2018-12-01 00:04:18 <winny> can apk use rsync mirrors directly?
2018-12-01 00:08:43 <danieli> winny: no
2018-12-01 00:08:50 <danieli> why would you even want to do that?
2018-12-01 00:10:30 <winny> that's a good question
2018-12-01 00:10:44 <danieli> just use a http:// mirror
2018-12-01 00:10:48 <winny> mainly thought it'd be easier to set up a rsync mirror than set up a http daemon
2018-12-01 00:11:21 <danieli> fair enough, that's true
2018-12-01 00:11:32 <winny> not that setting up a http daemon is hard or anything
2018-12-01 00:11:35 <winny> just lazy lol
2018-12-01 00:11:43 <danieli> are you setting up a mirror for yourself?
2018-12-01 00:11:54 <winny> it's for a stand alone vm system for a security lab
2018-12-01 00:12:12 <winny> (not to be networked)
2018-12-01 00:13:06 <danieli> gotcha - well, you'd have to set up some httpd for apk to actually use it
2018-12-01 00:13:43 <winny> indeed, seems straight forward. Had just set my host system's apk path to the directory path, as i wasn't able to use rsync:// and thought i had misconfigured rsync
2018-12-01 00:14:10 <clandmeter> http more difficult than rsync?
2018-12-01 00:14:22 <winny> very slightly lol, nothing of noteworthy
2018-12-01 00:14:33 <clandmeter> i mean even busybox httpd would suffice
2018-12-01 00:15:06 <danieli> it's hardly more difficult unless you aren't used to rsync, to be fair
2018-12-01 00:15:26 <danieli> I suppose the logic is that extra daemon requires more time and effort
2018-12-01 00:17:24 <winny> at any rate, i think I missed the documentation for the /etc/apk/repositories file, is that in a seperate package?
2018-12-01 00:17:39 <danieli> apk-tools
2018-12-01 00:18:01 <danieli> clandmeter: i actually don't think there is an apk-tools-doc?
2018-12-01 00:18:31 <clandmeter> i dont think there is.
2018-12-01 00:18:40 <clandmeter> maybe there is a pr about it
2018-12-01 00:18:40 <winny> thanks
2018-12-01 00:18:49 <clandmeter> or it was abuild
2018-12-01 00:18:54 <winny> 
ACTION adds todo to package rlwrap

2018-12-01 00:19:09 <danieli> there's really no doc for the alpine linux package manager?
2018-12-01 00:19:18 <clandmeter> the wiki
2018-12-01 00:19:22 <danieli> that's kind of.. shocking
2018-12-01 00:19:31 <danieli> 
ACTION prays to the openbsd gods

2018-12-01 00:19:31 <winny> hmmm, the wiki is nice, but it's not 'online documentation'
2018-12-01 00:19:40 <winny> 
ACTION contemplates downloading the wiki

2018-12-01 00:20:14 <clandmeter> we would like to have some sort of "official" online documentation.
2018-12-01 00:20:24 <clandmeter> but we need somebody to manage that.
2018-12-01 00:20:39 <SpaceToast> I'd be interested, assuming we version manage it and write it in asciidoc
2018-12-01 00:20:42 <danieli> i'd like manpages more first
2018-12-01 00:21:10 <clandmeter> SpaceToast, thats absolutely fine
2018-12-01 00:21:10 <SpaceToast> danieli: you can always build the asciidoc -> man page :D
2018-12-01 00:21:23 <SpaceToast> (ignore the part where it requires a very specific format :) )
2018-12-01 00:21:36 <danieli> yeah no, keep me out of this.
2018-12-01 00:21:53 <SpaceToast> clandmeter: alright; I'm not really officially part of the project or anything, but if the idea's approved definitely ping me
2018-12-01 00:22:28 <SpaceToast> there's lots of stuff missing in the wiki re: apkbuild for example (no mention of default_openrc() that I could see, for one)
2018-12-01 00:22:36 <danieli> how does joining Alpine even work?
2018-12-01 00:22:44 <SpaceToast> but a wiki is sufficiently "not official" that I haven't bothered to fix it (before doing other things)
2018-12-01 00:23:47 <danieli> yeah, it contains *a lot* of subjective and inconsistent pages, outdated ones, you name it. I occasionally look at the Arch wiki and sigh
2018-12-01 00:23:56 <clandmeter> SpaceToast, i like your idea.
2018-12-01 00:24:15 <clandmeter> we prefer to manage as much as possible in git
2018-12-01 00:24:19 <SpaceToast> mhm
2018-12-01 00:24:52 <SpaceToast> clandmeter: my 4 gripes with alpine so far are docs, booting process, lack of proper git in abuild and that edge is considered "unstable" (I personally prefer the rolling release model)
2018-12-01 00:24:57 <SpaceToast> if I can help with docs, I'm definitely in :)
2018-12-01 00:25:06 <SpaceToast> (or with any of those other things)
2018-12-01 00:25:11 <danieli> edge is the development branch
2018-12-01 00:25:17 <SpaceToast> (the last thing that annoys me is bad/old openrc initscripts, but I'm already working on that)
2018-12-01 00:25:43 <SpaceToast> danieli: I'd be fine with making a new branch ala "pre-edge" or something, I just don't like hard releases (bug-fix updates only) for most use-cases (aka outside of servers)
2018-12-01 00:25:43 <clandmeter> the docs project never really took off.
2018-12-01 00:25:58 <clandmeter> cause nobody took the step to organize it.
2018-12-01 00:25:59 <danieli> it never really materialized properly imo.
2018-12-01 00:26:02 <danieli> exactly
2018-12-01 00:26:09 <SpaceToast> clandmeter: are there any preferences unto how to structure it?
2018-12-01 00:26:23 <SpaceToast> I could start literally now, on my time off, assuming it'll go somewhere
2018-12-01 00:26:33 <clandmeter> we had a few discussions, but to be honest we are open to ideas.
2018-12-01 00:26:41 <SpaceToast> mhm
2018-12-01 00:26:43 <danieli> we were thinking to have official docs @ docs.a.o, more tutorial-type and non-critical things @ wiki.a.o
2018-12-01 00:26:52 <SpaceToast> clandmeter: can you correct me if I'm wrong in this -
2018-12-01 00:26:57 <danieli> as to the implementation itself, nothing really materialized
2018-12-01 00:27:32 <SpaceToast> alpine is divided into a couple "important" parts - abuild / apk-tools / etc; mkinitfs; and the policy unto how to package things (i.e separate everything)
2018-12-01 00:28:18 <SpaceToast> so if I was to write "official docs", it'd be divided into abuild-related manpage-compatible files, general "how to use" docs (ala gentoo handbook), and deeper technical details (e.g how the boot process works)
2018-12-01 00:28:42 <clandmeter> SpaceToast, i think you are pretty close.
