm.com/2012/09/16/installing-arch-linux-with-luks-system-encryption-under-lvm-and-bios-gpt-with-grub2/
2015-04-26 13:01:06 <sillysausage> hmm, i could use a sdcard as a ramdisk instead of a usb stick like i am at the moment
2015-04-26 13:05:46 <crow> sillysausage but your partition seems does not looks like correct alignmend partitions?
2015-04-26 13:06:34 <sillysausage> it should be
2015-04-26 13:09:49 <sillysausage> Number  Start    End         Size        File system  Name                 Flags
2015-04-26 13:09:50 <sillysausage>  1      2048s    4096s       2049s                    BIOS boot partition  bios_grub
2015-04-26 13:09:52 <sillysausage>  2      6144s    409600s     403457s     ext2         Linux filesystem
2015-04-26 13:09:55 <sillysausage>  3      411648s  125045390s  124633743s               Linux LVM            lvm
2015-04-26 13:09:56 <crow> well if the first secort is the 1049 then it is not ok, as you cannot devide by 8.
2015-04-26 13:09:56 <sillysausage> hmm you might be right crow
2015-04-26 13:10:05 <sillysausage> i was trying to check it with https://www.techpowerup.com/articles/other/157
2015-04-26 13:10:33 <sillysausage> Number  Start    End       Size      File system  Name                 Flags
2015-04-26 13:10:34 <sillysausage>  1      1.00MiB  2.00MiB   1.00MiB                BIOS boot partition  bios_grub
2015-04-26 13:10:36 <sillysausage>  2      3.00MiB  200MiB    197MiB    ext2         Linux filesystem
2015-04-26 13:10:38 <sillysausage>  3      201MiB   61057MiB  60856MiB               Linux LVM            lvm
2015-04-26 13:10:40 <sillysausage> is that wrong?
2015-04-26 13:15:18 <sillysausage> well this seems to work crow
2015-04-26 13:15:25 <sillysausage> (parted) align-check opt 1
2015-04-26 13:15:27 <sillysausage> 1 aligned
2015-04-26 13:15:29 <sillysausage> (parted) align-check opt 2
2015-04-26 13:15:31 <sillysausage> 2 aligned
2015-04-26 13:15:33 <sillysausage> (parted) align-check opt 3
2015-04-26 13:15:35 <sillysausage> 3 aligned
2015-04-26 13:15:44 <sillysausage> so surely if it was not aligned parted would have reported otherwise
2015-04-26 13:15:59 <crow> sillysausage well i dont know but i read that the start sector should be aligmend.. here is from my hdd http://sprunge.us/bdLQ
2015-04-26 13:16:20 <darkfader> sillysausage: lvm should manage the alignment on created lv's anyway. or at leasst there's some opts in lvm.conf for it
2015-04-26 13:16:36 <sillysausage> hmm
2015-04-26 13:17:10 <crow> http://sprunge.us/KJii
2015-04-26 13:19:12 <crow> sillysausage but i guess gptfdisk works with 2048 as start already http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM#Partitioning
2015-04-26 13:19:30 <crow> there you will see my partition on an VMs in time i wrote that wiki
2015-04-26 13:20:00 <sillysausage> well i looked at this https://bbs.archlinux.org/viewtopic.php?pid=1359312#p1359312
2015-04-26 13:20:13 <sillysausage> > I align everything to MiB boundaries. That works for every drive I had thus far.
2015-04-26 13:20:19 <sillysausage> and that's what i did there didn't i?
2015-04-26 13:20:53 <sillysausage> as for the partitions in LVM im not sure about that im an LVM noob
2015-04-26 13:21:02 <sillysausage> but it is nice to know darkfader that automatically did it
2015-04-26 13:41:47 <crow> sillysausage i am also not expter just reading about ssd and aligment
2015-04-26 13:43:00 <crow> can you sprunge the : sudo sfdisk -d /dev/sda and the sudo fdisk -l -u /dev/sda (change sda to your ssd drive)
2015-04-26 13:51:26 <sillysausage> https://dpaste.de/tRB3 that what you looking for crow?
2015-04-26 13:53:10 <crow> sillysausage well thats looks just fine, but in your first paste here in channel you paste: (14:45:40) (sillysausage) 1 1049kB 2098kB 1049kB BIOS boot partition bios_grub
2015-04-26 13:53:21 <crow> that 1049 as start sector is/was not good
2015-04-26 13:53:55 <crow> start sector from your all 3 partition is devided by 8 just fine without any ,
2015-04-26 13:54:51 <sillysausage> that's what parted tiold me
2015-04-26 13:54:58 <sillysausage> *told
2015-04-26 13:55:21 <sillysausage> oh that was the size
2015-04-26 13:55:31 <sillysausage> Number  Start   End     Size    File system  Name                 Flags
2015-04-26 13:55:33 <sillysausage>  1      1049kB  2098kB  1049kB               BIOS boot partition  bios_grub
2015-04-26 13:55:38 <sillysausage> it's a 1MiB partition
2015-04-26 13:55:55 <sillysausage> Number  Start    End       Size      File system  Name                 Flags
2015-04-26 13:55:57 <sillysausage>  1      1.00MiB  2.00MiB   1.00MiB                BIOS boot partition  bios_grub
2015-04-26 13:56:17 <sillysausage> so i think that's okay
2015-04-26 13:57:27 <sillysausage> so i think im going to commit and buy a pi2 for my alpine router
2015-04-26 14:07:34 <BitL0G1c> I've added GPT support to setup-disk & my partition script https://it-offshore.co.uk/linux/alpine-linux
2015-04-26 14:08:08 <crow> sillysausage :) ok the size, well Start/End was a bit weird in that case
2015-04-26 14:08:27 <crow> BitL0G1c i know, i checkd your video 10min before
2015-04-26 14:08:42 <crow> BitL0G1c i dont need luks, but rest looks nice
2015-04-26 14:09:06 <BitL0G1c> :) yes - makes installing a bit quicker
2015-04-26 14:09:09 <crow> dont see reason in my case to crypt something for server on local infra
2015-04-26 14:09:26 <crow> BitL0G1c aligment is checked to?
2015-04-26 14:09:47 <BitL0G1c> I followed the instruction on the arch linux wiki
2015-04-26 14:09:55 <BitL0G1c> & coded that
2015-04-26 14:11:10 <BitL0G1c> have a look at gpt_create()
2015-04-26 14:11:40 <BitL0G1c> I use sgdisk to setup gpt
2015-04-26 14:12:25 <sillysausage> does the pi2 interface support vlan tagging?
2015-04-26 14:12:37 <sillysausage> i would think most things would these days
2015-04-26 14:15:04 <sillysausage> looks like it does. https://opusmagnus.wordpress.com/2014/10/04/create-a-tor-only-vlan-with-a-raspberry-pi/
2015-04-26 14:15:05 <sillysausage> neat idea.
2015-04-26 14:16:10 <crow> sillysausage my rpi2 like rpi1 is just laying arround doing nothing jet :)
2015-04-26 14:16:37 <crow> BitL0G1c thanks, ill try to use code from your script and update the instruction i wrote here if needed http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM
2015-04-26 14:19:05 <BitL0G1c> ;-)
2015-04-26 14:48:26 <ahills> sillysausage: you're going to use an rpi... as a router???
2015-04-26 14:50:04 <sillysausage> ah rp2, yeah it's only my own home gateway
2015-04-26 14:50:46 <sillysausage> i know it will be limited at around 30Mbit, but my internet doesn't go any faster than that
2015-04-26 16:05:30 <ramonsoares> hello, i have a question about apk dev packages, have any way of install ONLY -dev packages?
2015-04-26 16:06:18 <ramonsoares> ex.: postgresql-dev install all postgresql packages as deps,
2015-04-26 16:41:49 <crow> BitL0G1c does luks need vfat (sgdisk code EF00) partition? or why you are creating this one?
2015-04-26 16:52:00 <BitL0G1c> the various GPT guides I saw had it as fat32 (so you can boot other oses) - luks doesnt need it
2015-04-26 16:53:19 <BitL0G1c> if you will only be booting linux - it could be ext whatever
2015-04-26 16:58:19 <crow> ok
2015-04-26 18:24:08 <crow> minor changes applayed to wiki: http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM
2015-04-26 18:52:07 <crow> if i am using syslinux, should not be there syslinux.cfg ? (on old run from ram install there was such file). find / -name syslinux.cfg does not find anything. (needed to add serial output from boot cmdline egg to append: console=ttyS0,115200
2015-04-26 18:54:32 <crow> ok seems now its extlinux.conf :) http://wiki.alpinelinux.org/wiki/Enable_Serial_Console_on_Boot#extlinux.conf_.282.5.2B.29
2015-04-26 19:47:08 <crow> someone knows what is MODULE_0 in lm_sensors? here my stuff: http://sprunge.us/gIeU
2015-04-26 20:04:55 <crow> seems a bug https://bugs.gentoo.org/show_bug.cgi?id=480018 , the HWMON_MODULES should be renamed to MODULE_0 inside /etc/conf.d/lm_sensors . here my stuff which are working http://sprunge.us/ZhcC
2015-04-27 12:45:20 <johnf1911> hello
2015-04-27 12:45:44 <johnf1911> I'm trying to use opennhrp and strongswan in a fairly basic phase I NHRP configuration
2015-04-27 12:45:59 <johnf1911> wondered if there's anyone around who knows opennhrp
2015-04-27 12:52:04 <ncopa> johnf1911: we have some here who knows a bit, yes
2015-04-27 12:52:31 <johnf1911> ok, cool, let me continue
2015-04-27 12:52:40 <johnf1911> I'm trying to use strongswan and not racoon
2015-04-27 12:52:54 <ncopa> i dont think strongswan currently works with opennhrp atm
2015-04-27 12:53:01 <johnf1911> I don't need anything too fancy, no dynamic tunnel initiation
2015-04-27 12:53:28 <johnf1911> as I understand it the main issues with strongswan/opennhrp stem from needing to build new tunnels dynamically
2015-04-27 12:53:31 <johnf1911> but I don't need that
2015-04-27 12:53:39 <johnf1911> https://lists.strongswan.org/pipermail//dev/2014-August/001017.html
2015-04-27 12:53:58 <johnf1911> I see some recentish discussion on the subject
2015-04-27 12:54:10 <johnf1911> I'm probably better off switching to racoon I'm guessing
2015-04-27 12:54:28 <johnf1911> but I feel like I'm very close to having it functional
2015-04-27 12:55:21 <ncopa> opennhrp needs the ipsec layer do ipsec in transport mode and not tunnel mode
2015-04-27 12:55:32 <ncopa> the tunnel is done with gre
2015-04-27 12:56:00 <ncopa> so you will have to configure strongswan to do ipsec in transport mode for gre traffic
2015-04-27 12:56:09 <ncopa> and you should be fine - theoretically
2015-04-27 12:56:11 <johnf1911> yes, that's done
2015-04-27 12:56:40 <ncopa> what is the error message you get?
2015-04-27 12:56:40 <johnf1911> I have nice ip proto 50 traffic flowing between the two
2015-04-27 12:56:46 <johnf1911> I don't get an error message at all
2015-04-27 12:57:14 <ncopa> what does opennhrp logs say?
2015-04-27 12:57:46 <johnf1911> the Cisco device is basically ignoring the nhrp requests
2015-04-27 12:58:00 <ncopa> oh its against a cisco
2015-04-27 12:58:03 <ncopa> hum
2015-04-27 12:58:29 <ncopa> iirc there are some bugs in (some?) ciscos
2015-04-27 12:58:40 <ncopa> and i think there are a config option for opennhrp to emulate it
2015-04-27 12:59:22 <ncopa> i also know not all kernels have been good
2015-04-27 12:59:31 <ncopa> some have needed patching
2015-04-27 12:59:45 <fabled> johnf1911, there's no public integration of opennhrp with strongswan at the moment
2015-04-27 13:00:02 <fabled> i suspect there's one by some commercial entity
2015-04-27 13:00:15 <fabled> but i'm intending to write the quagga plugin soonish
2015-04-27 13:00:26 <fabled> in fact i'm already coding it; it's just nowhere complete yet
2015-04-27 13:00:42 <johnf1911> ah, I see
2015-04-27 13:00:46 <johnf1911> so, for me
2015-04-27 13:00:52 <johnf1911> the Cisco device doesn't report anything
2015-04-27 13:01:00 <fabled> it could various things
2015-04-27 13:01:12 <johnf1911> how do I turn up the debug on opennhrp
2015-04-27 13:01:20 <johnf1911> because all I'm getting is
2015-04-27 13:01:22 <fabled> the opennhrp/ipsec integration consist of two parts:
2015-04-27 13:01:32 <johnf1911> Mon Apr 27 12:39:23 2015 daemon.info opennhrp[2840]: Sending Registration Request to 172.24.12.1 (my mtu=0)
2015-04-27 13:01:36 <fabled>  - requesting the tunnels per dynamic tunnel
2015-04-27 13:01:43 <fabled>  - letting ipsec report back any dead peers
2015-04-27 13:01:57 <johnf1911> yes, this is only Phase I so there are no dynamic tunnels
2015-04-27 13:02:19 <johnf1911> it doesn't put any kind of info about the IPSec tunnels into the NHRP request, does it?
2015-04-27 13:02:31 <johnf1911> I'm thinking though, which I didn't know when I started this
2015-04-27 13:02:32 <fabled> no
2015-04-27 13:02:46 <johnf1911> that opennhrp is kind of a control script for racoon and quagga
2015-04-27 13:02:49 <fabled> nhrp and ipsec are disjoint oon protocol level
2015-04-27 13:02:57 <johnf1911> in that it's aware of the state
2015-04-27 13:03:02 <johnf1911> and driving the configuration
2015-04-27 13:03:03 <johnf1911> as such
2015-04-27 13:03:09 <fabled> yes, it's additionally driving kernel
2015-04-27 13:03:21 <johnf1911> I'm probably better off stopping messing around with strongswan
2015-04-27 13:03:24 <johnf1911> and switching to racoon
2015-04-27 13:03:28 <fabled> for the time being yes
2015-04-27 13:03:32 <fabled> it's non-trivial to set it up
2015-04-27 13:03:36 <johnf1911> even if strongswan is, from my reading, the current preferred ipsec stack
2015-04-27 13:03:44 <johnf1911> well, I do have it set up[
2015-04-27 13:03:48 <johnf1911> and it wasn't trivial :)
2015-04-27 13:03:59 <fabled> i mean the opennhrp integration part
2015-04-27 13:04:05 <johnf1911> but the Cisco device somehow doesn't like my registration request
2015-04-27 13:04:12 <johnf1911> yes, if I want something more than basic
2015-04-27 13:04:18 <johnf1911> I will have an external control script
2015-04-27 13:04:18 <fabled> perhaps it's not going ipsec encrypted?
2015-04-27 13:04:22 <johnf1911> it is
2015-04-27 13:04:39 <johnf1911> 12:37:12.265630 IP 10.2.0.26 > x.x.x.x:  ip-proto-50 184
2015-04-27 13:04:47 <johnf1911> ignore the private IP, it's not NATed
2015-04-27 13:05:08 <fabled> mmm. how is your opennhrp.conf like?
2015-04-27 13:06:33 <johnf1911> minimal
2015-04-27 13:07:33 <johnf1911> just a moment
2015-04-27 13:11:37 <johnf1911> https://gist.github.com/johnfzc/0eb2b5ec5dc663df5a47
2015-04-27 13:11:43 <johnf1911> these are my configurations
2015-04-27 13:11:52 <johnf1911> I have a packet capture of the NHRP packet created (with crypto off)
2015-04-27 13:12:07 <johnf1911> packet captures also confirm that the packet is being encrypted properly
2015-04-27 13:12:17 <johnf1911> including dec counters incrementing on the Cisco side
2015-04-27 13:13:02 <johnf1911> NBMA and Protocol addresses in the request are correct
2015-04-27 13:15:26 <johnf1911> any ideas?
2015-04-27 13:16:27 <fabled> try enable cisco debug?
2015-04-27 13:16:36 <johnf1911> yes, already done
2015-04-27 13:16:41 <johnf1911> debug ip nhrp packet shows nothing
2015-04-27 13:16:44 <johnf1911> er, no ip
2015-04-27 13:17:01 <fabled> is grekey matching?
2015-04-27 13:17:18 <fabled> that is how is the gre tunnel configured on cisco vs. linux?
2015-04-27 13:17:33 <johnf1911> fuck
2015-04-27 13:17:37 <johnf1911> ;_;
2015-04-27 13:17:47 <johnf1911> I thought about that but never actually did it
2015-04-27 13:17:49 <johnf1911> just a moment
2015-04-27 13:23:33 <johnf1911> wow, it's working
2015-04-27 13:23:36 <johnf1911> thank you very much for that
2015-04-27 13:23:43 <johnf1911> it was a pretty pathetic oversite on my part
2015-04-27 13:24:08 <johnf1911> hmm, so now I'm faced with a difficult decision
2015-04-27 13:24:18 <johnf1911> on if I continue with this maual hack job
2015-04-27 13:24:22 <johnf1911> or switch to racoon
2015-04-27 13:24:23 <johnf1911> :/
2015-04-27 13:27:32 <johnf1911> hmm, reading this list, I really wonder about the future health of racoon
2015-04-27 13:27:42 <johnf1911> https://lists.debian.org/debian-devel/2014/04/msg00075.html
2015-04-27 13:30:09 <fabled> johnf1911, i've been doing racoon coding for quite a bit, and am also in favor of switching to strongSwan
2015-04-27 13:30:49 <fabled> hopefully nhrp runs with strongSwan within few months
2015-04-27 13:31:06 <fabled> but it's still work-in-progress
2015-04-27 13:31:15 <johnf1911> I'm going to need to support the solution I deploy for the next three years minimum
2015-04-27 13:31:29 <johnf1911> I think my easiest path is probably to continue on the one that I'm on
2015-04-27 13:31:37 <johnf1911> and use strongswan
2015-04-27 13:31:46 <johnf1911> I will need my own control scripts no matter what
2015-04-27 13:31:51 <johnf1911> later, when you integrate proper support
2015-04-27 13:32:00 <johnf1911> I'll try to reduce the amount of hackery that's going on
2015-04-27 13:45:10 <johnf1911> fabled: how can I adjust the frequency of NHRP requests? I suspect that it's sending NHRP registration requests every minute
2015-04-27 13:45:33 <johnf1911> you don't seem to be following the cisco standard 1/3 of hold time after registration
2015-04-27 13:45:41 <johnf1911> (unless I'm wrong about the source of the traffic)
2015-04-27 13:52:01 <johnf1911> hmm, perhaps I take that back
2015-04-27 15:18:32 <darkorigen> hi, how can I modify grub and update them?
2015-04-27 15:57:40 <JonnyBGo_> hi. Anyone facing problems with alpine apk?
2015-04-27 15:57:47 <JonnyBGo_> I am always getting fetch http://dl-4.alpinelinux.org/alpine/v3.1/main/x86_64/APKINDEX.tar.gz ERROR: http://dl-4.alpinelinux.org/alpine/v3.1/main: IO ERROR WARNING: Ignoring APKINDEX.689bb31a.tar.gz: No such file or directory
2015-04-27 16:04:13 <clandmeter>  JonnyBGod_ which cmd did you run?
2015-04-27 16:04:27 <clandmeter> did you apk update?