2018-12-01 00:29:00 <SpaceToast> I think the biggest challenge would be keeping the docs up to date
2018-12-01 00:29:15 <clandmeter> the first things would be to properly document our in house solutions like you mention.
2018-12-01 00:29:19 <SpaceToast> (e.g I could write them to spec of abuild's current state, but keeping up with changes n that could be hard)
2018-12-01 00:29:27 <danieli> yeah, a lot of changes seem to happen without being documented anywhere but in the git commit history
2018-12-01 00:29:37 <SpaceToast> danieli: and in parallel of each other, becoming incompatible
2018-12-01 00:29:55 <SpaceToast> I recall setup-alpine having an order of operations snafu somewhere around the tail end, for example
2018-12-01 00:30:26 <SpaceToast> it might make sense to have a kind of "changelog" mailing list, where any changes to in-house solutions would be published (and make maintaining docs a lot easier)
2018-12-01 00:30:35 <SpaceToast> anyway, I have to go back to cleaning \o poke me if this goes anywhere
2018-12-01 00:31:44 <clandmeter> SpaceToast, we have an old #alpine-docs channel which was setup long time ago when it was mentioned. If you want please join.
2018-12-01 11:05:10 <mhavela> Hi all. Does anyone know if setup-apkrepos is broken on Alpine 3.8?
2018-12-01 11:05:27 <mhavela> seems the script looks for mirrors at https://mirrors.alpinelinux.org/mirrors.tx
2018-12-01 11:05:46 <mhavela> but I don't think that URL work
2018-12-01 11:26:23 <mhavela> Anyone knows what telnet client we have in alpine 3.8?
2018-12-01 11:27:03 <AinNero> does the nc applet from busybox do the job for you?
2018-12-01 11:34:00 <mhavela> not really. well - I could use nc, but it complicates things.
2018-12-01 11:34:20 <mhavela> but I just managed to figure it out
2018-12-01 11:34:42 <mhavela> telnet seems to be part of busybox-extras package
2018-12-01 11:35:08 <mhavela> Thanks for trying to help me @AinNero
2018-12-01 11:35:39 <mhavela> So that solves my telnet issue.
2018-12-01 11:42:30 <fisherthewol> exit
2018-12-01 12:27:09 <clandmeter> mhavela, url is incorrect?
2018-12-01 12:28:09 <_ikke_> it should be mirrors.txt
2018-12-01 12:28:28 <clandmeter> And telnet is in BusyBox extras
2018-12-01 12:28:42 <clandmeter> Ah already mentioned
2018-12-01 12:48:49 <mhavela> Might be my host is having some other issues why it doesn't reach mirrors.txt
2018-12-01 14:59:38 <laskin> Hello. I need guide on how to setup gre tunnel using /etc/network/interfaces. Is it available somewhere?
2018-12-02 13:12:20 <mepholic> z/b 11
2018-12-02 13:12:22 <mepholic> er
2018-12-02 16:54:29 <mpmc> Is there something wrong with the build servers? http://build.alpinelinux.org/ armhf is missing unit on edge.
2018-12-02 16:56:23 <_ikke_> build-edge-armhf?
2018-12-02 16:57:16 <mpmc> Lol, weird it said it was idle a second ago, now "failed".
2018-12-02 17:28:37 <mpmc> _ikke_: Any clue as to why some of the build logs are missing. I'm guessing failure to build?
2018-12-02 17:28:48 <_ikke_> mpmc: probably
2018-12-02 17:29:40 <mpmc> _ikke_: OK. I'll have to wait & hope that it get fixed.
2018-12-02 17:29:41 <feuerteufel> Aber irgendwas muss funktioniert haben, denn mein Kontobuch ist aktuell per heute :-))
2018-12-02 17:30:12 <_ikke_> feuerteufel: Dis is keine Deutsches kanall :)
2018-12-02 17:30:21 <_ikke_> (sorry for any broken german :P)
2018-12-02 17:30:34 <feuerteufel> Sorry wrong chennal
2018-12-02 17:30:37 <_ikke_> hehe
2018-12-02 17:33:53 <mpmc> _ikke_: There isn't much info on the build server setup of alpine. I'm guessing builds are triggered on git commit rather than automated. Who manages the build servers? I'm guessing ncopa ?
2018-12-02 17:34:51 <_ikke_> What's the difference between 'on git commit' and 'automated'?
2018-12-02 17:35:08 <_ikke_> mpmc: There are a couple of people who maintain the infrastructure
2018-12-02 17:35:37 <_ikke_> mpmc: https://git.alpinelinux.org/cgit/aports/tree/main/aports-build/aports-build
2018-12-02 17:36:25 <mpmc> _ikke_: Well, automated is a bad word, maybe nightly is better suited :)
2018-12-02 17:36:42 <_ikke_> so scheduled
2018-12-02 17:37:27 <mpmc> Yeah, sorry, half asleep here, not long woke up :p
2018-12-02 17:37:30 <_ikke_> rather than triggered
2018-12-02 17:37:43 <_ikke_> Not sure how that would improve things
2018-12-02 17:38:46 <mpmc> _ikke_: I'm just asking random questions :)
2018-12-02 17:38:53 <_ikke_> mpmc: note that we are working on improving the build process, it just takes time
2018-12-02 17:40:52 <SpaceToast> the main improvement I see we could haev atm is auto-bumping the dep tree
2018-12-02 17:41:07 <SpaceToast> (not sure if that happens or not, my own repo is on full manual mode :) )
2018-12-02 17:41:58 <mpmc> _ikke_: It's all good :) I tried to switch my vps to alpine but awall bailed with something along the lines of "unknown option --mask". pretty sure that was caused by the kernel being so old [2.6]!