2015-04-27 17:10:27 <JonnyBGod_> apk --update add openssl-dev
2015-04-27 17:10:45 <JonnyBGod_> on osx boot2docker
2015-04-27 17:15:27 <JonnyBGod_> same problem when running: apk update
2015-04-27 17:17:43 <clandmeter> JonnyBGod_: can you pastebin your repositories file?
2015-04-27 17:20:15 <JonnyBGod_> http://pastebin.com/NJQAugwA
2015-04-27 17:20:24 <JonnyBGod_> thats my dockerfile
2015-04-27 17:20:40 <JonnyBGod_> I tested with a lot of simpler commands as well
2015-04-27 17:21:53 <JonnyBGod_> and here is the console log: http://pastebin.com/xRzMgk3Z
2015-04-27 17:23:51 <ahills> JonnyBGod_: did you try a different mirror, just in case?
2015-04-27 17:58:31 <systmkor> JonnyBGod_: Why Johnny Still Can’t Encrypt :P http://www.chariotsfire.com/pub/sheng-poster_abstract.pdf
2015-04-27 19:13:47 <johnf1911> fabled: hmm, so I should add cisco to the line?
2015-04-27 19:15:34 <johnf1911> I'm afraid I didn't fully follow your response on the mailing list
2015-04-27 20:22:45 <Risk64> How can i format ext4 with alpine-mini-3.1.3-x86.iso booted?
2015-04-27 20:23:32 <crow> Risk64 http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM
2015-04-27 20:23:38 <crow> (part with ext4
2015-04-27 20:24:11 <crow> Risk64 i guess you need to: apk add e2fsprogs
2015-04-27 20:24:25 <crow> to be able to use mkfs.ext4
2015-04-27 20:24:39 <Risk64> yes.. but its not on the iso in apk/ folder..
2015-04-27 20:25:05 <Risk64> at least not with obvious name..
2015-04-27 20:28:16 <crow> Risk64 you did apk add e2fsprogrs like i wrote above?
2015-04-27 20:28:20 <crow> Risk64 /sbin/mkfs.ext4 is owned by e2fsprogs-1.42.12-r0
2015-04-27 20:28:49 <Risk64> says missing
2015-04-27 20:29:33 <crow> please sprunge.us it.
2015-04-27 20:32:11 <Risk64> i have no network there.. i just booted up the mini-iso and partitioned the hdd. Now i need to format it for the alpine-disk script.
2015-04-27 20:33:00 <Risk64> Previously i used alpine-setup but i don't want a swap-partition.
2015-04-27 20:33:36 <Risk64> so i partitioned so far and now need to format... the only packages i see for apk are on cdrom/apks. But there is no e2fsprogs
2015-04-27 20:34:13 <crow> well yea i guess you cant do it with mini-iso then
2015-04-27 20:34:22 <crow> can you ssh stuff to that box?
2015-04-27 20:34:33 <crow> or connect usb device?
2015-04-27 20:34:53 <crow> you could download package from here and install http://nl.alpinelinux.org/alpine/v3.1/main/
2015-04-27 20:35:14 <Risk64> i will set it up in virtualbox...
2015-04-27 20:35:49 <crow> why you are using mini-iso if i my ask
2015-04-27 20:36:25 <Risk64> to keep it as small as possible from the beginning. I didn't expect ext4 missing though..
2015-04-27 20:36:40 <Risk64> the 'setup-alpine'
2015-04-27 20:36:47 <Risk64>  script did nevertheless somehow
2015-04-27 20:36:58 <Risk64> this is my 2nd run.
2015-04-27 20:37:48 <Risk64> or it was vfat and i didn't noticed.. :)
2015-04-27 20:40:31 <bitshark_> hey hey
2015-04-27 20:40:34 <bitshark_> ncopa around
2015-04-27 20:41:13 <crow> Risk64 well here you can see that setup-disk is doing http://git.alpinelinux.org/cgit/alpine-conf/tree/setup-disk.in <- line 444 adds some packages which are needed for ext3,4 etc
2015-04-27 20:41:13 <bitshark_> or ne1 else
2015-04-27 20:41:34 <bitshark_> okay w0rd
2015-04-27 20:42:04 <bitshark_> HEH.
2015-04-27 20:42:23 <bitshark_> so my experimentz with Alpine Linux as desktop were a pretty decent success...   no problems at all with the glibc replacement
2015-04-27 20:42:51 <bitshark_> the only issue i ran into is problems related to GTK2 / GTK3 has dependencies on Systemd
2015-04-27 20:43:08 <Risk64> bitshark_: thx, i guess i will try to set up internet to get the missing packages...
2015-04-27 20:43:32 <bitshark_> what do ya mean
2015-04-27 20:43:50 <bitshark_> <-- dont understand , dense
2015-04-27 20:44:17 <Risk64> sorry, meant crow.. :)
2015-04-27 20:44:25 <bitshark_> oh ok
2015-04-27 20:44:26 <bitshark_> word
2015-04-27 20:44:50 <bitshark_> i just built CLFS so i know what u mean with the bootloaders packages
2015-04-27 20:44:50 <bitshark_> haha
2015-04-27 20:44:56 <bitshark_> for ext2/3/4
2015-04-27 20:45:29 <bitshark_> i kept getting kernel panics in vmware and virtualbox because the kernel config didn't have certain SCSI / SATA stuff compiled in as Y (not M)
2015-04-27 20:45:36 <ahills>   bitshark_: I'm pretty sure those are GNOME 3 dependencies, not GTK dependencies
2015-04-27 20:45:45 <bitshark_> finally i set vmware to IDE and my CLFS machine boots right up
2015-04-27 20:45:51 <bitshark_> oh yeah?
2015-04-27 20:45:59 <crow> bitshark_ i have here vmware workstation and dont see any kernel panics,
2015-04-27 20:46:00 <ahills> bitshark_: there's at least one project out there that stubs systemd's dbus interfaces, but I can't remember the name or where I found it
2015-04-27 20:46:11 <bitshark_> well i dunno .. i'm not an expert at this .. especially not at X
2015-04-27 20:46:18 <bitshark_> i just hate systemd and i want it out forever
2015-04-27 20:46:19 <crow> i dont use X :)
2015-04-27 20:46:28 <ahills> you probably don't want GNOME 3
2015-04-27 20:46:31 <ahills> because it requires systemd
2015-04-27 20:46:36 <bitshark_> naw i hate gnome3
2015-04-27 20:46:37 <bitshark_> its out
2015-04-27 20:46:43 <bitshark_> i used to use MATE (gnome2 fork)
2015-04-27 20:46:55 <bitshark_> that also may be out because it's getting harder and harder
2015-04-27 20:47:05 <Risk64> try lxqt
2015-04-27 20:47:15 <bitshark_> we're up to 69 unique userspace binaries for systemd alone
2015-04-27 20:47:20 <bitshark_> but anyway i did like mate... i may move on to lxqt
2015-04-27 20:47:22 <bitshark_> yes!
2015-04-27 20:47:25 <bitshark_> great minds
2015-04-27 20:47:48 <bitshark_> there's also enlightenment / kde derived stuff
2015-04-27 20:47:54 <ahills> where were you finding systemd dependencies?
2015-04-27 20:47:55 <bitshark_> that's not as tightly coupled to systemd
2015-04-27 20:47:59 <bitshark_> oh ahills
2015-04-27 20:48:02 <bitshark_> u gotta see this one hahah
2015-04-27 20:48:57 <bitshark_> http://i.imgur.com/5MchQde.png
2015-04-27 20:49:11 <bitshark_> --^ i think that's from the guys who forked debian and their buddies
2015-04-27 20:49:23 <bitshark_> not sure... but it shows dependencies in jessie i think
2015-04-27 20:49:45 <ahills> yeah, they're reusing some of the updated components from the GNOME3 environment
2015-04-27 20:50:26 <bitshark_> okay
2015-04-27 20:50:33 <bitshark_> is that whats going on?
2015-04-27 20:50:54 <bitshark_> wasnt really clear here... just was getting my LFS / CLFS build going
2015-04-27 20:51:08 <ahills> Those desktop environments are so needlessly complex...
2015-04-27 20:51:14 <bitshark_> then looking into the solution for what i'm gonna run on alpine desktop
2015-04-27 20:51:21 <ahills> Good luck on your crusade
2015-04-27 20:51:45 <ahills> I think I might run Alpine on my laptop with an Arch-OpenRC container for the stuff that needs it
2015-04-27 20:52:28 <bitshark_> yeah
2015-04-27 20:52:28 <bitshark_> well there's other experimentz goin on at the same time, heh
2015-04-27 20:52:34 <bitshark_> with freebsd and pc-bsd
2015-04-27 20:52:49 <bitshark_> no one is gonna pwn my computer with redhat malwarez tho
2015-04-27 20:53:09 <bitshark_> even if it's just crappily written schizo code
2015-04-27 20:53:19 <ahills> calm down there...
2015-04-27 20:53:37 <bitshark_> but yeh... so i checked out pc-bsd (freebsd with a nice installer and desktop env)
2015-04-27 20:53:48 <bitshark_> noticed there's something called lumina
2015-04-27 20:53:54 <bitshark_> that might be promising as a port to linux
2015-04-27 20:53:56 <bitshark_> http://wiki.pcbsd.org/index.php/Lumina
2015-04-27 20:54:36 <bitshark_> as alternative to lxqt and xfce
2015-04-27 21:23:18 <iwearsable> I feel really stupid to keep asking this, but I just can't figure out why my system isn't seeing updates.
2015-04-27 21:24:36 <iwearsable> A good example would be on the package search website, it shows linux-grsec at version 3.18.12-r1. If I run apk update and apk search '*linux-grsec*' I only see 3.14.37-1 (which I already have)
2015-04-27 21:25:47 <_ikke_> What version of alpine are you running, and what repository do you have?
2015-04-27 21:26:11 <iwearsable> 3.2.0_alpha1
2015-04-27 21:26:57 <iwearsable> http://dl-3.alpinelinux.org/alpine/v3.1/main
2015-04-27 21:27:23 <iwearsable> There's also edge/main and edge/testing enabled
2015-04-27 21:28:12 <_ikke_> Both 3.1 and edge only have 3.14.37-r0
2015-04-27 21:29:26 <_ikke_> edge/testing doesn't contain linux-grsec
2015-04-27 21:29:49 <BitL0G1c> apk add linux-grsec@edge
2015-04-27 21:30:09 <clandmeter> edge has linux-grsec-3.18.12-r1
2015-04-27 21:30:09 <_ikke_> BitL0G1c: http://dl-3.alpinelinux.org/alpine/edge/main/x86_64/ only shows 3.14.37-r1
2015-04-27 21:30:19 <clandmeter> dont use dl-3
2015-04-27 21:30:22 <clandmeter> its lagging
2015-04-27 21:30:22 <_ikke_> right
2015-04-27 21:30:25 <BitL0G1c> use the nl.alpine repo
2015-04-27 21:30:43 <_ikke_> that's the problem then
2015-04-27 21:30:46 <iwearsable> okay, let me try that
2015-04-27 21:35:19 <iwearsable> Switching to nl worked! 175 packages to upgrade
2015-04-27 21:35:40 <iwearsable> Thank you all, that had been confusing me for a few weeks now
2015-04-27 21:36:39 <clandmeter> we should probalby redirect dl-3. seems off for a long time already.
2015-04-27 21:36:55 <_ikke_> Who is maintaining it?
2015-04-27 21:37:51 <clandmeter> dont know
2015-04-27 21:39:23 <nangel> 
ACTION here... I'll look at dl-3

2015-04-27 21:51:59 <nangel> iwearsable: sorry, dl-3 had a dns issue and wasn't resolving rsync.a.o - its fixed now.  Sorry for wasting your time.
2015-04-27 23:15:35 <nangel> dl-3 should be up-to-date
2015-04-27 23:18:56 <clandmeter> grazie
2015-04-27 23:21:54 <Risk64> When installing 'webmin' it asks to pick a specific OS optimization. Is there a better choice than 'Generic Linux'? See here: http://pastebin.com/2y8gEfY5
2015-04-27 23:28:01 <nangel> mm... http://wiki.alpinelinux.org/wiki/Webmin has no guidance
2015-04-27 23:29:16 <Risk64> yep.. i have taken generic linux so far..
2015-04-27 23:30:28 <Risk64> it also tries to create /etc/rc.d/init.d/webmin that is not available. Needs some extra steps.
2015-04-27 23:57:34 <crow> Risk64 you could try acf instead webmin (setup-acf) http://wiki.alpinelinux.org/wiki/Managing_ACF
2015-04-28 00:06:53 <Risk64> i did so already.. its ok but i want to see webmin as well.
2015-04-28 00:07:00 <Risk64> crow
2015-04-28 00:07:31 <Risk64> i'm running in virtualbox to take a look around
2015-04-28 00:08:01 <Risk64> webmin is still installing (since 20min). I'm not sure if it hangs... the VM is not very fast.
2015-04-28 02:47:34 <bitshark_> yo
2015-04-28 02:47:47 <bitshark_> whens ncopa around
2015-04-28 02:47:52 <bitshark_> ne one know
2015-04-28 02:49:18 <ahills> ncopa is both always and never here
2015-04-28 02:49:27 <ahills> ncopa is the force that binds alpine linux together
2015-04-28 02:51:11 <bitshark_> yeah
2015-04-28 02:51:13 <bitshark_> haha
2015-04-28 02:51:49 <bitshark_> well i've been looking around for good lightweight window managers that aren't so strongly coupled with systemd
2015-04-28 02:52:16 <bitshark_> i found a couple of really cool ones
2015-04-28 02:52:16 <bitshark_> one in particular is called EDE2  (Equinox)
2015-04-28 02:52:17 <bitshark_> it's great
2015-04-28 02:52:25 <bitshark_> running it now instead of MATE
2015-04-28 02:53:21 <bitshark_> http://equinox-project.org/
2015-04-28 02:53:39 <bitshark_> check out this comparison of linux desktop memory footprints...  https://eerielinux.wordpress.com/tag/equinox/
2015-04-28 02:53:42 <bitshark_> fascinating
2015-04-28 02:55:50 <ahills> you're looking for dwm
2015-04-28 02:55:58 <ahills> git clone git://git.suckless.org/dwm
2015-04-28 02:55:58 <bitshark_> yeah?
2015-04-28 02:55:59 <ahills> cd dwm
2015-04-28 02:56:00 <bitshark_> k
2015-04-28 02:56:01 <ahills> make
2015-04-28 02:56:05 <bitshark_> i'll try it now
2015-04-28 02:56:08 <ahills> edit config.h to configure it
2015-04-28 02:56:16 <ahills> http://dwm.suckless.org/
2015-04-28 02:56:16 <bitshark_> thx
2015-04-28 02:56:19 <ahills> np
2015-04-28 02:56:34 <ahills> oh, you'll probably also want to install dmenu with it
2015-04-28 02:56:50 <bitshark_> oh that's pretty slick
2015-04-28 02:57:14 <ahills> I've been using dwm exclusively for years now, everything else just seems to fall short
2015-04-28 02:57:24 <ahills> (as everyone says about everything they like)
2015-04-28 03:01:57 <bitshark_> is it actively maintained still
2015-04-28 03:02:02 <bitshark_> or abandonware
2015-04-28 03:03:36 <bitshark_> hm hang on gonna try dwm
2015-04-28 03:03:38 <bitshark_> brb
2015-04-28 03:25:30 <iwearsable> @bitshark, wow that equinox project looks like Windows 2000.
2015-04-28 03:25:41 <iwearsable> I love my OpenBox, but that's kind of cool
2015-04-28 06:08:39 <ncopa> morning
2015-04-28 06:27:46 <kocka> moin
2015-04-28 07:09:06 <ScrumpyJack> moin
2015-04-28 07:53:54 <sillysausage> hmm i seem to be having some troubles booting my raspberry pi
2015-04-28 07:54:02 <sillysausage> I couldn't find much in the way of documentation except for http://forum.alpinelinux.org/comment/228#comment-228
2015-04-28 07:54:20 <sillysausage> http://elinux.org/R-Pi_Troubleshooting#Red_power_LED_does_not_light.2C_nothing_on_display
2015-04-28 07:54:55 <sillysausage> i formatted it with fdisk like so http://wiki.alpinelinux.org/wiki/Create_a_Bootable_Compact_Flash#Format_CF_Card
2015-04-28 07:56:24 <sillysausage> fdisk layout/contents of card
2015-04-28 07:56:25 <sillysausage> https://dpaste.de/49dW
2015-04-28 07:56:47 <sillysausage> i wonder what i'm missing.
2015-04-28 07:58:35 <sillysausage> i'm using a HDMI to HDMI plug so that shouldn't be the issue
2015-04-28 08:03:39 <sillysausage> i might try using the noobs image to see that it works
2015-04-28 08:08:27 <crow> sillysausage i used this on my rpi1 http://sprunge.us/Lgah
2015-04-28 08:08:31 <sillysausage> http://wiki.alpinelinux.org/wiki/RPI_Video_Receiver
2015-04-28 08:09:14 <sillysausage> well i did set it up on my archlinux machine
2015-04-28 08:09:28 <sillysausage> i dont have another alpinelinux computer, and certainly not a one that is arm based
2015-04-28 08:09:47 <sillysausage> but your fdisk setup looks same as mine
2015-04-28 08:10:50 <sillysausage> apk add dosfstools
2015-04-28 08:10:52 <sillysausage> mkdosfs -F32 /dev/sdc1
2015-04-28 08:10:54 <sillysausage> unless that is different to
2015-04-28 08:10:57 <sillysausage> mkfs.vfat
2015-04-28 08:11:00 <crow> you have rpi1 or 2?
2015-04-28 08:11:04 <sillysausage> rp2
2015-04-28 08:11:38 <sillysausage> :S hope that alpinelinux works on the rp2
2015-04-28 08:12:32 <sillysausage> i made sure to plug in a keyboard incase that was stopping it from booting, (using good old leopold mechanical)
2015-04-28 08:12:56 <sillysausage> the adaptor was purchased with it and labeled as RP2 adaptor
2015-04-28 08:13:01 <sillysausage> i bought it from element14
2015-04-28 08:13:16 <sillysausage> so i doubt its a power issue
2015-04-28 08:13:28 <crow> sillysausage you used this img? 2015-04-20 16:44:08 <fabled> and the rpi image is at: http://dev.alpinelinux.org/~tteras/alpine-rpi-150420-armhf.rpi.tar.gz
2015-04-28 08:13:44 <sillysausage> nope
2015-04-28 08:13:48 <sillysausage> i used the one on the download page
2015-04-28 08:13:57 <sillysausage> i was just thinking about that
2015-04-28 08:14:01 <sillysausage> when looking at your output
2015-04-28 08:14:13 <fabled> sillysausage, rpi2 is not supported on the image on the official web imageuntil AL 3.2
2015-04-28 08:14:19 <fabled> use the dev.a.o link for now
2015-04-28 08:14:28 <crow> 2015-04-20 16:37:42 <fabled> and now uploading universal rpi image that works on all rpi models (a,a+,b,b+,b2)
2015-04-28 08:14:28 <crow> 2015-04-20 16:44:08 <fabled> and the rpi image is at: http://dev.alpinelinux.org/~tteras/alpine-rpi-150420-armhf.rpi.tar.gz
2015-04-28 08:14:54 <fabled> there's also http://dev.alpinelinux.org/~tteras/alpine-rpi-150424-armhf.rpi.tar.gz
2015-04-28 08:15:01 <sillysausage> alpine-rpi-3.1.3-armhf.rpi.tar.gz is not new enough?