2018-12-02 17:43:24 <_ikke_> SpaceToast: I guess that should be part of a dev toolkit (akin to abuild), rather than the build system
2018-12-02 17:43:43 <SpaceToast> _ikke_: yeah, I'm being a bit more general
2018-12-02 17:44:04 <SpaceToast> are there any docs for aports-build? (yeah, I know, look who's asking ;) )
2018-12-02 17:44:14 <SpaceToast> (if there aren't I'll add that to the list of what to write up)
2018-12-02 17:44:58 <_ikke_> SpaceToast: afaik, there isn't
2018-12-02 17:45:13 <_ikke_> And it's probable scheduled for replacement anyway
2018-12-02 17:48:15 <mpmc> So what do you guys use alpine for? :)
2018-12-02 17:49:17 <_ikke_> currently I have a VPS running alpine
2018-12-02 17:49:39 <mpmc> _ikke_: Cool, what provider/kernel? :p
2018-12-02 17:49:49 <_ikke_> mpmc: scaleway
2018-12-02 17:50:04 <_ikke_> 4.5.7-std-3
2018-12-02 17:50:23 <mpmc> _ikke_: Do you use awall?
2018-12-02 17:50:35 <SpaceToast> all of my personal servers non-desktop/laptop systems run alpine
2018-12-02 17:50:52 <SpaceToast> I also have my own repo so I don't have to wait for things to get upstreamed (don't actually use it though, no guarantees)
2018-12-02 17:51:17 <SpaceToast> the main difficulty re: laptops etc is lack of some packages (I'm slowly working on it), some boot stuff, and the lack of plasma5 (my preferred graphical environment)
2018-12-02 17:51:27 <_ikke_> mpmc: no, just plain iptables ATM
2018-12-02 17:53:00 <mpmc> alpine could use a simple firewall app tbh, awall is great but requires a lot of config editing.
2018-12-02 17:53:38 <_ikke_> does it?
2018-12-02 17:54:21 <SpaceToast> I haven't tried using it, but from what I saw it's basically shorewall, but it generates the output (for init.d/iptables to use) rather than doing... whatever shorewall does
2018-12-02 17:54:25 <SpaceToast> I personally use ferm though
2018-12-02 17:54:36 <SpaceToast> and await bpfilter
2018-12-02 17:57:44 <_ikke_> Just updated kernel to 4.15.11
2018-12-02 18:00:15 <mpmc> Any reason why the package filter doesn't allow searching by description?
2018-12-02 18:00:39 <_ikke_> you mean pkgs.a.o?
2018-12-02 18:01:44 <mpmc> _ikke_: Yarr.
2018-12-02 18:02:04 <_ikke_> No idea myself
2018-12-02 18:05:43 <mpmc> <_< went pirate there for no reason :/
2018-12-02 18:27:48 <mps> SpaceToast: yes, ferm works quite nice
2018-12-02 18:28:13 <mps> I use it for many years on different distros
2018-12-02 18:33:02 <boules> Congrats on your web design. It's pleasing.
2018-12-02 18:41:51 <_ikke_> boules:  thanks for the compliment
2018-12-03 09:19:03 <jwh> mpmc: ufw is in testing
2018-12-03 13:29:23 <misterunknown> Hi. I have a problem with snmpd. It won't start, saying: "Error opening specified endpoint "udp:0.0.0.0:161""
2018-12-03 13:30:03 <misterunknown> I tried all different combinations, but it does not start at all. Not with "udp:161" or with "udp:127.0.0.1:161".
2018-12-03 13:30:08 <misterunknown> Any ideas?
2018-12-03 13:30:52 <misterunknown> (Alpine Edge)
2018-12-03 13:36:41 <programmerq> misterunknown▸ googling the error message shows a similar error message where the solution was that the init script was explicitly passing something different on the command line when it started snmpd. I doubt that's what's going on, but check to see what command is actually getting called when you run it.
2018-12-03 13:37:02 <programmerq> if you have anything set in SNMPD_OPTS, then the openrc script will pass it in
2018-12-03 13:37:31 <programmerq> it's just a guess
2018-12-03 13:45:44 <misterunknown> programmerq: Unfortunately not. I tried starting snmpd from command line, with -Lo to see error message in stdout.
2018-12-03 13:45:50 <misterunknown> Same behaviour.
2018-12-03 13:51:14 <mps> misterunknown: do you start it as root or normal user?
2018-12-03 13:52:05 <misterunknown> mps: I tried it as root...
2018-12-03 13:52:34 <misterunknown> Thats also what the init script does, if I see that correctly
2018-12-03 13:52:36 <mps> init script or 'snmpd' command
2018-12-03 13:53:01 <misterunknown> mps: Both.
2018-12-03 13:53:18 <mps> can you check that something isn't already listening on udp 161 port
2018-12-03 13:54:10 <misterunknown> mps: It is not, at least I don't see anything. But snmpd does some things which I don't understand: http://termbin.com/cazc
2018-12-03 13:55:24 <misterunknown> In line 2258 it tries to bind the port, which fails because the address is already in use.
2018-12-03 13:55:45 <misterunknown> But in line 2163 the process already binds this port...
2018-12-03 13:56:00 <misterunknown> So it seems like the process tries to bind the port twice...
2018-12-03 13:56:27 <mps> look at the end of file you posted: ind(8, {sa_family=AF_INET, sin_port=htons(161), sin_addr=inet_addr("127.0.0.1")}, 16) = -1 EADDRINUSE (Address in use)
2018-12-03 13:56:53 <mps> EADDRINUSE
2018-12-03 13:57:14 <misterunknown> mps: Correct, but it successfully binds the same port some lines earlier.
2018-12-03 13:59:17 <misterunknown> (btw. termbin.com is amazing... Usable via netcat :-P )
2018-12-03 13:59:32 <mps> misterunknown: bind successfully with port 161? can't see
2018-12-03 14:00:14 <mps> oh, soryy
2018-12-03 14:00:29 <mps> sory, need better search in browser
2018-12-03 14:00:32 <misterunknown> mps: Line 2164: >> bind(9, {sa_family=AF_INET, sin_port=htons(161), sin_addr=inet_addr("127.0.0.1")}, 16) = 0
2018-12-03 14:00:42 <misterunknown> Ah ok :)
2018-12-03 14:00:58 <mps> I seen, my search '
2018-12-03 14:00:59 <Shiz> might be a duplicate interface definition then?
2018-12-03 14:01:28 <mps> yes, looks like it tries to bind two times to same port
2018-12-03 14:01:59 <misterunknown> You mean two agentAddress statements in the configuration?