2015-04-28 08:15:02 <sillysausage> ah
2015-04-28 08:15:02 <sillysausage> :D looks like that's the issue then
2015-04-28 08:15:03 <fabled> it has the sdcard mmc bug fixed
2015-04-28 08:15:16 <sillysausage> right shall do then
2015-04-28 08:15:24 <fabled> rpi2 needs newer kernel
2015-04-28 08:15:34 <sillysausage> ah
2015-04-28 08:15:35 <fabled> that is a kernel built specially for rpi2
2015-04-28 08:17:50 <sillysausage> i guess ill go with 150424 that's the newest one
2015-04-28 08:24:18 <sillysausage> feels pretty dirty not verifying an .asc :D
2015-04-28 08:24:18 <sillysausage> and not downloading over https
2015-04-28 08:24:18 <sillysausage> guess i'm inviting the NSA into my RP2
2015-04-28 08:24:18 <sillysausage> a little bit of QUANTUM here and there and everywhere
2015-04-28 08:24:19 <sillysausage> f35a7f5d1b37a7155154d1dfd0f5dd58208071ed524d160987a4d1057c9fa6d1  alpine-rpi-150424-armhf.rpi.tar.gz
2015-04-28 08:25:31 <sillysausage> https://en.wikipedia.org/wiki/QUANTUM as long as  it's a little black cat
2015-04-28 08:25:35 <sillysausage> and not that brown gremlin thing
2015-04-28 08:31:12 <sillysausage> how do i stop it trying to set the systemclock
2015-04-28 08:31:20 <sillysausage> because obviously rps don't have an RTC
2015-04-28 08:40:28 <sillysausage> to decrypt an encrypted provision file can i use something like:
2015-04-28 08:40:30 <sillysausage> openssl aes-256-cbc -d -a -in secrets.txt.enc -out secrets.txt.new
2015-04-28 08:43:02 <sillysausage> openssl enc -d -aes-256-cbc -in gateway.apkovl.tar.gz.aes-256-cbc -out gateway.apkovl.tar.gz
2015-04-28 08:43:04 <sillysausage> that worked :D
2015-04-28 09:10:37 <sillysausage> :(
2015-04-28 09:10:48 <sillysausage> looks like ppp-pppoe is not available form arm
2015-04-28 09:11:03 <sillysausage> how am i supposed to dial a ppp connection to my modem without that :(
2015-04-28 09:11:43 <sillysausage> fabled: i wonder why that is not there
2015-04-28 09:12:59 <sillysausage> http://dl-7.alpinelinux.org/alpine/v3.2/main/armhf/
2015-04-28 09:13:01 <sillysausage> http://dl-7.alpinelinux.org/alpine/edge/testing/armhf/
2015-04-28 09:13:06 <sillysausage> iputils is missing too
2015-04-28 09:19:26 <sillysausage> so that's ppp-pppoe, iputils, dhcp, unbound, ca-certificates, curl, tlsdate in short i need hmm
2015-04-28 09:21:03 <sillysausage> these packages are available for x86_64, how difficult is it to make arm versions?
2015-04-28 09:23:25 <sillysausage> http://wiki.alpinelinux.org/w/index.php?title=Special%3ASearch&profile=default&search=apk&fulltext=Search
2015-04-28 09:23:37 <sillysausage> i didn't see anything there about how to compile APKs from source
2015-04-28 09:26:01 <crow> http://wiki.alpinelinux.org/wiki/Abuild_and_Helpers
2015-04-28 09:28:30 <crow> sillysausage and http://wiki.alpinelinux.org/wiki/Include:Setup_your_system_and_account_for_building_packages
2015-04-28 09:29:35 <sillysausage> ah
2015-04-28 09:29:42 <sillysausage> i might build them then and see if i have any issues
2015-04-28 09:29:54 <sillysausage> im guessing they werent built cos someone didnt want them
2015-04-28 09:30:18 <sillysausage> i hope they don't need patching to work on arm :P
2015-04-28 09:31:55 <crow> well setup build stuff on rpi2 and build these packages yourself
2015-04-28 09:40:16 <crow> check if right ARCH inside APKBUILD is to match arm
2015-04-28 09:49:20 <fabled> sillysausage, setup-apkcache, and apk add
2015-04-28 09:49:30 <fabled> iputils, dhcp, unbound, ca-certificates and curl exist for sure
2015-04-28 09:49:39 <fabled> ppp-pppoe might have build failure
2015-04-28 09:50:06 <fabled> oh
2015-04-28 09:50:08 <fabled> sillysausage, use http://dl-7.alpinelinux.org/alpine/edge/main/armhf/
2015-04-28 09:50:13 <fabled> v3.2 is not built yet
2015-04-28 09:51:23 <crow> oh i missed that he did paste v3.2 :)
2015-04-28 10:24:19 <sillysausage> yeah
2015-04-28 10:24:19 <sillysausage> should i build them then fabled or can you do it and push it to 3.2 branch
2015-04-28 10:24:19 <sillysausage> because if i do it, it's not going to end up in the repo :P
2015-04-28 10:46:40 <crow> sillysausage what do you have in /etc/apk/repositories ?
2015-04-28 10:47:03 <crow> just do: setup-apkcache then: apk update && apk add ppp-ppoe
2015-04-28 10:50:33 <sillysausage> yeah i have the repositories, but it seems the packages just don't exist for the armf arch
2015-04-28 10:50:45 <sillysausage> i didn't use setup-apkcache but i did manually fill out /etc/apk/repositories
2015-04-28 10:50:50 <sillysausage> like i have working on my x86_64 machine
2015-04-28 10:52:14 <sillysausage> eg
2015-04-28 10:52:15 <sillysausage> localhost:~# cat /etc/apk/repositories
2015-04-28 10:52:18 <sillysausage> /media/mmcblk0p1/apks
2015-04-28 10:52:20 <sillysausage> http://dl-7.alpinelinux.org/alpine/v3.2/main
2015-04-28 10:52:22 <sillysausage> #http://dl-7.alpinelinux.org/alpine/edge/main
2015-04-28 10:52:24 <sillysausage> @testing http://dl-7.alpinelinux.org/alpine/edge/testing
2015-04-28 10:52:29 <sillysausage> i think it's because i need to build them
2015-04-28 10:54:32 <sillysausage> ah
2015-04-28 10:54:36 <sillysausage> i think its a mirror problem
2015-04-28 10:54:42 <sillysausage> because curl exists in http://nl.alpinelinux.org/alpine/v3.2/main/armhf/
2015-04-28 10:55:01 <sillysausage> i thought those mirrors were rsynced periodically?
2015-04-28 10:57:39 <sillysausage> so ca-certificates, curl are in there
2015-04-28 10:57:59 <sillysausage> ah ppp-pppoe is in there too
2015-04-28 10:58:36 <sillysausage> so looks like i still need to compile iputils, dhcp, unbound, tlsdate
2015-04-28 11:09:26 <sillysausage> crow: where do i pull the source packages from
2015-04-28 11:09:38 <sillysausage> doesn't abuild need some sort of spec file
2015-04-28 11:09:58 <sillysausage> i know these packages must have one because they exist for x86_64
2015-04-28 11:10:04 <sillysausage> i guess what im looking for is like a SRPM
2015-04-28 11:15:00 <sillysausage> https://ssl.nettworks.org/repo/browse/alpine-aports-mirror/main
2015-04-28 11:15:08 <sillysausage> ah that looks like what i was trying to find
2015-04-28 11:16:10 <sillysausage> oh actually i think the main one is git://dev.alpinelinux.org/aports
2015-04-28 11:19:56 <crow> sillysausage well your repository is wrong
2015-04-28 11:20:21 <crow> yous the one which was there before. you are downloading stuff for v3.2 which isnt even released.
2015-04-28 11:21:09 <crow> and the arm repository is not the same as the x86_64
2015-04-28 11:21:23 <crow> http://dl-7.alpinelinux.org/alpine/v3.2/main vs http://dl-7.alpinelinux.org/alpine/edge/main/armhf/
2015-04-28 11:22:21 <sillysausage> oh i've got both
2015-04-28 11:22:28 <sillysausage> the reason being is because dnscrypt is in testing
2015-04-28 11:22:53 <sillysausage> what i was saying dl-7.alpinelinux.org has less packages than nl.alpinelinux.org
2015-04-28 11:23:08 <sillysausage> i changed it the other day when there was the isp issues with the main one
2015-04-28 11:23:22 <sillysausage> i guess as this is a devel thing i might be best using the main one and not a mirror
2015-04-28 11:24:33 <crow> sillysausage then use this one http://nl.alpinelinux.org/alpine/edge/main/armhf/
2015-04-28 11:30:01 <sillysausage> yeah i am now
2015-04-28 11:30:10 <sillysausage> and its got most of the packages, just need to compile the last few
2015-04-28 11:30:16 <sillysausage> iputils, dhcp, unbound, tlsdate
2015-04-28 11:30:19 <sillysausage> and i'm good to go :D
2015-04-28 11:30:22 <sillysausage> ill tell you how it goes.
2015-04-28 11:30:53 <sillysausage> oh wait they are in there
2015-04-28 11:32:42 <crow> apk update after changing repositors
2015-04-28 11:33:32 <sillysausage> yup
2015-04-28 11:33:36 <sillysausage> well i just used the pinning method
2015-04-28 11:33:58 <sillysausage> least i know now how to compile things for alpine-linux :D
2015-04-28 12:34:37 <crow> which repository you pined and why?
2015-04-28 12:59:08 <remjey> Hi, just before I try and install alpine on my main PC, is there anybody here who's tried to use the ati proprietary graphics drivers ?
2015-04-28 12:59:43 <clandmeter> remjey: nobody
2015-04-28 13:00:52 <clandmeter> a more detailed answer would be, they will not work because they are linked against glibc.
2015-04-28 13:00:59 <ahills> remjey: Nope, but when I'm done applying for jobs and can get back to hacking, I'll be putting Alpine on a desktop with an AMD card
2015-04-28 13:01:22 <remjey> clandmeter: that is the kind of answer I was expecting, tbh
2015-04-28 13:02:00 <remjey> ahills: okay then, maybe I'll be doing to anyway
2015-04-28 13:02:01 <clandmeter> isnt the os drive in decent
2015-04-28 13:02:07 <remjey> *too
2015-04-28 13:02:26 <clandmeter> driver
2015-04-28 13:02:29 <remjey> clandmeter: it's decent but some steam games I have just won't display correctly
2015-04-28 13:02:42 <remjey> though necrodancer works fine
2015-04-28 13:02:45 <ahills> Will Steam run without glibc?
2015-04-28 13:02:57 <remjey> ahills: yeah in a chroot, according to the alpine wik
2015-04-28 13:02:58 <remjey> i
2015-04-28 13:03:45 <ahills> Ah, clever
2015-04-28 13:03:58 <ahills> Too bad you can't run the drivers in a chroot
2015-04-28 13:04:23 <ahills> You actually might be able to do something with KVM and a virtual machine
2015-04-28 13:05:45 <remjey> I'd prefer a more simple and lighter solution :p
2015-04-28 13:07:50 <clandmeter> remjey: i know somebody did some work on the nvidia drivers to get them working on musl. maybe there is somebody doing the same for ati
2015-04-28 13:08:02 <clandmeter> better ask in #musl
2015-04-28 13:46:01 <BitL0G1c> clandmeter - xf86-video-ati works for me on AMD Trinity / HD 7690
2015-04-28 13:47:37 <ahills> BitL0G1c: does that one actually perform well with 3D?
2015-04-28 13:48:02 <clandmeter> BitL0G1c: i was under the impression ati drivers are/were in a decent state, but i never play games.
2015-04-28 13:48:23 <BitL0G1c> not tried playing any games
2015-04-28 14:29:53 <ramonsoares> ncopa: about the postgresql patch, libpq need be the first subpackage or the package-dev will move all .h files.
2015-04-28 14:31:55 <clandmeter> ramonsoares: why do you want to ship the headers with the library?
2015-04-28 14:32:41 <ramonsoares> clandmeter: because i need to compile 3rd party packages
2015-04-28 14:33:03 <clandmeter> whats wrong with postgresql-dev?
2015-04-28 14:33:06 <ahills> ramonsoares: the -dev package is exactly for that...
2015-04-28 14:33:16 <ramonsoares> and -dev install half of world
2015-04-28 14:33:38 <ramonsoares> i only need headers i dont need contrib, doc, ...
2015-04-28 14:33:59 <clandmeter> we do not optimize our dev packages
2015-04-28 14:34:47 <ramonsoares> with my patch i can install only libpq package and i can build any 3rd party packages
2015-04-28 14:36:47 <ramonsoares> perhaps the best option to be a libpq-dev, but this conflict with -dev package build function
2015-04-28 14:39:44 <ramonsoares> i'm using alpine to build small docker containers and install all origin deps and subpackages with *-dev let my containers huge
2015-04-28 14:40:50 <clandmeter> i understnad your issue, but your patch isnt the correct fix.
2015-04-28 14:41:00 <ncopa> the -dev stuf is not intended for runtime
2015-04-28 14:41:50 <ncopa> the general idea is 'if you need -dev at runtime, then you are likely doing something wrong'
2015-04-28 14:41:54 <clandmeter> i do think on small devices, you can run into trouble with so many -dev deps.
2015-04-28 14:42:42 <ncopa> ramonsoares: it was a convenience thing that we pull in * with -dev
2015-04-28 14:43:24 <ncopa> we have deon that with everything else mostly to save maintenance work of properly splitting -dev packages
2015-04-28 14:44:02 <ramonsoares> i understand
2015-04-28 14:44:12 <ncopa> i suppose we could do an exception for libpq-dev
2015-04-28 14:44:35 <ncopa> but i am not really interested in doing manual splitting of all 2000+ -dev packages we have
2015-04-28 14:45:11 <ncopa> sorry
2015-04-28 14:45:13 <ncopa> its only 905
2015-04-28 14:45:27 <ramonsoares> maybe add a dev_deps=... in APKBUILD
2015-04-28 14:45:36 <ncopa> we have that
2015-04-28 14:45:43 <ncopa> depends_dev
2015-04-28 14:45:44 <nangel> isn't the "right way" to compile the 3rd party package into an apk?
2015-04-28 14:46:02 <nangel> so the build pulls in everything, but then the 3rd party apk doesn't need the headers.
2015-04-28 14:46:12 <ncopa> correct
2015-04-28 14:46:41 <ncopa> i think what ramonsoares is interested in is to do the compile step in a docker image
2015-04-28 14:46:52 <ncopa> and that docker image he wants keep as small as possible
2015-04-28 14:47:22 <ramonsoares> but in apkbuild generated to -dev pkgs include all origin subpackages
2015-04-28 14:47:46 <ncopa> yes, to save manual work
2015-04-28 14:48:19 <ncopa> basically, the -dev package is same as 'make install'
2015-04-28 14:49:19 <ahills> ramonsoares: if you have a special case, it is actually not too difficult to set up aports automatically, and just apply your patch to postgresql's APKBUILD
2015-04-28 14:49:27 <ahills> the apk system is very clean
2015-04-28 14:49:42 <ramonsoares> an ideia:
2015-04-28 14:49:58 <ramonsoares> subpackages="$pkgname-dev $pkgname-doc $pkgname-dbg ..."
2015-04-28 14:50:08 <ramonsoares> dev_subpackages="..."
2015-04-28 14:51:16 <ramonsoares> and in generated -dev apkbuild, only include packages in that list
2015-04-28 14:53:15 <ramonsoares> maybe can be a good workaround
2015-04-28 14:55:01 <ncopa> ramonsoares: is it only postgresql that is the problem?
2015-04-28 14:55:05 <ncopa> or is it in general
2015-04-28 14:55:30 <ncopa> oh
2015-04-28 14:55:44 <ncopa> its probably the -dbg stuff that makes things big
2015-04-28 14:55:53 <ramonsoares> readline-dev install docs...
2015-04-28 14:56:38 <ncopa> how bad is it
2015-04-28 14:56:43 <ncopa> how big does it get
2015-04-28 14:56:49 <ncopa> and how big would you expect it
2015-04-28 14:57:03 <ramonsoares> in my option -dev only need install link and headers files
2015-04-28 14:57:13 <ncopa> that is correct
2015-04-28 14:57:39 <ncopa> and potentially static libs
2015-04-28 14:58:25 <ramonsoares> the problem is it include all origin package subpackages
2015-04-28 14:58:51 <ramonsoares> maybe not need it
2015-04-28 14:59:32 <ncopa> becuase its convenient and save us the manual work to specify and maintain the exact list manually
2015-04-28 15:00:30 <ncopa> how big does your docker container become when you pull in postgresql-dev?
2015-04-28 15:01:32 <ncopa> i think the major problem is the -dbg
2015-04-28 15:02:23 <ramonsoares> nothing, I remove it after compiling.
2015-04-28 15:02:35 <ncopa> readline-doc is only 500k
2015-04-28 15:03:15 <ncopa> so its only a temp problem?
2015-04-28 15:03:24 <ncopa> the way we do it now saves us much work
2015-04-28 15:03:34 <ramonsoares> the "problem" is download 50MB to 800kb needed
2015-04-28 15:03:54 <ncopa> ok
2015-04-28 15:04:23 <ncopa> oh
2015-04-28 15:04:25 <ncopa> its late
2015-04-28 15:04:27 <ncopa> i need to go
2015-04-28 15:04:36 <ncopa> i think if we exclude the -dbg from -dev
2015-04-28 15:04:42 <ncopa> we would solve a lot of it
2015-04-28 15:06:12 <ramonsoares> i will play with abuild source later
2015-04-28 15:06:46 <ramonsoares> try some ideas to make it better
2015-04-28 15:07:15 <ncopa> i belive that the majority of those 50MB downloaded is the -dbg
2015-04-28 15:07:59 <ramonsoares> but that's not my point.
2015-04-28 15:08:04 <clandmeter> why do we include -doc in -dev?
2015-04-28 15:08:21 <ncopa> because it was simple to implement
2015-04-28 15:08:41 <ncopa> and -doc does not add that much overhead compared to gcc and other
2015-04-28 15:08:59 <ncopa> -doc normally don't add much overhead
2015-04-28 15:09:01 <clandmeter> i know, but we micro optimize eveything
2015-04-28 15:09:07 <ncopa> not everything
2015-04-28 15:09:09 <ncopa> not -dev :)
2015-04-28 15:09:26 <clandmeter> if there isnt a reason to include it, why not remove it?
2015-04-28 15:09:34 <ramonsoares> give me 1 or 2 day to play with abuild src
2015-04-28 15:10:04 <ramonsoares> maybe i can found a good way :p
2015-04-28 15:11:23 <ncopa> http://sprunge.us/WPaM
2015-04-28 15:11:44 <ncopa> unfortunally, the builders are running the v3.2 build
2015-04-28 15:11:51 <ncopa> we cannot really change that at this stage
2015-04-28 15:12:03 <ncopa> unless we want delay the v3.2 release with another week
2015-04-28 15:12:18 <ahills> ETA on 3.2?
2015-04-28 15:12:32 <ncopa> we should do the release next tuesday
2015-04-28 15:12:36 <ahills> awesome!
2015-04-28 15:12:42 <ncopa> but we squeezed in 3.18 kernel late
2015-04-28 15:12:52 <ncopa> and we had issues with arm kernel
2015-04-28 15:13:03 <ncopa> so we lost at least 1-2 weeks there
2015-04-28 15:13:12 <ahills> is the armhf only for rpi or for other similar arm platforms too?
2015-04-28 15:13:23 <ncopa> rpi, rpi2 and wandboard
2015-04-28 15:13:31 <ncopa> are the ones we have tested
2015-04-28 15:13:33 <ahills> ok
2015-04-28 15:13:49 <ahills> I will give it a shot with a beaglebone black soon
2015-04-28 15:13:58 <ncopa> nice
2015-04-28 15:14:19 <ncopa> ramonsoares: i believe http://sprunge.us/WPaM will solve 80% of the problem
2015-04-28 15:14:26 <ramonsoares> yep
2015-04-28 15:14:37 <ncopa> without adding any manual maintenance work
2015-04-28 15:15:11 <ncopa> its just too bad we didnt do it 1 month ago
2015-04-28 15:15:21 <ramonsoares> :p
2015-04-28 15:15:27 <darkfader> wow
2015-04-28 15:15:33 <ramonsoares> i start play with alpine this week
2015-04-28 15:15:43 <ncopa> seriously, i really dont want delay v3.2 release for this
2015-04-28 15:15:45 <darkfader> if i could switch my arm thinggies to alpine that would make my life so much better
2015-04-28 15:16:00 <ncopa> but i willing to push the change soon after v3.2
2015-04-28 15:16:09 <ncopa> ok i really have to go now
2015-04-28 15:16:10 <ncopa> see u
2015-04-28 15:16:12 <darkfader> ncopa: bye
2015-04-28 15:16:14 <ncopa> and thanks for you feedback
2015-04-28 15:16:22 <ramonsoares> ncopa: o/
2015-04-28 15:16:23 <ahills> alpine has made my linux life so much better
2015-04-28 15:16:43 <ahills> now if only those damn upstream developers would stop depending on glibc!
2015-04-28 15:17:01 <clandmeter> they do?
2015-04-28 15:17:38 <clandmeter> they should target the correct def
2015-04-28 15:19:59 <ahills> some packages, e.g. Guile, only build against glibc
2015-04-28 15:20:21 <ahills> for Guile it's specifically because their garbage collector library they depend on doesn't work with musl
2015-04-28 15:20:38 <ahills> guile 2.0+ I should specify
2015-04-28 15:20:46 <ahills> that may have changed, I haven't tried to build it in a while
2015-04-28 15:20:59 <clandmeter> yeah, some need glibc specifics, although i dont see it that much.
2015-04-28 15:21:09 <ahills> nope, and definitely not for my servers
2015-04-28 15:21:17 <ahills> I haven't run into a single package that won't build for those
2015-04-28 17:43:09 <bitshark__> hey
2015-04-28 17:46:38 <bitshark> my laptop distro is almost dead, haha
2015-04-28 17:47:26 <bitshark> somehow ubuntu 14.04 got its apt sources changed to 14.10... so its halfway between the two revisions and barely workz
2015-04-28 17:47:26 <bitshark> i gotta reinstall
2015-04-28 17:48:22 <bitshark> trying to decide (A) whether to use alpine or my custom-compiled linux from scratch, and (B) whether to layer nix on top of alpine apt , or whether to build the entire os and toolchain as nix packages in haskall
2015-04-28 17:48:31 <bitshark> oh and who can forget most importantly
2015-04-28 17:48:55 <bitshark> (C) what window manager i should use which lookz the illest but has the least dependencies on systemd now and forever
2015-04-28 17:49:11 <bitshark> maybe like enlightmentment or awesomewm or sumt
2015-04-28 17:49:14 <bitshark> fluxbox
2015-04-28 17:49:16 <bitshark> i dunno
2015-04-28 17:50:30 <bitshark> goona go evaluate some window managers
2015-04-28 17:57:49 <clandmeter> bitshark: dunno about wm, but alpine has good support for xfce4
2015-04-28 17:57:59 <ahills> bitshark: did you look at dwm?
2015-04-28 17:58:06 <ahills> it's pretty "ill"
2015-04-28 18:19:03 <bitshark> hey hey
2015-04-28 18:19:13 <bitshark> ahills:  i tried dwm but i was confused
2015-04-28 18:19:24 <bitshark> it gave me like 10 buckets at the top of the screen at that was it
2015-04-28 18:19:48 <bitshark> plus my computer is on the fritz becuase it's halfway upgraded bwetween ubuntu 14.04 and 14.10
2015-04-28 18:20:13 <bitshark> so like , icewm wouldnt work, but lxde would work
2015-04-28 18:20:22 <bitshark> it was random
2015-04-28 18:21:03 <bitshark> so that's why i'm reinstalling my OS...  it's been getting worse for a few weekz cause the apt repo is corrupt
2015-04-28 18:24:09 <bitshark> i'm more trying to pick a laptop platform for like everydays tasks and for work.... criteria is (A) no windows ,  (B) no systemd or stuff related to it ,  (C) a nice desktop / GUI , can be minimalits,  (D) need to be able to run developer stuff like PyCharm and WebStorm etc  (E) Stuff should work out of the box and be fast,  (F)  Good package management,  (G) Patched kernel to support ZFS on linux
2015-04-28 18:24:25 <_ikke_> bitshark: What is the problem with systemd :P
2015-04-28 18:24:29 <bitshark> i actualy briefly considered FreeBSD (thru PC-BSD, the friednly freebsd) ... And i got installed to a USB stick and it didn't support my 3 year old laptop
2015-04-28 18:24:36 <bitshark> video driver ... i had to use vesa
2015-04-28 18:25:11 <bitshark> so FreeBSD is out , even tho i like it, but it's hardware support for laptops is awful.. it's an intel i915 chipset... super common
2015-04-28 18:25:14 <_ikke_> I use awesomewm and happy with it
2015-04-28 18:26:29 <ahills> awesomewm and dwm both you will need to read the manual
2015-04-28 18:26:38 <ahills> they have good manuals, but they are not GNOME clones
2015-04-28 18:27:08 <bitshark> anywayz i'm down do either (A) Alpine Linux , with or without addition of Nix as a package manager.... or (B) convert my 64-it LFS build into a USB stick installer, then use nix package management (or maybe gentoo portage or arch pacman)  on top of the LFS base
2015-04-28 18:27:25 <bitshark> okay
2015-04-28 18:27:40 <ahills> what is nix and why wouldn't you just use apk?
2015-04-28 18:27:42 <bitshark> yeah i'm most interested in the window manager so thanks
2015-04-28 18:27:56 <bitshark> oh it's a side project i will explain in a sec
2015-04-28 18:28:16 <bitshark> i saw the stuff about dwm and awesomewm and i liked how it make the windows take up screen real estate
2015-04-28 18:28:17 <ahills> ok, don't worry about that, I now know that your question is outside my scope
2015-04-28 18:28:35 <bitshark> instead off getting hidden... but it seems like u gotta edit config files to have it look the way you like
2015-04-28 18:28:39 <ahills> tiling window managers
2015-04-28 18:28:45 <_ikke_> I only have a 20px top bar, the rest is all used for windows
2015-04-28 18:28:51 <ahills> dwm doesn't have a config file per se
2015-04-28 18:28:54 <ahills> you edit the header and recompile
2015-04-28 18:28:55 <_ikke_> lua is very easy to learn
2015-04-28 18:28:59 <ahills> yeah it is
2015-04-28 18:29:22 <ahills> alt-shift-enter to launch a new terminal in dwm, bitshark
2015-04-28 18:29:22 <bitshark> yeah... that sounds pretty sick..  i think to get a feel for it , 'll need to try it with a known working config or on a live cd or sumt
2015-04-28 18:29:24 <ahills> that should get you started
2015-04-28 18:29:29 <ahills> alt-p to run a program
2015-04-28 18:29:33 <bitshark> so i can see what it could be
2015-04-28 18:29:33 <ahills> man dwm for the rest
2015-04-28 18:29:37 <bitshark> oh yeah that was the other thing
2015-04-28 18:29:39 <_ikke_> bitshark: Yeah, I recommend starting with the default config
2015-04-28 18:29:42 <bitshark> i forgot about that
2015-04-28 18:29:43 <bitshark> lua
2015-04-28 18:29:44 <ahills> the builtin dwm config is 99% fine
2015-04-28 18:29:54 <_ikke_> ahills: Do you use alt as mod key?
2015-04-28 18:29:58 <ahills> you may not like firefox being sent to tag 9
2015-04-28 18:29:58 <ahills> yep
2015-04-28 18:30:02 <bitshark> so u can make plugins i suppose
2015-04-28 18:30:08 <_ikke_> I use the 'flag'
2015-04-28 18:30:09 <bitshark> yeah
2015-04-28 18:30:14 <bitshark> i have alot of mod keys
2015-04-28 18:30:18 <ahills> that's too far for my thumb :P
2015-04-28 18:30:19 <bitshark> to go in and out of vms
2015-04-28 18:30:27 <bitshark> cuz they catch the keyboards attention on focus
2015-04-28 18:30:31 <_ikke_> but alt is used in many other programs
2015-04-28 18:30:36 <bitshark> and i dont use mouse
2015-04-28 18:30:37 <ahills> not my programs
2015-04-28 18:30:41 <bitshark> well not much
2015-04-28 18:31:01 <bitshark> i use the defaults tho
2015-04-28 18:31:09 <bitshark> except for one change prety much
2015-04-28 18:31:12 <ahills> default programs?
2015-04-28 18:31:19 <bitshark> ctrl-alt-arrowkey for desktop
2015-04-28 18:31:24 <bitshark> no for hotkeys
2015-04-28 18:31:29 <_ikke_> I never go to the desktop
2015-04-28 18:31:33 <ahills> what's that the default for
2015-04-28 18:31:53 <bitshark> like ctrl-alt-arrow changes virtual desktops in most every gnome-derived wm
2015-04-28 18:32:09 <_ikke_> awesome uses tags
2015-04-28 18:32:20 <_ikke_> mod + [1-9]
2015-04-28 18:32:20 <ahills> awesome and dwm are not gnome-derived ;)
2015-04-28 18:32:22 <bitshark> okay
2015-04-28 18:32:29 <ahills> there's no "desktop"
2015-04-28 18:32:31 <bitshark> yeah that's why i'm interested in them
2015-04-28 18:32:32 <bitshark> heheh
2015-04-28 18:32:59 <_ikke_> I have fixed tags for certain programs
2015-04-28 18:33:00 <bitshark> cause the way i look at it, the further away i am from gnome, polkt ,etc... the less systemd is gonna affect me for next 48-36 months
2015-04-28 18:33:39 <_ikke_> I use arch, one of the early adopters for systemd, and really like it
2015-04-28 18:33:44 <bitshark> k
2015-04-28 18:33:49 <ahills> I liked systemd until it massacred my server
2015-04-28 18:33:57 <_ikke_> What happened
2015-04-28 18:33:59 <ahills> Now I run my Arch machines with OpenRC instead
2015-04-28 18:34:07 <buckley310> i like systemd for desktops but i have had bad experiences with it on servers
2015-04-28 18:34:17 <_ikke_> Running it on my server without problems
2015-04-28 18:34:18 <ahills> long story short, journald crashed but would only log its crashes to the journal, and journalctl couldn't retrieve the journal
2015-04-28 18:34:32 <ahills> journald crashed every time it tried to start, but there was no way to find out how or why
2015-04-28 18:34:33 <bitshark> well i dont want to make this into systemd... i dont think less of ppl that use it, but it's not 4 me, heh
2015-04-28 18:34:34 <bitshark> and i like gentoo and openrc
2015-04-28 18:34:35 <bitshark> yeah?
2015-04-28 18:34:45 <_ikke_> sure
2015-04-28 18:34:54 <bitshark> man that's unfortunate
2015-04-28 18:35:06 <bitshark> i remeember reading about how they buitl that journal from scratch
2015-04-28 18:35:12 <bitshark> instead of using something ots ...
2015-04-28 18:35:21 <bitshark> why reinvent the wheel it just makes bugz
2015-04-28 18:35:23 <bitshark> heh
2015-04-28 18:35:24 <ahills> and it turns out that as a part of that, or maybe unrelated but just at the same time, it erased my raid configuration
2015-04-28 18:35:31 <bitshark> oh gawd
2015-04-28 18:35:38 <bitshark> that's like the stuff of work nightmares
2015-04-28 18:35:42 <ahills> so I had to manually extract important data from the filesystem
2015-04-28 18:35:46 <ahills> now my server runs alpine
2015-04-28 18:35:46 <_ikke_> ahills: Isn't it perhaps the other way around?
2015-04-28 18:35:59 <_ikke_> problems with your raid config causing journald to crash?
2015-04-28 18:36:05 <ahills> _ikke_: fortunately only the /usr/share/man directory and subdirectories seemed to have been affected
2015-04-28 18:36:14 <bitshark> "i wake up in a cold sweat.. the raid array said RAID0 not RAID1-0... this couldnt be happening..."
2015-04-28 18:36:37 <bitshark> haha mann
2015-04-28 18:36:58 <bitshark> well regardless of whether its systemd or not, i didnt want it  , even if it's not problematic
2015-04-28 18:37:14 <bitshark> and for like data integrity
2015-04-28 18:37:20 <bitshark> i'm all over ZFS on linux
2015-04-28 18:37:30 <_ikke_> One issue I have with openrc is that it doens't detect when a service is stopped outside of it's control (at least, with the services I used)
2015-04-28 18:37:46 <bitshark> instead of RAID
2015-04-28 18:37:46 <ahills> if you want that kind of oversight, consider daemontools
2015-04-28 18:38:07 <bitshark> hmm..
2015-04-28 18:38:32 <bitshark> i guess i'm just used to the status quo  ... where sysvinit was a bunch of shell scripts and after that you're on your own
2015-04-28 18:38:53 <bitshark> it's obviously not innovative or future proof, heh
2015-04-28 18:38:55 <ahills> systemd added three or four nice features
2015-04-28 18:39:12 <ahills> openrc and daemontools together provide all I need, without the other problems, like pwning my syslog
2015-04-28 18:39:46 <bitshark> but i just think that there are enuf brilliant ppl around in the linux / bsd communities, there could have been a better solution w/o redhat's 2 gold star ppl who messed up pulseaudio for like 2 years
2015-04-28 18:39:50 <ahills> _ikke_: the raid problem is a mystery because the journal corruption meant I didn't get any information about the system leading up to the incident
2015-04-28 18:40:06 <_ikke_> Yeah, I can imagine
2015-04-28 18:40:20 <ahills> i also was running on JFS, which was a mistake
2015-04-28 18:40:38 <ahills> indefinitely delaying journal writes meant I almost lost data out of the raid problem
2015-04-28 18:40:47 <bitshark> ahills: what RAiD were u using
2015-04-28 18:40:53 <bitshark> and hw or software
2015-04-28 18:40:57 <bitshark> and did hw have batteriez
2015-04-28 18:41:09 <ahills> software
2015-04-28 18:41:11 <ahills> mdadm
2015-04-28 18:41:14 <bitshark> k
2015-04-28 18:41:14 <ahills> it works great
2015-04-28 18:41:20 <bitshark> what mode
2015-04-28 18:41:23 <ahills> mirrored
2015-04-28 18:41:27 <bitshark> ohh
2015-04-28 18:41:33 <bitshark> that shouldnt be such a big ris
2015-04-28 18:41:35 <bitshark> k
2015-04-28 18:41:41 <bitshark> on ssds or hdds
2015-04-28 18:41:50 <ahills> right, well, it turns out one of the disks was literally two months behind the other
2015-04-28 18:42:00 <bitshark> so hdds
2015-04-28 18:42:01 <bitshark> ?
2015-04-28 18:42:01 <_ikke_> whut
2015-04-28 18:42:02 <ahills> and the two months was exactly the amount of missing journal information
2015-04-28 18:42:18 <bitshark> ahh
2015-04-28 18:42:26 <ahills> in any case, that plus systemd upgrades requiring a reboot means I don't run systemd on servers anymore
2015-04-28 18:42:31 <bitshark> so it might implicate systemd, but its not definitive
2015-04-28 18:42:45 <ahills> it's just a correlation
2015-04-28 18:42:46 <bitshark> could also be the kernel
2015-04-28 18:42:48 <ahills> the events are related
2015-04-28 18:42:49 <bitshark> or the hd
2015-04-28 18:42:50 <bitshark> yeah
2015-04-28 18:42:54 <bitshark> i see
2015-04-28 18:42:56 <ahills> it wasn't the HD
2015-04-28 18:43:00 <ahills> the HDs are both happily running alpine
2015-04-28 18:43:22 <crow> ahills openrc on archlinux?
2015-04-28 18:43:24 <ahills> I've revised my server architecture considerably, though, moving to xen
2015-04-28 18:43:30 <ahills> yeah my arch servers run openrc
2015-04-28 18:43:38 <ahills> it's just way easier to use
2015-04-28 18:43:45 <bitshark> yeah
2015-04-28 18:43:49 <bitshark> i love openrc
2015-04-28 18:43:54 <ahills> even on servers that don't require uptime
2015-04-28 18:43:57 <bitshark> so much better than sysvinit
2015-04-28 18:44:05 <bitshark> ajd no controversy of systemd
2015-04-28 18:44:58 <bitshark> i dig openrc as well, espeially for servers
2015-04-28 18:45:15 <erethon> am I missing something obvious here? I can't boot alpine from a usb drive created by dd'ing the .iso to the drive
2015-04-28 18:45:23 <bitshark> hm
2015-04-28 18:45:29 <bitshark> u shoould be able to
2015-04-28 18:45:32 <_ikke_> erethon: The way they build the image only works for cdroms
2015-04-28 18:45:35 <bitshark> altho i havent tried it with albine
2015-04-28 18:45:36 <ahills> erethon: yeah, you need to process the ISO further
2015-04-28 18:45:45 <ahills> erethon: it's on the wiki but might not be on the front, hold on
2015-04-28 18:45:45 <bitshark> it depends on how the ISO is created
2015-04-28 18:45:52 <clandmeter> erethon: our isi is not hybrid
2015-04-28 18:45:56 <bitshark> oh
2015-04-28 18:46:05 <bitshark> so there's your answer.. no u cant without a few quick mods
2015-04-28 18:46:06 <clandmeter> you need to copy the contents to fat32 and use syslinux
2015-04-28 18:46:28 <ahills> http://wiki.alpinelinux.org/wiki/Installing_Alpine_on_USB
2015-04-28 18:46:37 <ahills> http://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB
2015-04-28 18:46:39 <ahills> one of those two
2015-04-28 18:46:40 <bitshark> man there really should be an alpine image that you can dd
2015-04-28 18:46:47 <ahills> I need to get back to work or I'd be more helpful
2015-04-28 18:46:50 <_ikke_> ahills: Do you know systemctl daemon-reexec
2015-04-28 18:46:51 <erethon> that makes sense, although a link from the wiki on installing on usb, redirected me to the Arch wiki for dd etc
2015-04-28 18:46:52 <_ikke_> ?