2018-12-03 14:03:14 <mps> something like that. some years passed when I last time worked with snmpd
2018-12-03 14:04:25 <misterunknown> This is my snmpd.conf http://termbin.com/p3bc
2018-12-03 14:04:41 <misterunknown> (Comments are stripped out)
2018-12-03 14:06:46 <mps> could be stupid idea, but could you try to remome (comment) 'agentAddress' from config
2018-12-03 14:06:59 <mps> s/remome/remove/
2018-12-03 14:09:51 <misterunknown> mps: Hm, interesting. I removed the "agentAddress" directive, and see this: http://termbin.com/c69d
2018-12-03 14:10:38 <misterunknown> Then snmpd binds udp:127.0.0.1:161 and at the end tries to bind udp:0.0.0.0:161 which also fails with "EADDRINUSE"
2018-12-03 14:11:14 <mps> fyi, just added snmp to and started it with default config without probem (AL stable, 3.8.1)
2018-12-03 14:12:05 <misterunknown> mps: Well, I have some 3.8 systems, too. I try to run it there...
2018-12-03 14:13:40 <mps> did you added someting in /etc/conf.d/snmpd
2018-12-03 14:17:14 <misterunknown> mps: Nope. The only relevant line in this file is: >> OPTS="${OPTS} -LSwd -Lf /dev/null"
2018-12-03 14:17:42 <misterunknown> mps: I tried it on an alpine 3.8 system, which works fine...
2018-12-03 14:17:55 <misterunknown> So maybe a bug?
2018-12-03 14:19:51 <mps> could be. post a issue at bugs.a.o if you find something about it
2018-12-03 14:20:25 <misterunknown> bugs.a.o?
2018-12-03 14:20:44 <_ikke_> https://bugs.alpinelinux.org
2018-12-03 14:24:20 <misterunknown> There is no "edge" in "affected versions". Is it right to pick 3.9.0?
2018-12-03 14:29:04 <danieli> i don't think 3.9.0 is officially released yet
2018-12-03 14:30:41 <misterunknown> danieli: Yes, but this version is available in the select box. So I guess this is "edge" right now?
2018-12-03 14:32:03 <mps> misterunknown: init script and conf.d file are changed a lot in last commit, should be checked if new version works
2018-12-03 14:59:09 <misterunknown> mps: Well, I use the latest init.d and conf.d files, as far as I see.
2018-12-03 14:59:48 <TBB> hum hum. does anyone use gtk-ssh-askpass with openssh on Alpine?
2018-12-03 15:13:45 <TBB> (that is, the graphical password entry tool from openssh-askpass that's in edge-testing)
2018-12-03 16:33:13 <auxin> I enabled access to 'testing' packages and performed an upgrade, which rendered my system useless, how can I undo this fiasco?
2018-12-03 16:36:37 <auxin> I'm currently starting over the installation on a separate disk, but I spent a loooot of time on the previous one and would like to rescue it.
2018-12-03 17:23:30 <pheoxy> is there a way to get vfio modules in mkinitfs?
2018-12-03 17:23:43 <TBB> yes there is
2018-12-03 17:24:11 <TBB> look at /etc/mkinitfs/features.d/ and you'll quickly figure out how it works; also look at /etc/mkinitfs/mkinitfs.conf
2018-12-03 17:24:50 <TBB> damn, that auxin person left already, could've given him a tip or two
2018-12-03 17:24:59 <pheoxy> oh ok I'll have a look, was looking at https://github.com/alpinelinux/mkinitfs/blob/master/Makefile . opps
2018-12-03 17:25:49 <AinNero> no, its an runtime configuration thing
2018-12-03 17:25:52 <AinNero> not build-time
2018-12-03 17:26:26 <AinNero> pheoxy: what TBB said
2018-12-03 17:49:20 <pheoxy> TBB, AinNero, ok thanks
2018-12-03 17:53:08 <pheoxy> TBB, So I looked in cat /etc/mkinitfs/features.d/virtio.modules for "vfio_pci vfio vfio_iommu_type1 vfio_virqfd" but it only has "block/virtio* virtio virtio_net". Am I missing something?
2018-12-03 17:53:26 <pheoxy> there is no vfio.modules
2018-12-03 17:53:59 <mps> pheoxy: you can create it
2018-12-03 17:54:12 <pheoxy> how would I do that?
2018-12-03 17:54:50 <mps> create vfio.modules file
2018-12-03 17:55:36 <mps> and add vfio to /etc/mkinitfs/mkinitfs.conf
2018-12-03 17:55:40 <pheoxy> in /etc/mkinitfs/features.d/ ?
2018-12-03 17:56:07 <mps> yes
2018-12-03 17:57:23 <mps> can't remember exactly, but I did something similar about year ago
2018-12-03 17:58:27 <mps> or, you can add needed vfio modules to /etc/mkinitfs/features.d/virtio.modules
2018-12-03 17:59:56 <pheoxy> I think I'll make a vfio.modules file to keep it closer to vanilla
2018-12-03 18:00:11 <SpaceToast> pheoxy: what exactly does vfio do, relative to virtio?
2018-12-03 18:01:10 <pheoxy> SpaceToast, allows you to assign devices in virtual machines
2018-12-03 18:01:19 <pheoxy> SpaceToast, from the host
2018-12-03 18:01:53 <SpaceToast> oh so more passthrough-style than paravirt, gotcha
2018-12-03 18:01:56 <SpaceToast> ty :)
2018-12-03 18:02:13 <pheoxy> where would I find the kernel/drivers folder to assign the right files?
2018-12-03 18:02:49 <pheoxy> yeh, want to pass through some pci devices and usb ports
2018-12-03 18:03:52 <pheoxy> planning on running a baremetal server with pci passthrough so I can create a pseudo lab for some hobbies
2018-12-03 18:04:46 <pheoxy> never ming found kernel in /lib/modules
2018-12-03 18:06:15 <pheoxy> is the asterisk in a mkinitfs module such as mean all files beginning with virtio?
2018-12-03 18:07:37 <SpaceToast> I suspect it's a shell glob
2018-12-03 18:08:01 <pheoxy> so... yes?
2018-12-03 18:09:23 <SpaceToast> only files that would match a shell glob, if I am correct ;)
2018-12-03 18:09:38 <SpaceToast> /some/*virtio would not match /other/virtio, for example
2018-12-03 18:10:37 <pheoxy> SpaceToast, so to compare https://bpaste.net/show/59ff12e2ca4c
2018-12-03 18:11:01 <pheoxy> I could get rid of some with just vfio* ?
2018-12-03 18:12:22 <SpaceToast> I'm not a maintainer/dev of mkinitfs, but I suspect something like `kernel/driver/vfio/vfio*` could do
2018-12-03 18:12:26 <SpaceToast> but there's still mdev_vfio, etc
2018-12-03 18:12:39 <pheoxy> mdev_vfio?
2018-12-03 18:50:59 <pheoxy> SpaceToast, so like this https://gist.github.com/Pheoxy/16298df2737f261746a0d5d0bb79a71d
2018-12-03 18:51:35 <SpaceToast> erm
2018-12-03 18:51:37 <SpaceToast> ¯\_(ツ)_/¯
2018-12-03 18:51:40 <SpaceToast> why don't you try it and see if it works?
2018-12-03 18:52:11 <pheoxy> SpaceToast, im currently using the computer I would I have to try it on
2018-12-03 18:52:18 <pheoxy> running archlinux
2018-12-03 18:53:44 <pheoxy> was setting it up on usb on my laptop that doesn't support iommu to see if worked and modinfo showed all required modules
2018-12-03 18:54:00 <pheoxy> so reboot and test?
2018-12-03 18:54:04 <SpaceToast> sure~
2018-12-03 18:54:20 <SpaceToast> but again, I'm not the maintainer for mkinitfs, I'm just guessing based on other entries
2018-12-03 18:54:28 <pheoxy> would be quitting irc so wish me luck
2018-12-03 19:07:14 <pheoxy> back
2018-12-03 19:07:36 <SpaceToast> \o
2018-12-03 19:09:06 <pheoxy> it seems to have worked, I'll have to setup a modprobe options thing like this "options vfio-pci ids=10de:13c2,10de:0fbb" or something?
2018-12-03 19:09:40 <SpaceToast> I've never used vfio, so that's not really for me (even the prior wasn't really, but eh)
2018-12-03 19:10:09 <pheoxy> eh, sounding boards help even if you don't really do much
2018-12-03 19:10:42 <pheoxy> wait are modules built into the kernel or are we loading them in with mkinitfs?
2018-12-03 19:11:31 <SpaceToast> erm
2018-12-03 19:11:40 <SpaceToast> if a module was built into the kernel, it wouldn't be a module
2018-12-03 19:12:01 <SpaceToast> mkinitfs's features.d copies the modules from your live system over into the initramfs
2018-12-03 19:12:09 <SpaceToast> so that they can then be used inside of that
2018-12-03 19:13:12 <pheoxy> ok, reading so many wikis was just confusing me thanks for clearing that up
2018-12-03 19:17:09 <SpaceToast> pheoxy: I'm currently working on getting alpine official documentation, and mkinitfs should be an entry in the developer handbook :)
2018-12-03 19:17:21 <SpaceToast> it might take a while to get there though, so right now that's not much help
2018-12-03 19:18:07 <TBB> what's vfio, by the way?
2018-12-03 19:19:32 <pheoxy> TBB, easiest explanation https://wiki.archlinux.org/index.php/PCI_passthrough_via_OVMF#Enabling_IOMMU
2018-12-03 19:20:15 <pheoxy> sorry left heading in just go to the top
2018-12-03 19:20:40 <TBB> no worries, the link alone tells me enough :)
2018-12-03 19:27:44 <pheoxy> can you diskless on x86_64?
2018-12-03 19:28:30 <SpaceToast> yes, but there are lots of "gotchas"
2018-12-03 19:28:40 <SpaceToast> (also diskless = no initramfs customization for you)
2018-12-03 19:28:46 <pheoxy> ah
2018-12-03 19:29:28 <pheoxy> data mode allow initramfs?
2018-12-03 19:29:40 <TBB> definitely yes
2018-12-03 19:29:57 <pheoxy> why would diskless have no mkinitfs?
2018-12-03 19:29:59 <SpaceToast> depends what you mean by data mode
2018-12-03 19:30:15 <SpaceToast> if you mean what they call that in the wiki, then no, you can't have your own either
2018-12-03 19:30:30 <SpaceToast> the reason is that your initramfs is like your kernel - it's loaded by the bootloader
2018-12-03 19:30:38 <SpaceToast> the bootloader, and initrd, and kernel are all on your readonly medium
2018-12-03 19:30:39 <pheoxy> I aim to run alpine from a usb in readonly and mount to ram on boot
2018-12-03 19:30:43 <SpaceToast> yeah
2018-12-03 19:30:49 <SpaceToast> when you run alpine from a usb, where is your initramfs?
2018-12-03 19:30:58 <SpaceToast> (hint: it's on the usb)
2018-12-03 19:31:11 <SpaceToast> now you could obviously make your own custom usb with a custom initramfs, but that's not quite the same thing, is it?
2018-12-03 19:31:41 <pheoxy> so "lbu commit -d" after "mkinitfs" wouldn't solve it?
2018-12-03 19:32:04 <SpaceToast> nope
2018-12-03 19:32:10 <SpaceToast> first off, suddenly you're backing up /boot (??)
2018-12-03 19:32:18 <SpaceToast> but that's not what the bootloader will look at
2018-12-03 19:32:32 <SpaceToast> usually, lbu stuff happens inside the initramfs, so you end up with a chicken and egg problem
2018-12-03 19:32:37 <SpaceToast> (and in this case, the answer is that the usb comes first)
2018-12-03 19:32:47 <pheoxy> so diskless doesn't create a boot partition?
2018-12-03 19:32:54 <SpaceToast> you're booting off the usb
2018-12-03 19:33:00 <SpaceToast> that is the whole point, as you said
2018-12-03 19:33:03 <pheoxy> yeh
2018-12-03 19:33:11 <pheoxy> im using sys mode at the moment
2018-12-03 19:33:13 <SpaceToast> the usb must either magically know about your /boot partition, despite being generic to everyone, on the bootloader level
2018-12-03 19:33:19 <SpaceToast> (which it doesn't)
2018-12-03 19:33:24 <SpaceToast> or it will load the bundled initrd
2018-12-03 19:33:27 <SpaceToast> (which it does)
2018-12-03 19:33:38 <SpaceToast> once it loads the bundled initrd, whatever you have on your disk doesn't matter
2018-12-03 19:34:57 <pheoxy> i think I'm missing something underlying here, you can't reboot after 'mkinitfs', 'lbu commit -d' is what your saying?
2018-12-03 19:35:12 <pheoxy> it just keeps the old initramfs?
2018-12-03 19:35:54 <SpaceToast> I'm saying that even if you modified lbu to backup your initramfs it wouldn't make any difference, ever.