2015-04-28 18:46:52 <bitshark> and also a livecd for ppl to try out
2015-04-28 18:47:06 <bitshark> so ahills u asked me about 'why would [i] use a different package manager (called nix) instead of apk?'   short answer:  check out nix
2015-04-28 18:47:18 <bitshark> it's stateful package mgmt .. no circular dependencies
2015-04-28 18:47:21 <erethon> anyway, thank you guys
2015-04-28 18:47:26 <bitshark> it's like container technology for package mgmt
2015-04-28 18:47:34 <bitshark> http://nixos.org
2015-04-28 18:48:05 <ahills> _ikke_: yes, what about it?
2015-04-28 18:48:11 <bitshark> i like the premise but their OS is bloated .. . hence why i hang out here
2015-04-28 18:48:11 <bitshark> hah
2015-04-28 18:48:25 <bitshark> well i just built a 64 bit non-systemd CLFS system
2015-04-28 18:48:28 <bitshark> linux from scratch
2015-04-28 18:48:32 <ahills> ok, well, have you tried apk?
2015-04-28 18:48:33 <_ikke_> ahills: It seems that it alows you to update systemd without rebooting
2015-04-28 18:48:36 <bitshark> i was considering using nix to build on top of that
2015-04-28 18:48:39 <ahills> it is probably close to the perfect package manager
2015-04-28 18:48:40 <bitshark> yeah i like apk
2015-04-28 18:48:41 <bitshark> alot
2015-04-28 18:48:45 <ahills> also, I don't want to make my own distribution
2015-04-28 18:48:51 <bitshark> i just mean nix lets u do stuff u can do in apk
2015-04-28 18:48:55 <ahills> clandmeter and ncopa do a lot of work so I don't have to
2015-04-28 18:49:39 <bitshark> oh see i dont mind... like 'my own distro' is actually preferable for some clients because it means u aren't dependent on canonicalt to keep those repos up
2015-04-28 18:50:17 <bitshark> so i dont know if nix is any good in 'real life' but i plan on doinga test run
2015-04-28 18:50:21 <bitshark> soon
2015-04-28 18:50:22 <_ikke_> With aports, it's really easy to build your own packages
2015-04-28 18:50:24 <ahills> yes, but if I've learned anything about running servers over the past decade and a bit, it's that organizations of multiple people are generally more reliable than one-man organizations like myself
2015-04-28 18:50:35 <bitshark> either in alpine or in my clfs box
2015-04-28 18:50:46 <bitshark> yeah
2015-04-28 18:51:05 <bitshark> thatz true... but there's all sorts of complexity as im sure u know
2015-04-28 18:51:22 <bitshark> security ppl, billing ppl, legacy systems, customer needs, etc etc
2015-04-28 18:51:27 <bitshark> flexibility always helps
2015-04-28 18:51:37 <bitshark> but yeah the nix is an experiment
2015-04-28 18:51:51 <bitshark> not like a proclamation... i have no idea whether its practical
2015-04-28 18:51:59 <bitshark> but the devops ppl seem  to like it
2015-04-28 18:52:17 <bitshark> but i heard u gotta go, so if u do, thx for the info
2015-04-28 18:53:19 <bitshark> anyway with apk... i like it enough i started building a GUI in Qt5
2015-04-28 18:53:24 <bitshark> to answer someone else's question
2015-04-28 18:53:24 <bitshark> heheh
2015-04-28 18:53:43 <bitshark> ikke: what's aports... ive heard of it
2015-04-28 18:53:52 <bitshark> is it worth reading about
2015-04-28 18:54:47 <_ikke_> bitshark: It's alpines build system
2015-04-28 18:55:48 <Tekki> How can one fix things when the running kernel (uname -r 3.14.37-0-grsec) and the installed modules (ls /lib/modules 3.14.37-3-grsec) mismatch, don't get loaded, and I cannot get internet/USB working to try to fix things through apk
2015-04-28 18:56:08 <_ikke_> Tekki: Usually reboot
2015-04-28 18:57:17 <_ikke_> The modules are already updated, but you are still runing the older kernel
2015-04-28 18:58:48 <Tekki> I've already rebooted thrice, hasn't fixed anything, I tried to continue this .. adventure of an install after a pause of several days, and this is the problem I'm facing now
2015-04-28 18:59:33 <clandmeter> check your exlinux.conf
2015-04-28 19:00:52 <clandmeter> sorry extlinux.conf
2015-04-28 19:02:29 <clandmeter> it could be that it didnt update the bootloader correctly. in that case i would boot alpine iso and chroot and apk fix the kernel
2015-04-28 19:09:40 <Tekki> Might have to try that then, thanks in any case
2015-04-28 19:14:12 <bitshark> ikke: thanks
2015-04-28 19:14:21 <bitshark> i'll definitely look at it
2015-04-28 19:14:41 <bitshark> Tekki sometimes theres a crypto key
2015-04-28 19:14:45 <bitshark> for modules
2015-04-28 19:14:51 <bitshark> an openssl cert
2015-04-28 19:14:54 <bitshark> did u compile the kernel yourself
2015-04-28 19:15:10 <bitshark> i always disable that crap
2015-04-28 19:15:39 <Tekki> Uh, no, just did a minimal install for my laptop. I'm still very much a learning user.
2015-04-28 19:16:11 <bitshark> oh okay
2015-04-28 19:16:11 <bitshark> minimal install of alpine?
2015-04-28 19:16:11 <bitshark> and the kernel versions mismatch?
2015-04-28 19:16:14 <Tekki> I've had plenty of stuff to figure out already that's been giving me headaches, haven't had the time to jump to compiling stuff yet
2015-04-28 19:16:23 <bitshark> oh no i understand completely
2015-04-28 19:16:26 <bitshark> dont do it! hah
2015-04-28 19:16:39 <bitshark> i just asked cuz i had that problem the other night and pulled like chunks of hair out as well
2015-04-28 19:17:00 <bitshark> where the kernel modules now need to hae a digital signature in new kernels to get loaded
2015-04-28 19:17:15 <_ikke_> since when?
2015-04-28 19:17:21 <bitshark> but itz brand new feature i think
2015-04-28 19:17:35 <bitshark> well brand new in last 18-24 months
2015-04-28 19:17:45 <bitshark> but i havent recompiled a kernel in about 5 years til yesterday so
2015-04-28 19:17:54 <bitshark> i used lets see.. 3.14.xx
2015-04-28 19:17:55 <_ikke_> I never compiled a kernel myself
2015-04-28 19:18:11 <bitshark> and i took the config file from ubuntu 14.04 as a 'base'
2015-04-28 19:18:32 <bitshark> there were some new options.. including a few related to kernel modules and crypto signing them
2015-04-28 19:18:48 <bitshark> i noticed because during the compile it failed out in the logs... because i didnt have openssl installed
2015-04-28 19:19:09 <bitshark> make (in the kernel directory)   wanted acccess to openssl to generate certificates to sign all the kernel modules
2015-04-28 19:19:22 <bitshark> so they dont get attacked by malware / backdoorz etc
2015-04-28 19:19:42 <bitshark> but i didnt have openssl and it sounded like a recipe for a headache so i turned it off
2015-04-28 19:19:47 <bitshark> the feature
2015-04-28 19:21:18 <bitshark> i made an interesting chart on building linux system from scratch...  kernel compile was the third longest task
2015-04-28 19:21:22 <bitshark> at about 20-30 minutes
2015-04-28 19:21:50 <bitshark> longest was third pass of gcc (where you use a compiler to compile the compiler for the 3rd tme)  ... that took almost 3 hours
2015-04-28 19:21:54 <bitshark> haha
2015-04-28 19:22:07 <bitshark> on an 8gb quad cpu KVM vps
2015-04-28 19:41:30 <Tekki> Um, now after I tried using the ISO, doing the basic RAM setup, installing cryptsetup to open my LUKS / partition, mounting the / and /boot partitions and using chroot instructions from the wiki
2015-04-28 19:41:53 <Tekki> I'm trying to use apk and I'm getting "ERROR: Unable to lock database: No such file or directory" + "ERROR: Failed to open apk database: No such file or directory"
2015-04-28 19:45:12 <BitL0G1c> I could never get apk to work in a chroot either - but a chroot is not necessary for an encrypted install
2015-04-28 19:46:47 <Tekki> BitL0G1c, I'm trying to fix my basic install that has the kernel and kernel modules mismatching (modules don't get loaded, internet/USB doesn't work)
2015-04-28 19:48:28 <bitshark> wait what
2015-04-28 19:48:31 <Tekki> I think I managed to get apk working earlier because I had to do this chrooting too when having trouble with initramfs, this error is new to me in any case and I'm pretty confused about how to continue
2015-04-28 19:48:45 <bitshark> just pass in stuff to your chroot
2015-04-28 19:48:49 <bitshark> with --bind
2015-04-28 19:49:12 <bitshark> you need to pass in your binaries that u want in the chrroot.. and u also need to pass in any shared libs into the chroot
2015-04-28 19:49:20 <bitshark> u can find them by doing like ldd cryptsetup
2015-04-28 19:49:30 <bitshark> then passing /bin , /lib, etc into the chroot
2015-04-28 19:49:42 <bitshark> but really alpine needs to have all this stuff automatic , wow
2015-04-28 19:49:53 <bitshark> maybe i should do some of this work if i use it as my main OS
2015-04-28 19:50:02 <bitshark> cause that sounds pretty hard to set up
2015-04-28 19:50:21 <bitshark> compared to typical distros like ubuntu
2015-04-28 19:51:21 <BitL0G1c> maybe --bind mount proc & dev
2015-04-28 19:51:34 <bitshark> mkdir /mnt/mychroot/bin;  mount --bind /bin /mnt/mychroot/bin;  mmkdir -p /mnt/mychroot/lib; mount --bind /mnt//mychroot/lib;   sudo chroot /mnt/mychroot
2015-04-28 19:51:35 <Tekki> Uhh.. So, first I created a directory /mnt/chroot, where I mounted my /dev/mapper/cryptedroot partition and then mounted /boot partition to /mnt/chroot/boot - after that I did the mount --bind stuff found here http://wiki.alpinelinux.org/wiki/LVM_on_LUKS (bottom of the page)
2015-04-28 19:51:57 <bitshark> yeah
2015-04-28 19:52:22 <bitshark> make sure that you run ldd on whatever binaries u want to runin the chrot
2015-04-28 19:52:22 <bitshark> like ldd cryptsetup
2015-04-28 19:52:22 <bitshark> make sure that it can 'see' its shared libaries
2015-04-28 19:52:38 <bitshark> btw i have never set up encrypted root partition theway u are doing
2015-04-28 19:53:08 <bitshark> I always do the following...    set up 4 partitions... 1 swap, 1 'root' , 1 'home' and 1 'misc'
2015-04-28 19:53:16 <BitL0G1c> chroot is not necessary for an encrypted install
2015-04-28 19:53:33 <bitshark> anyway i essentially install in plaintext then do the encrypted home afterwards
2015-04-28 19:53:45 <bitshark> by copying stuff around
2015-04-28 19:54:15 <bitshark> so tekki dont feel bad man, u are clearly very skilled
2015-04-28 19:54:31 <bitshark> i think bitlogicis right,i think chroot is just one way out of a few
2015-04-28 19:54:51 <bitshark> the way i do it is put two partitions side by side...  aka i install like normal but leave one free partition
2015-04-28 19:55:38 <Tekki> I still feel very stupid after all this, I would get nothing done without google-fu or wikis, and it's really hard for a newcomer to use Alpine without a wiki as comprehensive as Arch's (this is not a complaint, I understand the work needed for wikis)
2015-04-28 19:55:58 <bitshark_> the way i do it is put two partitions side by side...  aka i install like normal but leave one free partition
2015-04-28 19:56:11 <bitshark_> then i run like losetup and cryptsetup
2015-04-28 19:56:19 <bitshark_> i encrypt and mount the second parttion
2015-04-28 19:56:24 <bitshark_> copy all my files over from /
2015-04-28 19:56:51 <bitshark_> then close it... put an entry in fstab and crypttab for it... then i think i re-run grub to point grub at the new partition
2015-04-28 19:57:14 <bitshark_> or else i just edit grub.cfg and change the nubmer to point to the new parition
2015-04-28 19:57:39 <bitshark_> i think that's how i did it
2015-04-28 19:57:39 <bitshark_> anyway dont follow my procedure
2015-04-28 19:57:42 <BitL0G1c> i wrote a script to make encrypted disk installs easier - it's linked at the top of the lvm luks page
2015-04-28 19:57:46 <bitshark_> just be in mind there's like a bunch of ways to do ths. . . and honestly setting up encrypted /home is more important than FDE
2015-04-28 19:58:13 <BitL0G1c> i think with ram only install you can add packages manually to the cache
2015-04-28 20:01:25 <Tekki> BitL0G1c, I skipped any scripts for two reasons: I want to learn to do stuff (the previous installation on my laptop was LVM/LUKS Arch so I kind of did this thing earlier) and secondly I'm not a coder and can't really understand scripts - I want to see what they do before I use anything.
2015-04-28 20:01:52 <Tekki> I did the install that way the wiki page says, with LUKS and just skipped all the LVM stuff - I used BTRFS for the partitions, I can do resizing that way
2015-04-28 20:02:48 <bitshark_> oh hang on
2015-04-28 20:02:50 <Tekki> With this chrooting I'm just trying to fix the base minimal install, it's booting up but needs fixing because modprobe says the kernel modules fail to load
2015-04-28 20:03:01 <bitshark_> u arent tryig to resize an encrypted disk r u?
2015-04-28 20:03:05 <bitshark_> just checkin..
2015-04-28 20:03:10 <bitshark_> cuz that wont work
2015-04-28 20:04:09 <Tekki> No, I don't have any need to resize it no, it's an old Intel 40GB SSD, 200MB for /boot, rest for /
2015-04-28 20:04:15 <bitshark_> okay cool
2015-04-28 20:04:59 <bitshark_> well the process is basically like this... get your partition sized... get the device name.. "sda2"  ... then to be really fancy u overwrite it with zeros or with /dev/urandom
2015-04-28 20:05:06 <bitshark_> i always skip that cause it takes forever
2015-04-28 20:05:15 <bitshark_> and it's not that much extra security
2015-04-28 20:05:32 <Tekki> I used the wiki suggestion to wipe the partition with haveged, it's all been done already
2015-04-28 20:05:36 <bitshark_> then i do losetup
2015-04-28 20:05:40 <bitshark_> i think
2015-04-28 20:05:42 <bitshark_> no i dont
2015-04-28 20:06:00 <bitshark_> then i do cryptsetup luksOpen /dev/sda2 sda2
2015-04-28 20:06:13 <bitshark_> erm no first u do..
2015-04-28 20:06:19 <bitshark_> cryptsetup luksCreate /dev/sda2
2015-04-28 20:06:21 <bitshark_> there we go
2015-04-28 20:06:25 <bitshark_> then do the prompts
2015-04-28 20:06:33 <bitshark_> and once u put in key and all that crap
2015-04-28 20:06:43 <bitshark_> u get like /dev/mapper/sda2
2015-04-28 20:06:59 <bitshark_> u run your format on /dev/mapper/sda2  ... so like mkfs.ext2 /dev/mapper/sda2
2015-04-28 20:07:03 <bitshark_> then mount it if u want
2015-04-28 20:07:07 <bitshark_> then unmount it
2015-04-28 20:07:19 <bitshark_> then you finiish with cryptsetup luksClose /dev/sda2
2015-04-28 20:07:38 <bitshark_> to get it to prompt for pw on startup you need to edit /etc/fstab and /etc/cryptdisks
2015-04-28 20:08:09 <bitshark_> my encrypted home directory is on /chome and is symlinked to /home
2015-04-28 20:08:18 <bitshark_> the fdisk entry looks like this
2015-04-28 20:08:32 <bitshark_> .  /dev/mapper/chome	/chome	xfs	rw,discard,noatime,nodiratime,nobarrier00.
2015-04-28 20:08:50 <bitshark_> i also have encryptd swap... the fstab entry lookz like this ...
2015-04-28 20:09:04 <bitshark_> . /dev/mapper/swap	none	swap	sw,discard	0	0
2015-04-28 20:09:50 <bitshark_> and... the last thing is /etc/crypttab or whatever... mine  (which just does encrypted swap and encrypted home) looks like this...   on startup it prompts me for my /home password, and it generates a random password for the swap
2015-04-28 20:10:03 <bitshark_> . swap    /dev/disk/by-id/ata-Samsung_SSD_840_EVO_250GB_S1DBNSADB21648Y-part3     /dev/urandom    swap,cipher=aes-cbc-essiv:sha256,size=256,discard
2015-04-28 20:10:19 <bitshark_> . chome   /dev/disk/by-id/ata-Samsung_SSD_840_EVO_250GB_S1DBNSADB21648Y-part4     none    luks,discard
2015-04-28 20:10:25 <bitshark_> --^ my etc crypttab
2015-04-28 20:10:34 <bitshark_> so that covers encrypted swap and encrypted home
2015-04-28 20:10:42 <bitshark_> and there's no chrooting and u can do it on an installed system
2015-04-28 20:10:48 <bitshark_> with any filesystem u like
2015-04-28 20:11:51 <bitshark_> and lastly, that discard option is for ssds as im sure you prob know
2015-04-28 20:11:58 <Tekki> I appreciate you trying to help, but I've already done all that and it seems to be working well already, it's just that the kernel isn't and that's what I'm trying to fix with the ISO and chrooting
2015-04-28 20:12:14 <bitshark_> hey np man
2015-04-28 20:12:43 <bitshark_> where are you seeing chrooting
2015-04-28 20:12:48 <bitshark_> can u link that to me
2015-04-28 20:14:24 <Tekki> Well I figured I'd try doing apk update and apk fix linux-grsec, but that "Unable to lock database" apk error is all I've got
2015-04-28 20:19:10 <bitshark_> oh wait
2015-04-28 20:19:11 <bitshark_> try this
2015-04-28 20:19:16 <bitshark_> if u are getting an error like that
2015-04-28 20:19:29 <bitshark_> did u try looking at strace apk <error condition input>
2015-04-28 20:19:43 <bitshark_> and see if it tells u what file it's crapping out on
2015-04-28 20:19:52 <bitshark_> just an idea
2015-04-28 20:20:09 <BitL0G1c> after a new install you need to 'setup-interfaces'
2015-04-28 20:20:17 <BitL0G1c> & start networking
2015-04-28 20:20:31 <BitL0G1c> it's not enabled by default
2015-04-28 20:23:03 <Tekki> BitL0G1c, Isn't it a part of setup-alpine?
2015-04-28 20:25:36 <Tekki> bitshark_, "/bin/ash: strace: not found" so I think I'd first need apk to work to get that as a package
2015-04-28 20:25:51 <BitL0G1c> i think so (I never run setup-alpine)
2015-04-28 20:26:22 <Tekki> I've been using that as a new user, it's been helpful.
2015-04-28 20:26:27 <BitL0G1c> but you definitely need to start networking
2015-04-28 20:36:48 <_ikke_> setup-alpine does ask thing to setup the network
2015-04-28 20:44:00 <bitshark_> ohh yeah
2015-04-28 20:44:18 <bitshark_> u know what these guys may be on to something
2015-04-28 20:44:19 <bitshark_> tekki try ping 8.8.8.8
2015-04-28 20:44:19 <bitshark_> see if u get echo replies
2015-04-28 20:44:27 <bitshark_> then try ping google.com and see if u get echo replies
2015-04-28 20:44:31 <bougyman> why abuse google?
2015-04-28 20:44:34 <bougyman> ping 4.2.2.2
2015-04-28 20:44:38 <bitshark_> well u know
2015-04-28 20:44:39 <bougyman> take it out on level3, instead.