2018-12-03 19:36:12 <SpaceToast> but instead of just telling you to look into how booting works on linux, let's go over it step by step then
2018-12-03 19:36:18 <SpaceToast> you plug in your usb, and press the power button
2018-12-03 19:36:20 <SpaceToast> what happens?
2018-12-03 19:37:09 <pheoxy> hardware init?
2018-12-03 19:37:33 <SpaceToast> well you get POST, sure
2018-12-03 19:37:45 <SpaceToast> then, the bios decides what to exec directly
2018-12-03 19:38:02 <SpaceToast> it finds that your usb has an MBR (or that there's a direct EFI setup), and will run the bootloader
2018-12-03 19:38:20 <pheoxy> so once bootloader detects ro it wouldn't be able to change initramfs?
2018-12-03 19:38:27 <SpaceToast> sorry what?
2018-12-03 19:38:32 <SpaceToast> the bootloader doesn't detect ro
2018-12-03 19:38:42 <SpaceToast> the bootloader's job is to exec into a kernel, and read an initramfs
2018-12-03 19:38:51 <SpaceToast> it doesn't "change" anything, that's not its job
2018-12-03 19:39:01 <SpaceToast> the bootloader also has to be told where the kernel/initramfs is
2018-12-03 19:39:15 <SpaceToast> since this is the same bootloader on all alpine isos, it will only know about the kernel/initramfs on the usb stick
2018-12-03 19:39:20 <SpaceToast> so it enters the usb stick initramfs
2018-12-03 19:39:44 <SpaceToast> the initramfs finds your apkovl, and applies that on top of the root filesystem (which is also the generic one that comes with the iso)
2018-12-03 19:40:01 <SpaceToast> there's a pretty file called initramfs-${version} in there, but who's going to run it?
2018-12-03 19:40:05 <SpaceToast> why would it affect anything?
2018-12-03 19:45:15 <pheoxy> so your saying the initramfs on the usb stick for diskless mode never changes and then just applies apkvol on top of it for all packages and changes?
2018-12-03 19:45:28 <pheoxy> or do I need to reread something?
2018-12-03 19:45:29 <SpaceToast> on top of your rootfs*
2018-12-03 19:45:56 <SpaceToast> basically, editing the initramfs when booting in diskless mode requires either the bootloader to do things it does not do
2018-12-03 19:46:01 <SpaceToast> or for you to make your own bootable usb stick
2018-12-03 19:48:46 <pheoxy> is that why when I tested diskless mode on a raspberry pi the kernel always stayed on 4.14.69 and wouldnt upgrade? or was that because I never install linux-vanilla on it. Sorry new alpine and that happened a while ago so might be a separate thing
2018-12-03 19:49:13 <SpaceToast> this isn't an alpine thing, this is a computers thing
2018-12-03 19:49:16 <SpaceToast> and yes, that would be why
2018-12-03 19:49:36 <SpaceToast> your kernel is the first thing that gets executed, and the thing that decides which one to run is your boot medium, and your boot medium has one bundled
2018-12-03 19:49:38 <SpaceToast> (and is generic)
2018-12-03 19:49:55 <SpaceToast> if you want to upgrade your kernel, you just re-flash the boot medium
2018-12-03 19:50:01 <SpaceToast> (that's kind of the whole point of diskless mode)
2018-12-03 19:50:46 <pheoxy> ah
2018-12-03 19:51:18 <pheoxy> so as you were saying I would just need to create a custom usb in diskless mode with my own customized initramfs?
2018-12-03 19:51:47 <SpaceToast> well you don't have to create the custom usb in diskless mode, you can make it in any environment you'd like
2018-12-03 19:51:58 <SpaceToast> but... it's easier to just use sys mode at that point, isn't it?
2018-12-03 19:52:13 <pheoxy> wear and tear on the usb
2018-12-03 19:52:36 <pheoxy> also i'm just trying to see if i can do it
2018-12-03 19:53:14 <pheoxy> also I'm planning on running docker containers on it with zfs
2018-12-03 19:53:21 <pheoxy> so theres that
2018-12-03 19:53:42 <pheoxy> I was going to use zfs on seperate hdd with luks encryption
2018-12-03 19:54:29 <pheoxy> boot usb, mount in ram, unlock luks, mount zfs pool, start docker, boot up vms
2018-12-03 19:54:46 <pheoxy> although maybe docker in its own vm
2018-12-03 19:55:18 <pheoxy> that way I can just use straight alpine to startup everything and keep the underlying system clean
2018-12-03 19:55:34 <pheoxy> i have a tendacy to break distros experimenting
2018-12-03 19:56:05 <SpaceToast> your use-case is very confusing to me
2018-12-03 19:56:14 <pheoxy> yeh
2018-12-03 19:56:14 <SpaceToast> because I do not see the reasoning
2018-12-03 19:56:34 <SpaceToast> what's wrong with installing alpine on the disk, having a separate partition for zfs, and everything else being identical?
2018-12-03 19:56:43 <SpaceToast> you can even mount / ro if you'd like
2018-12-03 19:56:52 <pheoxy> I currently use arch for it but I break it and running it in a vm would solve alot of hassel
2018-12-03 19:57:05 <pheoxy> but yeh
2018-12-03 19:57:11 <pheoxy> would that reduce writes?
2018-12-03 19:57:19 <SpaceToast> not to the zfs pool
2018-12-03 19:57:29 <SpaceToast> but mounting / ro means no writes can happen unless you remount it rw...
2018-12-03 19:57:41 <pheoxy> zfs is fine it was going to be recieving alot of traffic anyway
2018-12-03 19:58:05 <SpaceToast> another thing that confuses me is the part where you're talking about what's typically very high end hardware use-cases (lots of VM and container workloads)
2018-12-03 19:58:11 <SpaceToast> but then you worry about things like write cycles
2018-12-03 19:58:25 <SpaceToast> instead of having a bunch of raid thrown in with regular SMART
2018-12-03 19:59:12 <pheoxy> I use to run centos off a usb as hardrive limit is 8 sata ports all used for zfs pool
2018-12-03 19:59:29 <pheoxy> went through 3 usbs in year
2018-12-03 19:59:32 <TBB> I can't help still being skeptical about zfs, don't even exactly know why
2018-12-03 20:00:05 <SpaceToast> TBB: it's very un-linuxy
2018-12-03 20:00:12 <pheoxy> i was hoping btrfs would get raid 5/6 working but...