2015-04-28 20:44:44 <bitshark_> whatever ur favorite ping is haha
2015-04-28 20:45:14 <bitshark_> cause sometimes apk will fail if (a) network and (b) wrong mirror or high latency / packet loss to mirror in like antarctica or whatever
2015-04-28 20:45:15 <bougyman> I built 4.2.2.2 in '95.
2015-04-28 20:45:22 <bitshark_> yeah?
2015-04-28 20:45:24 <bitshark_> nice
2015-04-28 20:45:26 <bougyman> so i've got a special affinity for it :)
2015-04-28 20:45:32 <bitshark_> i got a nameserver arond i still use
2015-04-28 20:45:33 <bougyman> it was GTE back then.
2015-04-28 20:45:35 <bitshark_> for like 15 years
2015-04-28 20:45:42 <bougyman> i'm sure it's not the same build now :)
2015-04-28 20:45:59 <bitshark_> i didnt build it but i saved it from 0day like 800 times
2015-04-28 20:46:04 <bitshark_> 207.172.3.8
2015-04-28 21:36:44 <bitshark_> hey guuys
2015-04-28 21:36:50 <bitshark_> how do i see files in a package with apk
2015-04-28 21:36:58 <bitshark_> im having trouble installing enlightenment from testing
2015-04-28 21:37:04 <bitshark_> something about libgif
2015-04-28 21:38:20 <bitshark_> okay -a got it
2015-04-28 23:01:17 <Mo0O> hi there
2015-04-28 23:26:11 <Mo0O> I'm trying to make an aport, which needs to download a tar.gz from github, I can download the tar.gz using `curl -O` but `abuild checksum` return an error: curl: (22) The requested URL returned error: 406 Not Acceptable
2015-04-28 23:26:24 <Mo0O> do you know how can I solve that?
2015-04-28 23:28:03 <BitL0G1c> it will probably be easier to make a proper APKBUILD - & let abuild download the source
2015-04-28 23:28:35 <BitL0G1c> clone aports & look at the APKBUILDs
2015-04-28 23:29:00 <BitL0G1c> I'll show you what a github sload looks like
2015-04-28 23:29:07 <Mo0O> BitL0G1c: that's exactly what I'm doing, an APKBUILD
2015-04-28 23:29:33 <Mo0O> 
ACTION has already clone aports

2015-04-28 23:30:24 <BitL0G1c> source="naxsi-$_ngx_naxsi_ver.tar.gz::https://github.com/nbs-system/naxsi/archive/$_ngx_naxsi_ver.tar.gz
2015-04-28 23:30:37 <Mo0O> I was using `curl -O` just to test if the file is downloadable ;)
2015-04-28 23:30:52 <BitL0G1c> just run "abuild checksum"
2015-04-28 23:31:20 <BitL0G1c> github normally stores the tar.gz without a filename - so:
2015-04-28 23:31:20 <Mo0O> 
ACTION try

2015-04-28 23:32:03 <BitL0G1c> source = "$pkgname-$pkgver.tar.gz::http://...xxxxxx"
2015-04-28 23:32:20 <BitL0G1c> note 2 x colons before HTTP
2015-04-28 23:33:17 <BitL0G1c> you want to see $pkgname-$pkgver.tar.gz in your /var/cache/distfiles
2015-04-28 23:33:32 <Mo0O> hmm doesn't work too, here's my work in progress APKBUILD: http://sprunge.us/RVDf
2015-04-28 23:33:35 <BitL0G1c> not $pkgver.tar.gz
2015-04-28 23:33:52 <Mo0O> yes, I have noted
2015-04-28 23:34:33 <Mo0O> BitL0G1c: did I miss something?
2015-04-28 23:36:25 <Mo0O> here a better version: http://sprunge.us/iUJd
2015-04-28 23:36:42 <Mo0O> but didn't fix the 406 error
2015-04-28 23:38:03 <Mo0O> here is the github download page: https://github.com/Normation/rudder-packages/tags
2015-04-28 23:38:48 <BitL0G1c> source="rudder-agent-$pkgver::https://github.com/Normation/rudder-packages/archive/$pkgver.tar.gz"
2015-04-28 23:38:53 <BitL0G1c> that works
2015-04-28 23:39:45 <Mo0O> realy o_O
2015-04-28 23:39:50 <Mo0O> not here -_-
2015-04-28 23:40:27 <BitL0G1c> for problems like this just copy the link location on github & compare it to the curl error message - you missed the - in rudder-packages
2015-04-28 23:40:48 <BitL0G1c> & the p
2015-04-28 23:41:28 <Mo0O> oh yes, you're right!
2015-04-28 23:41:40 <Mo0O> thanks so much
2015-04-28 23:41:48 <Mo0O> sorry btw
2015-04-28 23:42:38 <Authentic> What goes into making a package for Alpine? (Warning: very new)
2015-04-28 23:43:03 <Authentic> http://pkgs.alpinelinux.org/packages has mysql and postgresql, but lacks Mongodb
2015-04-28 23:45:19 <Mo0O> Authentic: just write it ;(
2015-04-28 23:45:22 <Mo0O> ;)
2015-04-28 23:45:24 <Mo0O> *
2015-04-28 23:46:55 <Authentic> I will need to look up how to do that then!
2015-04-28 23:47:12 <systmkor> Authentic: check out https://projects.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD?h=packages/mongodb
2015-04-28 23:47:22 <Mo0O> Authentic: http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2015-04-28 23:47:35 <systmkor> Authentic: however my guess is that you will run into a variety of roadblocks
2015-04-28 23:47:59 <Authentic> Mo0o: Thanks for the link!
2015-04-28 23:48:08 <Authentic> systmkor: oh i'm sure I will, nothing is ever easy..
2015-04-28 23:48:37 <Mo0O> np
2015-04-29 00:46:40 <packeteer> oh hai
2015-04-29 00:53:58 <packeteer> anybody got a Dockerfile for nginx (using alpine base) ?
2015-04-29 00:54:26 <bitshark> hey
2015-04-29 00:54:32 <bitshark> oh geez Docker
2015-04-29 00:54:48 <bitshark> i tried it and i didnt like it at alll
2015-04-29 00:55:14 <bitshark> what are you running packeteer... uwsgi with python or sumt else
2015-04-29 00:56:22 <bitshark> usually ppl using docker with nginx are serving apps not html
2015-04-29 01:11:08 <packeteer> just trying to build a base nginx image first off
2015-04-29 01:11:34 <packeteer> <still learning>
2015-04-29 01:17:50 <bitshark_-> oh hey there u r
2015-04-29 01:18:11 <bitshark_-> oh well i have strong opinions that are a liabiity
2015-04-29 01:18:25 <bitshark_-> so i cant really help , haha
2015-04-29 01:18:32 <bitshark_-> im of the opinion docker is a fad
2015-04-29 01:18:54 <bitshark_-> for a number of reasons... unless they release a totally revamped version of it
2015-04-29 01:21:34 <packeteer> yeah, i'm not a big fan, but it is an emerging tech that we're looking at, so needs must an all that
2015-04-29 01:22:17 <packeteer> it's not eneterprise ready tech / still beta / immature
2015-04-29 01:22:18 <packeteer> etc
2015-04-29 02:47:40 <bitshark___> so ive had some issuez with the alpine package manageement
2015-04-29 02:47:52 <bitshark___> i dont think i can use it for desktop stuff quite yet...
2015-04-29 02:47:59 <bitshark___> mainly a problem with the 'testing' repo
2015-04-29 02:49:32 <bitshark> well i think there are a few changes that need to go into apk
2015-04-29 02:49:46 <bitshark> first is it should display versions
2015-04-29 02:50:04 <bitshark> like apt info   has got to list when there are two versions installed
2015-04-29 02:50:17 <bitshark> it should prompt if u try to install two versions of same package etc
2015-04-29 02:50:37 <bitshark> cause of sym links getting clobbered
2015-04-29 02:50:43 <bitshark> in /usr/lib
2015-04-29 02:51:33 <bitshark> so apk info  should show like "<package_name>]t<package_version>\n"
2015-04-29 02:52:00 <bitshark> i'll add the code unless teres objections to that idea
2015-04-29 05:03:19 <soo> hi
2015-04-29 05:03:44 <soo> I've made a big mistake, I've remove /usr/*
2015-04-29 05:04:04 <soo> do you think it's possible to repair it?
2015-04-29 05:04:34 <soo> I've chrooted to the system
2015-04-29 05:04:52 <soo> the history file is explicit:
2015-04-29 05:05:03 <soo> : 1430281072:0;sudo rm -rf /usr/*
2015-04-29 05:05:06 <soo> -_-
2015-04-29 05:37:02 <packeteer> u can try to extract a base install image over the top
2015-04-29 05:52:15 <soo> yep, but there's no /usr on the iso
2015-04-29 05:56:00 <crow> soo what kind of install you did? maybe you can install again in some VM (virtualbox or vmware workstation) and copy /usr to your other one
2015-04-29 05:56:25 <crow> soo why did you did such cleanup :) ?
2015-04-29 05:57:19 <soo> crow: it's a fail, I was want to delete another folder
2015-04-29 05:57:39 <soo> but my fingers hit the wrong keys -_-
2015-04-29 05:57:45 <soo> don't know why
2015-04-29 05:57:50 <soo> stupid me
2015-04-29 05:57:55 <soo> ^^
2015-04-29 05:58:10 <soo> but it's a good idea
2015-04-29 05:58:40 <crow> dont do all as root :)
2015-04-29 05:59:56 <soo> yes, ;)
2015-04-29 06:00:05 <soo> hope this will give me a lesson
2015-04-29 06:13:54 <bitshark> hey
2015-04-29 06:14:03 <bitshark> ne one here
2015-04-29 06:14:16 <sooo> hi bitshark
2015-04-29 06:15:37 <bitshark> hey
2015-04-29 06:15:49 <bitshark> just was getting annoyed at apk haha
2015-04-29 06:16:07 <bitshark> it lets u install multiple versions of the same library
2015-04-29 06:16:17 <bitshark> but it calls them by the same name
2015-04-29 06:16:27 <sooo> right now I just need one :D
2015-04-29 06:16:38 <bitshark> ie.  apk add giflib
2015-04-29 06:17:05 <bitshark> so if i do apk add giflib   , it's possible i end up with two giflibs,  each a different version
2015-04-29 06:17:18 <bitshark> and the only way to tell is to query each one
2015-04-29 06:17:36 <sooo> weired
2015-04-29 06:18:29 <bitshark> anyway i patched apk so that when you type "apk info" it shows the package name in one column and version in the other
2015-04-29 06:18:38 <bitshark> so that helps
2015-04-29 06:18:46 <rnalrd> hi
2015-04-29 06:18:54 <sooo> yes, good idea bitshark
2015-04-29 06:18:58 <rnalrd> what do you have in repositories?
2015-04-29 06:18:59 <sooo> hi rnalrd
2015-04-29 06:19:50 <bitshark> kk
2015-04-29 06:19:57 <bitshark> hey rnalrd
2015-04-29 06:20:05 <bitshark> i had only one repository
2015-04-29 06:20:13 <bitshark> but if you switch between them at all
2015-04-29 06:20:24 <bitshark> you can get into an unstable state
2015-04-29 06:20:28 <bitshark> really easily
2015-04-29 06:20:38 <rnalrd> no surprise
2015-04-29 06:20:47 <bitshark> i uncommented testing, then tried to do apk add enlightenment
2015-04-29 06:21:09 <bitshark> it was silent but it duplicated a ton of libraries
2015-04-29 06:21:15 <bitshark> clobbering symlinks etc
2015-04-29 06:21:27 <rnalrd> try changing distro in /etc/apt/repositorie.d in debian based distro, and the result will be no different
2015-04-29 06:21:41 <bitshark> well that's true
2015-04-29 06:22:10 <bitshark> but with debian at least i can see when i install two copies of the same thing
2015-04-29 06:22:12 <rnalrd> when you change version, you should "apk update && apk upgrade -a" to make sure you switch entirely on new version
2015-04-29 06:22:20 <bitshark> okay
2015-04-29 06:22:25 <bitshark> i'll do that
2015-04-29 06:22:28 <bitshark> hang on
2015-04-29 06:22:39 <rnalrd> flag "-a" ensures that you don't have packages from other versions
2015-04-29 06:22:42 <bitshark> im actually in unstable state right now
2015-04-29 06:22:48 <bitshark> everything i do has to have --f
2015-04-29 06:23:07 <rnalrd> also, installing busybox-static and apk-static helps in such cases
2015-04-29 06:23:41 <sooo> I'm actualy without /usr now :D
2015-04-29 06:23:48 <bitshark> oh wait
2015-04-29 06:24:01 <sooo> what?
2015-04-29 06:24:01 <bitshark> it works it just warns whenever i do a command that enlightenment is missing from "world"
2015-04-29 06:24:10 <bitshark> like apk add openssh-client
2015-04-29 06:24:58 <bitshark> adds the package silently, but prints something about "Error: unsatisfied constraints : Enlightenment (missing)  required by world(Enlightenment)
2015-04-29 06:25:03 <bitshark> that's a confusing error message haha
2015-04-29 06:25:05 <bitshark> world
2015-04-29 06:25:12 <bitshark> bitshark meetz world
2015-04-29 06:25:14 <rnalrd> sounds like enlightment is not avail in the repo selected
2015-04-29 06:25:23 <rnalrd> but it was previously installed
2015-04-29 06:25:27 <bitshark> ohhh
2015-04-29 06:25:29 <rnalrd> so it complains
2015-04-29 06:25:33 <bitshark> yeah i gave up on enlightenment two hours ago
2015-04-29 06:25:38 <bitshark> it as in testing
2015-04-29 06:25:47 <bitshark> i gave up and switched back to main / regular
2015-04-29 06:25:57 <bitshark> by changing the list and doing apk update
2015-04-29 06:25:59 <rnalrd> apk del enlightment should fix it
2015-04-29 06:26:03 <sooo> you could use the @testing syntax
2015-04-29 06:26:23 <rnalrd> yeah, repo pinning is a useful feature if you need just few packages from another repo
2015-04-29 06:26:27 <bitshark> well apk del enlight i tried
2015-04-29 06:26:28 <bitshark> no nice
2015-04-29 06:26:31 <bitshark> no dice*
2015-04-29 06:26:37 <bitshark> its okay this is a test system
2015-04-29 06:26:39 <rnalrd> apk del enlightment --force
2015-04-29 06:26:41 <rnalrd> ?
2015-04-29 06:26:41 <bitshark> i have snapshots in vmware
2015-04-29 06:26:43 <sooo> this way you install only enlightment from testing
2015-04-29 06:26:50 <bitshark> just trying to get familiar
2015-04-29 06:27:00 <sooo> will work without --force normaly
2015-04-29 06:27:55 <bitshark_> k just tried apk update ; apk upgrade -a
2015-04-29 06:28:02 <bitshark_> the latter onnly works with -f
2015-04-29 06:28:12 <bitshark_> now its downgrading most of the system
2015-04-29 06:28:17 <rnalrd> k
2015-04-29 06:28:18 <bitshark_> i think thats the major issue
2015-04-29 06:28:26 <bitshark_> cuz it just downgraded like 25+ packages
2015-04-29 06:28:36 <bitshark_> noope up to 50+ now
2015-04-29 06:28:38 <bitshark_> wow okay
2015-04-29 06:28:42 <bitshark_> that had to be the problem
2015-04-29 06:28:55 <bitshark_> thanks for the tip, im definitely closer
2015-04-29 06:28:59 <bitshark_> btw
2015-04-29 06:29:10 <bitshark_> do u run alpine at home / on laptop / on desktop
2015-04-29 06:29:15 <bitshark_> and if not what do ya run
2015-04-29 06:29:17 <sooo> yes
2015-04-29 06:29:29 <bitshark_> okay rad
2015-04-29 06:29:31 <sooo> laptop
2015-04-29 06:29:42 <bitshark_> why did u pick alpine sooo
2015-04-29 06:30:07 <packeteer> it's le small
2015-04-29 06:30:12 <sooo> security for server, using on laptop to be familliar with
2015-04-29 06:32:53 <bitshark_> yeah
2015-04-29 06:33:18 <bitshark_> me too basically
2015-04-29 06:33:37 <bitshark_> also i wanted openrc, grsec, and support for video/sound etc
2015-04-29 06:34:24 <bitshark_> i almost moved to freeBSD after this systemd nerd-war  , but i learned that even the latest freebsd release
2015-04-29 06:34:30 <iwearsable>  < bitshark_> thanks for the tip, im definitely closer
2015-04-29 06:34:31 <bitshark_> doesn't support intel i915 chipset for video
2015-04-29 06:34:34 <iwearsable> What was the tip?
2015-04-29 06:34:42 <bitshark_> that's a common laptop video card from like 3 years ago
2015-04-29 06:34:43 <bitshark_> hah
2015-04-29 06:34:46 <iwearsable> I was watching intently then my internet went out
2015-04-29 06:34:47 <bitshark_> so i couldt use freebsd
2015-04-29 06:35:30 <iwearsable> I have problems with apk after changing repos (I had to change repo one of them was out of date)
2015-04-29 06:35:44 <iwearsable> No matter what I'm stuck in a loop of dependency hell now
2015-04-29 06:36:02 <bitshark_> oh i meant suggestion for "apk update && apk upgrade -a "
2015-04-29 06:36:24 <bitshark_> seems to be working with an -f tacked onthe end
2015-04-29 06:36:38 <bitshark_> well actually iwearsable
2015-04-29 06:36:57 <bitshark_> i'm preparing right now to set upa  bare-bones linux system with nix package manager
2015-04-29 06:37:13 <bitshark_> it is hybrid binary/source package manager that makes dependency hell impossible
2015-04-29 06:37:44 <iwearsable> ah, nevermind it sounded like similar issue
2015-04-29 06:37:50 <bitshark_> i'm still torn between trying it on my customized linux vs on alpine minimal install
2015-04-29 06:38:00 <bitshark_> but i'll let u know guys know how it goes
2015-04-29 06:38:26 <bitshark_> the nixos guys are annoying tho
2015-04-29 06:38:37 <bitshark_> i'm like "i want to use your tool not your OS"
2015-04-29 06:38:45 <bitshark_> and they get mad at me
2015-04-29 06:38:52 <bitshark_> cause i dont like their distro
2015-04-29 06:38:53 <bitshark_> heh
2015-04-29 06:39:00 <bitshark_> i like alpine 100x better
2015-04-29 06:39:35 <bitshark_> sprawl and distro do not belong inthe same sentence
2015-04-29 06:40:15 <bitshark_> alpine is fast as hell
2015-04-29 06:40:29 <iwearsable> lol, I like Alpine too except when I get stuck in this loop where installing alpine-desktop conflicts with some library
2015-04-29 06:40:32 <bitshark_> just needs to be a little more robust
2015-04-29 06:40:40 <bitshark_> alpine needs some work on pkg mgmt
2015-04-29 06:40:53 <bitshark_> almost everyting else about it is elite / solid
2015-04-29 06:40:56 <iwearsable> It's like I need to tell it to just download and update ALL PACKAGES again and it would be fine
2015-04-29 06:40:58 <bitshark_> except two major things
2015-04-29 06:41:04 <bitshark_> (1) installer  / setup
2015-04-29 06:41:10 <bitshark_> and lack of live cd for new users
2015-04-29 06:41:31 <bitshark_> and (2) alpine needs work on apk to make it harder to enter dependency hell
2015-04-29 06:41:55 <bitshark_> actually if nix works out well
2015-04-29 06:42:22 <bitshark_> i'll issue a pull request up to alpine as an alternative / adjunct to apk
2015-04-29 06:42:31 <bitshark_> to alpine's source control
2015-04-29 06:43:36 <iwearsable> I don't know much aboit nix but I see it get mentioned here and there. First time I saw Alpine was as an option at my VPS host it was an option
2015-04-29 06:44:19 <iwearsable> I need to reboot bbl
2015-04-29 06:44:40 <bitshark__> k
2015-04-29 08:50:17 <sillysausage> hmm, that's weird thunderbird isn't in any of the repositories
2015-04-29 08:50:40 <sillysausage> yet faenza-icon-theme-thunderbird-1.3.1-r4 a theme is
2015-04-29 08:51:21 <clandmeter> probably subpackaged?
2015-04-29 08:51:23 <sillysausage> doesn't seem to be called icedove either
2015-04-29 08:52:58 <clandmeter> we do not ship tunderbird
2015-04-29 08:53:29 <sillysausage> fair enough, maybe i should look into packaging that
2015-04-29 08:53:52 <sillysausage> along with electrum, you have bitcoin-qt but that would require me to have a copy of the blockchain :P
2015-04-29 08:54:27 <sillysausage> i'm actually quite surprised to see truecrypt was packaged :P
2015-04-29 08:54:53 <sillysausage> as "truecrypt"
2015-04-29 08:55:14 <sillysausage> guess i can use clawsmail instead
2015-04-29 09:12:33 <sillysausage> http://www.thewildbeast.co.uk/claws-mail/bugzilla/show_bug.cgi?id=2658
2015-04-29 09:12:52 <sillysausage> aww well seems no clients support carddav, I find it hard to believe it's not a more important feature
2015-04-29 09:13:10 <sillysausage> only thing on linux i've found that does it reliably is akonadi
2015-04-29 09:14:18 <sillysausage> https://bugzilla.mozilla.org/show_bug.cgi?id=546932 they never implemented it either
2015-04-29 09:15:51 <sillysausage> maybe i could implement one in python using pyCardDAV as a plugin
2015-04-29 09:16:36 <sillysausage> should be able to access it with some sort of hacky clawsmail plugin using claws-mail-plugins-python
2015-04-29 09:32:00 <sillysausage> i guess i could do this http://wiki.alpinelinux.org/wiki/LXC#Debian_template and then use icedove
2015-04-29 09:32:08 <sillysausage> does the debian template work with jessie?
2015-04-29 09:35:02 <sillysausage> https://wiki.debian.org/LXC#Debian_8_.22Jessie.22.2Ftesting
2015-04-29 10:50:16 <jomat> usually apk works quite fine, but if it doesn't, it's not too chatty about its reasons :-) https://0.jmt.gr/?689ddad4c787c31d#pLhA+S9TFtdNmVNNnT3J/TDb13aCfD7j4oSEv5u31jA=
2015-04-29 10:54:45 <bitshark_> i had trouble with it earlier
2015-04-29 10:55:02 <bitshark_> it will install identical packages (Same names) and doesn't flag em
2015-04-29 10:55:27 <bitshark_> it happened when i switched to testing and back
2015-04-29 10:55:41 <bitshark_> i actually made a patch for it
2015-04-29 10:55:51 <bitshark_> i gotta submit that to alpine
2015-04-29 10:56:00 <jomat> there's my error 404: https://0.jmt.gr/?f8a21348a46ffa42#9hmucI3rYlkfx911f2PENIFrQHcZrTxqcbfOmwDzTCo=
2015-04-29 10:56:24 <bitshark_> right now i'm actually prepping a scratch build to test with nix package manager
2015-04-29 10:56:33 <bitshark_> a custom lfs and nix... nothing else
2015-04-29 10:56:42 <bitshark_> it's either gonna be great or a disaster haha
2015-04-29 10:57:06 <jomat> hehe :-)
2015-04-29 10:58:26 <jomat> I think I should file a bug... and clean up my aports tree and submit more patches :-)
2015-04-29 10:59:54 <fabled> jomat, i think that is fixed in edge
2015-04-29 11:00:03 <fabled> what is apk --version ?
2015-04-29 11:00:11 <jomat> apk-tools 2.5.0_rc1, compiled for x86_64.
2015-04-29 11:00:13 <fabled> 2.6.0_rc4 should be fixed
2015-04-29 11:00:24 <jomat> ok, thx :-)
2015-04-29 11:01:09 <fabled> re: nix, it solves the wrong problem for me, but i understand some might want to use it
2015-04-29 11:01:18 <fabled> i don't want multiple side-by-side installed versions
2015-04-29 11:01:28 <fabled> i want single minimal system where most things are shared
2015-04-29 11:01:45 <fabled> seems not so many people want it anymore because it means a bit more work
2015-04-29 11:01:48 <fabled> and hard disks are cheap
2015-04-29 11:03:39 <bitshark_> fabled did u try nix for ahwile?
2015-04-29 11:03:49 <fabled> no, but i read the design philosophy
2015-04-29 11:03:53 <bitshark_> oh okay
2015-04-29 11:04:08 <bitshark_> well i'm here working with alpine for a few reasons
2015-04-29 11:04:13 <bitshark_> one is the one you mention
2015-04-29 11:04:19 <bitshark_> there's too much bloat in OSs
2015-04-29 11:04:25 <bitshark_> so i agree on that part
2015-04-29 11:04:41 <bitshark_> and that's why i like alpine and ive got it in a cloud deployment
2015-04-29 11:05:02 <bitshark_> but for desktop / development, im sick of the 'bloat' that comes from even regular packagem manager like apt-get
2015-04-29 11:05:04 <fabled> but it's tricky problem docker and other similar things try to solve it
2015-04-29 11:05:14 <bitshark_> and dependency hell
2015-04-29 11:05:15 <fabled> usually they do it by installing multiple versions or multiple distributions
2015-04-29 11:05:19 <bitshark_> good lord
2015-04-29 11:05:24 <fabled> the install is easier
2015-04-29 11:05:29 <bitshark_> so im trying nix with a clfs build i made
2015-04-29 11:05:30 <fabled> but maintenance becomes more problematic
2015-04-29 11:05:40 <fabled> as minimum you end up depending on more places to upgrade their things
2015-04-29 11:05:50 <bitshark_> i'm worried about the exact issue you mention... i worry nix will creep into the hard drive
2015-04-29 11:05:55 <bitshark_> eating up my expensive ssd
2015-04-29 11:05:56 <fabled> and to roll out system wide openssl upgrade would be impossible in simple way
2015-04-29 11:06:03 <fabled> you'd need all different image providers to update it
2015-04-29 11:06:11 <bitshark_> that said... (1) i hate docker so much
2015-04-29 11:06:13 <bitshark_> so that's out
2015-04-29 11:06:15 <bitshark_> haha
2015-04-29 11:06:29 <fabled> i understand how it serves certain purposes
2015-04-29 11:06:36 <fabled> but i think people use it for wrong things
2015-04-29 11:06:40 <bitshark_> (2) nix does deal with some of the deduplication so that could be useful
2015-04-29 11:06:42 <fabled> just because "hey it's fast to install X" on it
2015-04-29 11:06:51 <fabled> does not mean it's easy also to run and maintain it in long-term
2015-04-29 11:06:54 <bitshark_> yeah
2015-04-29 11:06:57 <bitshark_> exactly how i feel
2015-04-29 11:07:13 <fabled> considering the hype
2015-04-29 11:07:14 <bitshark_> most ppl that use it dont even bother to read the haskall
2015-04-29 11:07:20 <bitshark_> for nix
2015-04-29 11:07:23 <fabled> it there will be gazillion images in docker soon
2015-04-29 11:07:30 <fabled> and only few will get regular updates
2015-04-29 11:07:32 <bitshark_> and for docker it's essentially like a trendy version of vagrant
2015-04-29 11:07:36 <fabled> and when next openssl critical bug gets out
2015-04-29 11:07:39 <bitshark_> dont even get me started on docker haha
2015-04-29 11:07:42 <fabled> docker will blow up ;)
2015-04-29 11:07:44 <bitshark_> lol
2015-04-29 11:07:47 <bitshark_> i live in SF
2015-04-29 11:07:59 <bitshark_> i have to deal with autistic kids that worship the alter of docker
2015-04-29 11:08:26 <bitshark_> they are everywhere
2015-04-29 11:08:33 <bitshark_> ne wayz yeah to me... i actually tried docker for awhile
2015-04-29 11:08:49 <bitshark_> what i found funny tho was their 'philosophy' and manifesto & blogs etc
2015-04-29 11:08:58 <bitshark_> completely contradicted how 99% of people were using it
2015-04-29 11:09:01 <bitshark_> ie.
2015-04-29 11:09:37 <bitshark_> according to the 'Docker' principles, you are suppoed to have one app per container, and make them talk to one another (no one specifies how... IPV4 with double NAT? )
2015-04-29 11:09:40 <bitshark_> anyway
2015-04-29 11:09:55 <bitshark_> in reality ppl now have put entire OS's in docker containers running multiple services
2015-04-29 11:10:09 <bitshark_> ie. an ubuntu docker image with nginx flask and mysql
2015-04-29 11:10:36 <bitshark_> that's against their whole premise but they let go becuse they want adoption at any cost
2015-04-29 11:10:51 <bitshark_> i suppose the moral is... if u want that sort of use case
2015-04-29 11:11:00 <bitshark_> u can do it 100x faster with LXC or freebsd jails
2015-04-29 11:11:13 <bitshark_> or even with lightweight hypervisor like openvz
2015-04-29 11:12:31 <bitshark_> there's no need for docker aufs / overlayfs since deduplication is a major performance hog... also docker runs as root , so forgot running it for anything dealing with money
2015-04-29 11:13:00 <bitshark_> like payment gatewayz or billing web servicez
2015-04-29 11:18:03 <ncopa> sorry ppl. i think i managed to blow away half edge repo
2015-04-29 11:18:28 <bitshark_> haha
2015-04-29 11:27:54 <_ikke_> auch :P
2015-04-29 11:32:16 <_ikke_> ncopa: Did you see the redmine docker container I created?
2015-04-29 11:32:58 <ncopa> i saw it
2015-04-29 11:33:04 <ncopa> i havent had time to test it though
2015-04-29 11:33:21 <_ikke_> ok, no problem
2015-04-29 12:22:12 <mhoungbo> hi all
2015-04-29 12:23:27 <mhoungbo> Question : does a raspberry pi 2 version of alpine linux  available ?
2015-04-29 12:23:55 <_ikke_> I heard some people here talking about it the other day
2015-04-29 12:25:24 <mhoungbo> Thanks _ikke_
2015-04-29 12:28:49 <clandmeter> mhoungbo: should be
2015-04-29 12:29:09 <_ikke_> Can't find any references though
2015-04-29 12:29:23 <clandmeter> http://dev.alpinelinux.org/~tteras/
2015-04-29 12:29:36 <clandmeter> i think the uboot images should be compatbile
2015-04-29 12:29:44 <clandmeter> fabled: am i correct?
2015-04-29 12:29:58 <_ikke_> yeah, I found a mailing list message from him talking about it
2015-04-29 12:30:14 <fabled> yes. take the latest alpine-rpi imagae from there
2015-04-29 12:32:39 <ahills> _ikke_: link to your redmine docker?
2015-04-29 12:36:21 <_ikke_> https://github.com/ikke/docker-alpine-redmine
2015-04-29 12:38:38 <ncopa> re redmine
2015-04-29 12:38:47 <ncopa> i pushed ruby 2.2.2 the other day
2015-04-29 12:39:04 <ncopa> would be nice if someone could test that redmine docker to verify that it still works
2015-04-29 12:39:20 <_ikke_> I can recreate my container
2015-04-29 12:39:24 <_ikke_> I have running
2015-04-29 12:40:25 <ahills> _ikke_: quick nitpick, your build_container error messages both refer to different filenames than the script actually wants
2015-04-29 12:42:01 <_ikke_> ahills: ah ok
2015-04-29 12:42:52 <_ikke_> The first one seems correct? ./buildsettings
2015-04-29 12:42:59 <_ikke_> ie, the one in the current directory
2015-04-29 12:43:06 <ahills> no, it's build_settings
2015-04-29 12:43:09 <ahills> both times
2015-04-29 12:43:15 <_ikke_> oh, yeah, thanks
2015-04-29 12:43:31 <mhoungbo> clandmeter, nothing there
2015-04-29 12:44:32 <_ikke_> Someone able to test it?
2015-04-29 12:45:53 <_ikke_> It's at least running right now
2015-04-29 12:59:10 <ahills> I'll be able to test it later today
2015-04-29 15:11:55 <coredumb> http://pastebin.com/4bBbCznR got edge install with couple of packages in testing
2015-04-29 15:12:05 <coredumb> how should i expect to fix that ?
2015-04-29 15:12:59 <coredumb> after icu-libs it's libvpx complaining
2015-04-29 15:13:45 <clandmeter> coredumb: using edge?
2015-04-29 15:14:04 <coredumb> clandmeter: yes
2015-04-29 15:14:08 <clandmeter> which repo?
2015-04-29 15:14:26 <coredumb> http://mirror1.hs-esslingen.de/pub/Mirrors/alpine/edge/main
2015-04-29 15:14:29 <coredumb> http://dl-2.alpinelinux.org/alpine/edge/main
2015-04-29 15:14:38 <coredumb> should i remove one ?
2015-04-29 15:14:55 <clandmeter> use nl.a.o
2015-04-29 15:15:02 <clandmeter> and see if that changes
2015-04-29 15:15:27 <clandmeter> there is a lot going on in edge currently, so dont expect everything to work.
2015-04-29 15:18:10 <coredumb> clandmeter: ok let me see
2015-04-29 15:18:32 <coredumb> btw you still need mirrors or you're pretty much covered ?
2015-04-29 15:18:50 <coredumb> ok same result on nl.a.o
2015-04-29 15:22:06 <clandmeter> mirrors are always welcome.
2015-04-29 15:22:17 <clandmeter> but its not like we dont have any :)
2015-04-29 15:23:49 <coredumb> :)
2015-04-29 15:25:13 <clandmeter> coredumb: i dont get the error you are seeing
2015-04-29 15:25:33 <clandmeter> oh now i see
2015-04-29 15:26:09 <coredumb> cladmeter fwiw i have libreoffice packages from testing
2015-04-29 15:27:21 <darkfader> coredumb: btw i wondered if i should just pay you to wrap up a rudder agent package for alpine
2015-04-29 15:27:54 <darkfader> cfengine support was already added a year back or so
2015-04-29 15:28:13 <clandmeter> coredumb: could be its mixing repo's
2015-04-29 15:28:24 <clandmeter> icu is ahead in edge
2015-04-29 15:28:30 <coredumb> darkfader: paying me is always welcome
2015-04-29 15:28:37 <coredumb> and you know it
2015-04-29 15:28:39 <coredumb> :D
2015-04-29 15:28:41 <darkfader> would you just take the money or also work? :)
2015-04-29 15:28:59 <clandmeter> how much would i still get?
2015-04-29 15:29:16 <coredumb> clandmeter: well I only have nl.a.o
2015-04-29 15:29:26 <coredumb> clandmeter: two beers ?
2015-04-29 15:29:31 <darkfader> hehe
2015-04-29 15:29:40 <clandmeter> that doesnt sound like money
2015-04-29 15:29:42 <darkfader> bounty feature in redmine would be so greay
2015-04-29 15:29:44 <darkfader> *t
2015-04-29 15:29:52 <clandmeter> but fuck it, beer taste much better :D
2015-04-29 15:32:04 <clandmeter> oh did i just use the f* word. its time to go home...
2015-04-29 15:36:34 <coredumb> darkfader: the major PITA gonna be to relocate /opt somewhere else
2015-04-29 15:44:31 <darkfader> i'd rather break the aport rules for it
2015-04-29 15:44:39 <darkfader> it's a vendor package somewhat
2015-04-29 15:44:55 <darkfader> but yeah, same thing on freebsd
2015-04-29 15:45:37 <darkfader> upstreaming in alpine or freebsd would mean to not use /opt i guess ...
2015-04-29 15:49:10 <coredumb> darkfader: dunno if we can do it actually don't have enough abuild knowledge yet
2015-04-29 15:50:01 <coredumb> you can still pay me to think about it btw ^^
2015-04-29 15:50:36 <darkfader> i'll think about that
2015-04-29 15:51:11 <coredumb> i've already partly bribed my fellow Mo0O to check the requirements :P
2015-04-29 15:55:12 <fabled> we have few packages that use /opt, like rasperrypi
2015-04-29 15:55:22 <fabled> you can override the aports check in options
2015-04-29 15:55:36 <fabled> but it's last resort only
2015-04-29 15:56:37 <coredumb> fabled: ok i'll check that
2015-04-29 15:56:54 <coredumb> yeah for rudder-agent it would be last resort i guess
2015-04-29 15:57:33 <coredumb> I'll check that in details when I have some time
2015-04-29 16:46:44 <dalias> what package on alpine provides ffmpeg/libavcodec libs?
2015-04-29 16:46:47 <dalias> for devel
2015-04-29 16:50:42 <dalias> and how do i unpin packages that are pinned to a tag?
2015-04-29 18:33:59 <fabled> dalias, ffmpeg-dev
2015-04-29 18:34:04 <fabled> and apk add pkgname
2015-04-29 18:34:10 <fabled> resets any pinnings it had
2015-04-29 18:34:20 <fabled> you might need to do after that additional "apk upgrade -a" to downgrade the package
2015-04-29 18:34:39 <dalias> i have something broken...
2015-04-29 18:34:40 <dalias> http://sprunge.us/iMXV
2015-04-29 18:36:34 <fabled> packages existing only in @stable repository, but it's not enabled for that dependency
2015-04-29 18:37:06 <fabled> hope you have repository enabled without any @tags
2015-04-29 18:40:19 <dalias> why would they be only in stable?
2015-04-29 18:40:46 <fabled> what you have in repositories?
2015-04-29 18:41:03 <dalias> @stable http://nl.alpinelinux.org/alpine/v3.0/main
2015-04-29 18:41:03 <dalias> @testing http://nl.alpinelinux.org/alpine/edge/testing
2015-04-29 18:41:03 <dalias> http://nl.alpinelinux.org/alpine/edge/main
2015-04-29 18:41:32 <fabled> mmm
2015-04-29 18:41:44 <fabled> maybe it's bug, try moving edge/main repository to be the first one
2015-04-29 18:42:59 <dalias> ok
2015-04-29 18:43:13 <dalias> same thing still happens
2015-04-29 18:43:21 <fabled> hmmh
2015-04-29 18:43:38 <dalias> something must be pinned to @stable causing it
2015-04-29 18:43:47 <dalias> but i don't see how to determine or fix that :/
2015-04-29 18:43:52 <fabled> what does "apk policy ffmpeg" say?
2015-04-29 18:44:36 <dalias> http://sprunge.us/fJNA
2015-04-29 18:46:21 <fabled> perhaps some other package is holding librtmp or x264 back then
2015-04-29 18:46:29 <fabled> what does "apk upgrade -a --latest" say?