2018-12-03 20:00:15 <SpaceToast> which makes sense, it comes from an entirely different infrastructure
2018-12-03 20:00:21 <SpaceToast> I just run btrfs in raid1
2018-12-03 20:00:40 <SpaceToast> (which isn't mdadm raid1, it means that every block is guaranteed to have at least one duplicate on a different physical device, so it's /2, not /n)
2018-12-03 20:00:47 <TBB> on Solaris or FreeBSD, sure, why not as they don't have anything better anyway, but on Linux... still trusting xfs + lvm + cryptsetup
2018-12-03 20:01:26 <pheoxy> current hardware 8 sata ports with max of 2TB in each harddrive
2018-12-03 20:01:55 <pheoxy> and I nearly fill those
2018-12-03 20:02:34 <pheoxy> in raidz2
2018-12-03 20:03:00 <SpaceToast> also, you can always just put / on a zvol
2018-12-03 20:03:09 <pheoxy> also replacing bad hdd is a breeze
2018-12-03 20:04:04 <pheoxy> and use usb as boot still?
2018-12-03 20:04:20 <pheoxy> how would that affect current zfs pool though?
2018-12-03 20:04:52 <pheoxy> and would general performance drop?
2018-12-03 20:05:58 <pheoxy> to change my current zfs pool I would need to burrow a external hdd to backup data probably. current backup is only important stuff
2018-12-03 20:05:59 <SpaceToast> no...
2018-12-03 20:06:18 <pheoxy> or would I just add a ROOT dataset?
2018-12-03 20:06:21 <SpaceToast> also I'm not very familiar with zfs
2018-12-03 20:06:27 <SpaceToast> with btrfs this would be relatively trivial though
2018-12-03 20:10:43 <pheoxy> so if just ran in sys mode with a usb as the boot drive and root in zfs it would be 10 times easier?
2018-12-03 20:10:51 <SpaceToast> ????
2018-12-03 20:11:07 <SpaceToast> I mean I guess you could just mount your usb as /boot
2018-12-03 20:11:15 <SpaceToast> that would work just fine, and definitely be easier than making your own
2018-12-03 20:11:18 <SpaceToast> but I meant /boot on zfs as well
2018-12-03 20:11:48 <pheoxy> i suppose it save s screwing around with lbu and figuring out how to encrypt root on usb and unlock from ssh
2018-12-03 20:12:50 <pheoxy> would I have to change current partition and disk scheme with /boot on zfs?
2018-12-03 20:13:17 <pheoxy> all drives are disk encrypted
2018-12-03 20:14:12 <pheoxy> i guess I could purposely fail a drive and use that i think? then just add it back to the pool
2018-12-03 20:14:15 <SpaceToast> Again, not too familiar with zfs
2018-12-03 20:14:21 <SpaceToast> All I can say is it'd work fine on btrfs
2018-12-03 20:14:26 <pheoxy> sorry thinking out loud
2018-12-03 20:14:34 <SpaceToast> (not sure re: luks, again, haven't used that)
2018-12-03 20:14:55 <pheoxy> whats btrfs like with drives failing in a raid configuration?
2018-12-03 20:17:11 <SpaceToast> depends on the raid level
2018-12-03 20:17:23 <pheoxy> hmm you've got me interest in lvm raid again as well
2018-12-03 20:17:25 <SpaceToast> on raid1, it'll complain very loudly, and refuse to boot unless it's in degraded mode
2018-12-03 20:17:32 <SpaceToast> but replacing a drive is basically
2018-12-03 20:17:42 <pheoxy> does it bring the whole raid array down on one disk failure?
2018-12-03 20:17:51 <SpaceToast> no, it becomes degraded
2018-12-03 20:18:17 <pheoxy> does it also work on top of luks?
2018-12-03 20:19:05 <SpaceToast> actually, there's btrfs replace
2018-12-03 20:19:11 <SpaceToast> it should, but I don't use luks
2018-12-03 20:19:52 <TBB> I haven't had such a setup, but it should
2018-12-03 20:21:34 <pheoxy> apprently raidz2/raid56 on btrfs has huge data loss bugs
2018-12-03 20:24:12 <TBB> the impression I have is btrfs has failed to fulfill the expectations, but of course to make any statements I should have experience in using btrfs, which I don't
2018-12-03 20:24:51 <pheoxy> what do you use for raid?
2018-12-03 20:34:34 <TBB> RAID rule of thumb on Linux: always, always mdadm
2018-12-03 20:34:56 <TBB> unless you do really big storage and have an expensive support contract, that is
2018-12-03 21:05:41 <pheoxy> TBB, ok I've done some reading, it seems I'll have stick to zfs as only btrfs has copy on write and bcachefs is too new for me to test with my data. mdadm also doesn't have CoW which I need with my current setup
2018-12-03 21:05:47 <pheoxy> thanks for the suggestions though
2018-12-03 21:06:43 <pheoxy> I'll look back into mdadm when I can get a UPS for my setup
2018-12-03 21:09:29 <pheoxy> g2g see you around and thanks for the convo
2018-12-03 22:01:55 <geosmin> hm. alpine won't seem to boot on the new raspberry pi A+
2018-12-03 22:02:04 <geosmin> AFAIK it's the same SoC as the B+, just with less IO
2018-12-03 23:10:48 <danieli> geosmin: it was just released, I'm not sure what you're expecting here?
2018-12-04 03:42:57 <geosmin> danieli: i assumed it was the same hardware with less IO
2018-12-04 03:44:36 <danieli> seems it isn't
2018-12-04 03:48:44 <geosmin> indeed
2018-12-04 03:49:16 <geosmin> i'll look at raspbian's modules tomorrow and see if i can find any new ones
2018-12-04 03:49:34 <geosmin> is it also likely to need a new kernel or just the modules?
2018-12-04 06:06:40 <geosmin> so the 3 A+ has the same Cortex-A53 and broadcom wireless chip, any ideas as to why it wouldn't work just fine with the modules from the 3 B+?
2018-12-04 06:32:13 <mepholic> https://github.com/intel/nemu
2018-12-04 06:32:21 <mepholic> look at this terrible attempt at EEE by intel
2018-12-04 06:32:35 <mepholic> ~JUST EXTEND ACTUAL QEMU~
2018-12-04 06:59:53 <Michcioperz> I think their presented rationale is that they're trimming rather than extending
2018-12-04 09:02:16 <Greyztar> is there a way to get serial console to work with rpi without booting up through other means first?