2015-04-29 18:47:02 <dalias> upgrade -a is probably going to take a while... :-p
2015-04-29 18:47:33 <dalias> and i have some packages pinned that need to be
2015-04-29 18:47:44 <dalias> Xorg intel driver was broken shortly after 3.0 and never fixed :/
2015-04-29 18:47:47 <fabled> that's probably conflict somewhere
2015-04-29 18:48:03 <fabled> one of the pinnings likely hold back x264 / ffmpeg / librtmp
2015-04-29 18:48:50 <fabled> pinning was designed to allow selective upgrades; not selecting holdbacks. that's why the errors are misleading
2015-04-29 18:49:14 <fabled> holding back should work just fine. the errors just become like what you see
2015-04-29 18:49:47 <fabled> you might get the error with "apk upgrade --latest --simulate"
2015-04-29 18:49:52 <fabled> that looks better
2015-04-29 18:50:54 <bitshark__> oh hey guys
2015-04-29 18:51:03 <bitshark__> ncopa said he broke the repos
2015-04-29 18:51:05 <bitshark__> testing i think
2015-04-29 18:51:20 <bitshark__> so if somethings wierd,that's rpobablywhy
2015-04-29 18:51:56 <bitshark__> <bitshark_> there's no need for docker aufs / overlayfs since deduplication is a major performance hog... also docker runs as root , so forgot running it for anything dealing with money
2015-04-29 18:51:56 <bitshark__> <bitshark_> like payment gatewayz or billing web servicez
2015-04-29 18:51:56 <bitshark__> <ncopa> sorry ppl. i think i managed to blow away half edge repo
2015-04-29 18:51:56 <bitshark__> <bitshark_> haha
2015-04-29 18:51:56 <bitshark__> <bitshark_> that sucks
2015-04-29 18:51:57 <bitshark__> <bitshark_> there must be backups somewhwere tho
2015-04-29 18:52:08 <bitshark__> "edge" that's it
2015-04-29 18:53:27 <_ikke_> bitshark__: he broke edge
2015-04-29 18:54:13 <nangel> ncopa: you just need the apks?
2015-04-29 18:54:54 <nangel> dl-3 hasn't replicated yet. edge uses 18.8G on dl-3
2015-04-29 18:57:38 <dalias> bleh a forced add -u started removing all packages....
2015-04-29 18:57:47 <dalias> killed it quickly
2015-04-29 18:58:08 <dalias> and managed to save the list of packages i removed to pass back to add
2015-04-29 19:03:26 <fabled> dalias, "touch /etc/apk/interactive" make apk confirm the action
2015-04-29 19:04:05 <dalias> pretty ugly that _add_ can ever remove packages :-p
2015-04-29 19:04:22 <dalias> anyway now after removing firefox i finally got apk update -a working without conflicts
2015-04-29 19:04:27 <dalias> so maybe it can fix the system
2015-04-29 19:04:40 <dalias> it's running terribly slow
2015-04-29 19:05:44 <bitshark_> oh i made a patch for apk
2015-04-29 19:05:48 <bitshark_> where do i send pull request
2015-04-29 19:06:12 <bitshark_> it's just a usability thing to show package version by default n nice columns
2015-04-29 19:07:53 <fabled> email to alpine-devel@lists.alpinelinux.org
2015-04-29 19:09:09 <nangel> http://wiki.alpinelinux.org/wiki/Creating_patches
2015-04-29 19:09:28 <nangel> wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2015-04-29 19:09:38 <Risk64> WHy doesn't it honor my fstab settings (especially "noatime")? http://abload.de/img/mount5rusl.png
2015-04-29 19:13:12 <fabled> Risk64, it's hard disk install: mounting root happens before it can read fstab. there's kernel boot parameter for that. though it might able to consider doing remount after the initial mount with right flags.
2015-04-29 19:13:43 <systmkor> fabled: sup
2015-04-29 19:14:24 <Risk64> fabled: ok, thx. Will try remount.
2015-04-29 19:16:34 <Risk64> Works for 'noatime' but no sign of 'data=writeback'. It doesn't show up.
2015-04-29 19:21:06 <Risk64> dmesg says its mounted in writeback mode so should be ok (even though its not listed in 'mount' output).
2015-04-29 19:25:21 <bitshark_> ok thanks
2015-04-29 19:25:31 <bitshark_> appeciate that
2015-04-29 21:02:15 <iwearsable> I've been fighting my system for a few days now, making things worse with APK. Seems to be centered around "unsatifiable contrants" involving libvpx. If I go and force anything, trying to fix it, I remove alpine-desktop
2015-04-29 21:02:47 <iwearsable> Also now I can't run Firefox no matter what, complains about missing libicuil8n.so.55
2015-04-29 21:03:18 <iwearsable> And I'm not sure where that is supposed to come from, package search (file search) turns up nothing
2015-04-29 21:04:45 <iwearsable> This started because I was using the third repository instead of nl, it was lagging behind. After switching to nl, and doing updates, I lost mplayer. Then I kept breaking more stuff after that trying to get mplayer to work... who knows how much I've screwed up now lol
2015-04-29 21:05:16 <iwearsable> I'm not sure if better off just to reinstall because I'm really stuck now without Firefox
2015-04-29 21:06:25 <iwearsable> It was having problems with icu-libs too, near the start, but I cleared that up (and apparently broke other stuff). Hard to remember everytnig I've done
2015-04-29 21:09:58 <dalias> iwearsable, yeah that seems to have been the source of my problems too....
2015-04-29 21:10:15 <dalias> trying to fix it now
2015-04-29 21:10:19 <iwearsable> These kinds of problems really confuse me
2015-04-29 21:12:18 <iwearsable> I'm going to try some things, if it works I'll let you know
2015-04-29 21:27:44 <dalias> almost got a usable system again
2015-04-29 21:27:54 <dalias> i removed the crap that wanted the old libvpx (mplayer :)
2015-04-29 21:28:02 <dalias> but i broke a lot of stuff in the mean time... :/
2015-04-29 21:31:04 <iwearsable> Yeah, I've been there already. Is the problem that mplayer wants new libvpx and firefox 37 wants old libvpx?
2015-04-29 21:32:47 <iwearsable> At the moment I can't even fetch APKINDEX :/ break time
2015-04-29 21:34:03 <bitshark__> iwearsable sorry to hear that
2015-04-29 21:34:18 <bitshark__> i made a patch to apk cause i had package problems
2015-04-29 21:35:45 <iwearsable> No big deal really, I have many computers
2015-04-29 21:37:59 <iwearsable> I installed apk-tools-static and can update again
2015-04-29 21:39:04 <iwearsable> Actually, not sure that had anything to do with it, or your patch
2015-04-29 21:43:12 <iwearsable> Bah, I'm just going to reinstall later
2015-04-29 21:43:12 <iwearsable> Too confused
2015-04-29 21:45:05 <dalias> iwearsable, other way around
2015-04-29 21:53:46 <iwearsable> Oh jeez my internet is cycling again (not an alpine linux problem, my terrible ISP). It's one of those days haha
2015-04-29 22:05:08 <coredumb> iwearsable, dalias you on edge ?
2015-04-29 22:05:12 <iwearsable> Yes
2015-04-29 22:05:18 <coredumb> got same problem as you on edge
2015-04-29 22:05:55 <coredumb> clandmeter: ^ seems i'm not the only one hitting this icu/libvpx issue
2015-04-29 22:21:43 <iwearsable> I got firefox to work by downloading the ports tree, going to main/icu, abuild checksum && abuild -r -d (-d was necessary or it fails on dependency check)
2015-04-29 22:23:09 <iwearsable> Then after it builds the package install with apk add /path/to/package/icu-.apk
2015-04-30 00:03:44 <iwearsable> I don't know if worth mentioning, but I tried to build mplayer from the aports tree and it fails
2015-04-30 00:03:56 <iwearsable> common.mak:45: recipe for target 'libavcodec/rv34.o' failed
2015-04-30 00:37:07 <iwearsable> Arch updated their mplayer to build against libvpx 1.4.0. It looks like they download ffmpeg-2.6.2 and use that when building mplayer instead of the included ffmpeg
2015-04-30 00:37:11 <iwearsable> https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/mplayer&id=e0a147caa4659cf1723b7caf9d950f0de12acb1c
2015-04-30 00:41:16 <iwearsable> I could probably make a new APKBUILD that works, but I don't know if it would be proper and I've never contributed like that before. I figure I would probably cause more problems as my knowledge is not that great
2015-04-30 00:43:41 <iwearsable> It'
2015-04-30 00:44:15 <iwearsable> It's one thing to make something that works for me, but a whole nother thing to understand the proper way
2015-04-30 00:47:47 <iwearsable> But is the main problem that mplayer need to be built against libvpx 1.4.0? I could make a bugreport if that's what all of this stems from
2015-04-30 00:48:40 <iwearsable> If there isn't already one
2015-04-30 00:59:19 <packeteer> so, alpine has ports ?
2015-04-30 01:37:00 <iwearsable> packeteer: http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2015-04-30 01:41:38 <packeteer> thx
2015-04-30 01:52:14 <packeteer> well sheeeit, corporate proxy prohibits access to the git repo
2015-04-30 01:53:42 <packeteer> belay that, http works
2015-04-30 01:54:06 <packeteer> and I spoke to soon
2015-04-30 02:07:58 <packeteer> ps. http access to the repo would be nice
2015-04-30 10:24:34 <jomat> weird, apk upgrade didn't upgrade musl, had to apk add musl manually to get it updated to get rid of Error relocating /usr/lib/libstdc++.so.6: get_nprocs: symbol not found
2015-04-30 12:10:21 <ScrumpyJack> I'm considering upgrading from 3.1 to 3.2. anything i should look out for?
2015-04-30 12:10:53 <ScrumpyJack> ( normally i live on the edge, but this host I try to keep "stable" )
2015-04-30 12:27:58 <clandmeter> ScrumpyJack: 3.2 is edge
2015-04-30 13:32:29 <sillysausage> ncopa: do you use claws-mail-plugins-rssyl?
2015-04-30 13:33:17 <sillysausage> there was a bug with the dates showing up as Thu,  1 Jan 1970 00:00:00 GMT for all feeds that has been apparently fixed http://comments.gmane.org/gmane.mail.sylpheed.claws.general/59967
2015-04-30 13:33:24 <sillysausage> http://git.claws-mail.org/?p=claws.git;a=blobdiff;f=src/plugins/rssyl/libfeed/date.c;h=92f5a860dae3132ff8476eabf19f7dc58f9052d9;hp=c37756d44f72adef6b79922d49c2b192141671fa;hb=72510f27b10889e119b42804835e8beaf70bc18d;hpb=d4fa1c076942efbf17094123c14530d607fb6d09
2015-04-30 13:34:29 <sillysausage> mm that url was longer than i thought: http://git.claws-mail.org/?p=claws.git;a=commit;h=72510f27b10889e119b42804835e8beaf70bc18d
2015-04-30 13:36:11 <ncopa> hi
2015-04-30 13:36:19 <ncopa> i dont use the rssyl
2015-04-30 13:37:05 <ncopa> sillysausage: you want this patch? http://git.claws-mail.org/?p=claws.git;a=patch;h=72510f27b10889e119b42804835e8beaf70bc18d
2015-04-30 13:37:10 <Phiphler> if I want to use Alpine as a Xen dom0, do I download the regular ISO or the Xen ISO? I'm new to hypervisors so I'm trying to get it right from the start
2015-04-30 13:37:20 <sillysausage> sure ncopa
2015-04-30 13:37:33 <sillysausage> because most of my rss feeds are showing up as 1970 :P
2015-04-30 13:37:45 <sillysausage> and you're the maintainer of it, it would be very nice if you could commit it
2015-04-30 13:38:02 <sillysausage> kind of breaks the point of rss if i can't tell what article is new :D
2015-04-30 13:38:06 <ncopa> i'm on it
2015-04-30 13:38:14 <sillysausage> :D
2015-04-30 13:38:19 <sillysausage> thanks
2015-04-30 14:18:46 <ncopa> sillysausage: claws-mail pushed
2015-04-30 14:18:53 <sillysausage> yay
2015-04-30 14:25:30 <darkfader> Phiphler: i tnd to use the normal iso and add the xen bits, but you might be best off doing some testing with the dom0 iso
2015-04-30 14:25:46 <darkfader> just if it doesn't work, then dont hesitate switching
2015-04-30 14:29:17 <crow> darkfader and xen bits are? installing xen kernel?
2015-04-30 14:29:57 <darkfader> crow: are you phiphler?
2015-04-30 14:30:14 <darkfader> we've some documentation in the wiki, sec
2015-04-30 14:30:40 <darkfader> http://wiki.alpinelinux.org/wiki/Xen_Dom0
2015-04-30 14:30:54 <darkfader> my version is a little more, bonding etc. but that is not really relevant at start
2015-04-30 14:36:49 <darkfader> found it, http://confluence.wartungsfenster.de/display/Adminspace/Alpine+Linux+Xen+tips ; but not updated since 2013
2015-04-30 14:55:45 <Phiphler> darkfader: thanks, Alpine seems like a good dom0 for many reasons. Now we just gotta choose what do use for the network bridge...
2015-04-30 15:20:00 <crow> darkfader no i am not, i just read some stuff about xen to
2015-04-30 15:46:03 <sillysausage> ncopa: how long till say that package is pushed to http://nl.alpinelinux.org/alpine/v3.1/main/x86_64/
2015-04-30 15:46:25 <sillysausage> oh wait wrong repo
2015-04-30 15:46:34 <sillysausage> it's in edge
2015-04-30 16:09:29 <sillysausage> ncopa: how do i install that one package from edge
2015-04-30 16:09:47 <sillysausage> seems to be installing it from main, do i need to use all of claws-mail from edge?
2015-04-30 16:10:32 <sillysausage> $ sudo apk upgrade claws-mail-plugins-rssyl@edge
2015-04-30 16:10:37 <sillysausage> seems to still use the cached one
2015-04-30 16:11:06 <sillysausage> ie claws-mail-plugins-rssyl-3.11.0-r0.ddab57b3.apk
2015-04-30 16:16:20 <clandmeter> apk add pkg@edge
2015-04-30 16:18:46 <sillysausage> sudo apk add claws-mail-plugins-rssyl@edge
2015-04-30 16:19:22 <sillysausage> doesn't seem to be installing this: http://nl.alpinelinux.org/alpine/edge/main/x86_64/claws-mail-plugins-rssyl-3.11.1-r2.apk
2015-04-30 16:20:15 <fireglow> OT: Has anybody had success in running Alpine Linux under bhyve?
2015-04-30 16:26:36 <clandmeter> sillysausage: did you pin that repo in repositries?
2015-04-30 16:26:40 <sillysausage> yes
2015-04-30 16:26:52 <sillysausage> eg
2015-04-30 16:26:56 <sillysausage> $ cat /etc/apk/repositories
2015-04-30 16:26:58 <sillysausage> /media/usb/apks
2015-04-30 16:27:00 <sillysausage> http://nl.alpinelinux.org/alpine/v3.1/main
2015-04-30 16:27:02 <sillysausage> @edge http://nl.alpinelinux.org/alpine/edge/main
2015-04-30 16:27:04 <sillysausage> @testing http://nl.alpinelinux.org/alpine/edge/testingy
2015-04-30 16:27:11 <clandmeter> apk update
2015-04-30 16:27:13 <sillysausage> oops there's no y on the end, that happened when i pasted
2015-04-30 16:27:16 <sillysausage> yeah i did that
2015-04-30 16:27:46 <sillysausage> but when i do
2015-04-30 16:27:48 <sillysausage> sudo apk add claws-mail-plugins-rssyl@edge
2015-04-30 16:27:51 <sillysausage> it just installs the one from stable
2015-04-30 16:28:24 <sillysausage> $ ls -la /media/usb/cache/ |grep claws-mail-plugins-rssyl
2015-04-30 16:28:26 <sillysausage> -rwxr-xr-x    1 root     root         56337 Apr 29 18:34 claws-mail-plugins-rssyl-3.11.0-r0.ddab57b3.apky
2015-04-30 16:28:46 <clandmeter> apk search -v claws-mail-plugins-rssyl
2015-04-30 16:28:46 <Jean-Scotch> apk add claws-mail-plugins-rssyl@testing should be the one
2015-04-30 16:28:53 <sillysausage> ah.
2015-04-30 16:29:14 <clandmeter> :)
2015-04-30 16:29:22 <sillysausage> yeah i think i needed to install it from testing
2015-04-30 16:29:27 <sillysausage> and i just noticed it's 3.11.1
2015-04-30 16:29:35 <sillysausage> (late here) that's why!
2015-04-30 16:32:18 <sillysausage> hmm.
2015-04-30 16:33:42 <sillysausage> still when i install claws-mail@testing and then try to do sudo apk add claws-mail-plugins-rssyl@edge i get: https://dpaste.de/G4EH
2015-04-30 16:33:59 <sillysausage> would i need to remove all those things
2015-04-30 16:34:03 <sillysausage> and install them from testing?
2015-04-30 16:34:35 <sillysausage> as in install weechat-1.1.1-r2 from @edge
2015-04-30 16:34:44 <sillysausage> and glib-networking
2015-04-30 16:42:33 <ScrumpyJack> clandmeter: silly me. of course it is
2015-04-30 16:44:00 <sillysausage> okay so the main problem seems to be some kind of conflict https://dpaste.de/BMfM
2015-04-30 16:44:31 <sillysausage> even though neither nettle or gnutls is installed
2015-04-30 17:05:44 <sillysausage> oops internet died.
2015-04-30 17:09:54 <fireglow> aww, this happens trying to boot the iso in bhyve: https://pastee.org/vcpwf
2015-04-30 17:14:47 <andyshinn> Is it safe to have a list of repositories in the apk repositories file for fail over purposes? Is it documented somewhere how this works?
2015-04-30 17:16:31 <andyshinn> The wiki just says that it is valid. But I'm wondering if it just goes down the list in case of error or other
2015-04-30 18:37:26 <Mo0O> hi there, do you know where I can find the list of option we can pass to abuild in APKBUILD?
2015-04-30 21:25:12 <systmkor> Mo0O: what abuild in APKBUILD?
2015-04-30 21:38:55 <dalias> does alpine not have good broad-coverage multilingual font packages?
2015-04-30 21:39:19 <dalias> yes it's easy to drop ttf files in ~/.fonts
2015-04-30 21:39:46 <dalias> but it would be really nice to be able to just apk add and not have to worry about encountering "hex squares"
2015-04-30 22:11:34 <Mo0O> systmkor: yes in APKBUILD, sound like I can overwrite the fact we can install in /opt
2015-04-30 22:13:03 <Diftraku> dalias: everyone loves hex squares, broken scandies, seagulls and the magic questionmarks
2015-04-30 22:18:33 <dalias> diftraku, :
2015-04-30 22:18:34 <dalias> :)
2015-04-30 22:19:20 <dalias> even if i can't read the text, i'd rather see it rendered correctly and at least immediately recognize what language it is than have to try to read the tiny digits in the hex boxes and remember which unicode block that is
2015-04-30 23:07:52 <systmkor> Mo0O: for me I thought it yelled at you during the build phase if you you tried installing to /opt or /usr/local/[sbin,bin]
2015-04-30 23:09:56 <systmkor> Mo0O: if it isn't documented in the wiki, that should be changed
2015-04-30 23:12:41 <ryuo> Is there some way to install alpine to a ZFS or BTRFS root?