2018-12-04 09:04:27 <jn__> you can (in principle) edit the command line offline. it's somewhere in the firmware partition
2018-12-04 09:04:56 <jn__> (the kernel command line)
2018-12-04 09:06:29 <danieli> config.txt or something, iirc?
2018-12-04 09:06:37 <Greyztar> im not too savvy,you mean the cmdline.txt?
2018-12-04 09:06:45 <danieli> if it's there, probably
2018-12-04 09:07:21 <Greyztar> says console=tty1 what should i replace it with?
2018-12-04 09:08:24 <danieli> something along these lines https://elinux.org/RPi_Serial_Connection#Method_3.2C_manual_configuration_.28complex.29
2018-12-04 09:08:27 <jn__> console=ttyS0 i guess
2018-12-04 09:08:47 <danieli> it's easier to just google this stuff unless someone here knows the raspberry pi config off the top of their head
2018-12-04 09:08:49 <jn__> oh, see link above
2018-12-04 09:09:06 <Greyztar> thanks
2018-12-04 09:10:01 <danieli> basically console=/dev/$tty,$baudrate
2018-12-04 09:11:31 <Greyztar> danieli: cheers thanks alot for the help (,")
2018-12-04 12:56:18 <hiro> good day
2018-12-04 12:56:30 <_ikke_> hi
2018-12-04 12:56:33 <hiro> what magic is required to make this work: apk info -r
2018-12-04 12:56:59 <hiro> example issue:
2018-12-04 12:57:00 <hiro> apk info -R pidgin-otr|grep otr
2018-12-04 12:57:07 <hiro> pidgin-otr-4.0.2-r0 depends on:
2018-12-04 12:57:07 <hiro> so:libotr.so.5
2018-12-04 12:57:16 <hiro> but reverse on libotr doesn't show pidgin-otr
2018-12-04 12:58:22 <danieli> libotr is provided by libotr-dev / libotr3-dev
2018-12-04 12:58:34 <danieli> s/libotr is/libotr.so/
2018-12-04 12:58:48 <hiro> when i do apk info -P libotr, it does show libotr.so.5
2018-12-04 12:58:57 <SpaceToast> hiro: do you have both installed?
2018-12-04 12:59:01 <hiro> SpaceToast: no
2018-12-04 12:59:04 <SpaceToast> that would do it
2018-12-04 12:59:11 <hiro> ah!
2018-12-04 12:59:12 <hiro> thanks
2018-12-04 12:59:17 <SpaceToast> my understanding is that -R (depends) uses the package index
2018-12-04 12:59:25 <SpaceToast> while -r (reverse depends) uses your running system information
2018-12-04 12:59:35 <hiro> that would explain it indeed
2018-12-04 13:02:04 <hiro> hmm, apk search has no -P (provides)
2018-12-04 13:02:21 <hiro> what's the easiest way to find the package that provides the .so that packages depend on?
2018-12-04 13:02:35 <hiro> like when i do apk info -R and get like so:libotr.so.5
2018-12-04 13:04:11 <SpaceToast> hiro: try `apk search libotr.so`
2018-12-04 13:04:13 <hiro> ah, search automatically also considers package contents for search input?
2018-12-04 13:04:19 <SpaceToast> I believe so, yes
2018-12-04 13:04:27 <SpaceToast> I haven't looked into the sources, but that seems to be the behavior
2018-12-04 13:04:41 <hiro> good default, i rarely saw that anywhere
2018-12-04 13:04:50 <SpaceToast> it's quite useful, yes
2018-12-04 13:04:59 <SpaceToast> you can also look over on http://pkgs.alpinelinux.org which offers further ways to search
2018-12-04 13:05:01 <hiro> i was looking for the 5th time over apk search --help output to find a hidden option for it ;)
2018-12-04 13:05:28 <hiro> my browser is fucked right now - as usual
2018-12-04 13:05:40 <hiro> that's why i'm "wasting" time with alpine secretly
2018-12-04 13:06:33 <hiro> yeah, i verified and apk info -r works for installed packages
2018-12-04 13:07:11 <hiro> might be worth noting in the --help bec. the asymmetry between -r and -R might confuse others than just me
2018-12-04 13:07:24 <SpaceToast> hiro: we're reviving the official documentation project
2018-12-04 13:07:36 <SpaceToast> it'll take a while for me to get to the "manual pages and stuff like that" section,
2018-12-04 13:07:37 <hiro> "List all installed packages depending on PACKAGE"
2018-12-04 13:07:40 <SpaceToast> but I'll try and keep it in mind
2018-12-04 13:07:52 <hiro> SpaceToast: thanks!
2018-12-04 13:08:05 <SpaceToast> I'll also keep the particularities of search in mind :)
2018-12-04 13:33:47 <mpmc> SpaceToast: Do you come with Jam or Peanut butter?
2018-12-04 13:34:01 <SpaceToast> only with the perfect amount of toasty UV rays
2018-12-04 13:37:55 <mps> hiro: apk add apk-file
2018-12-04 13:38:17 <mps> Command line front end for pkgs.alpinelinux.org
2018-12-04 14:56:11 <hiro> so when does stuff go from /testing into /edge ?
2018-12-04 14:56:28 <hiro> or i guess /community (wrong level)
2018-12-04 14:58:04 <_ikke_> hiro: mostly up to the maintainer of that packae
2018-12-04 14:58:06 <_ikke_> package
2018-12-04 14:58:29 <_ikke_> but it's at least expected that the package is actually tested to work properly
2018-12-04 14:58:40 <_ikke_> (so not that it just builds)
2018-12-04 14:58:46 <TBB> I think that's something that could use a better process
2018-12-04 14:59:34 <TBB> something like, say, people somehow signaling the project the package has been tested by them, and with enough confirmations the package gets a promotion... how would that sound?
2018-12-04 15:00:11 <SpaceToast> TBB: being in community (is at least supposed to) implies a bit more than just "it works" too
2018-12-04 15:00:20 <SpaceToast> technically, community packages are supposed to have official support for ~6 months
2018-12-04 15:00:34 <SpaceToast> (or was that main?, either way, a period of time)
2018-12-04 15:00:51 <_ikke_> main is even longer afaik
2018-12-04 15:00:56 <SpaceToast> yeah, iirc main is 2 years
2018-12-04 15:01:08 <SpaceToast> so you really don't want the maintainers disappearing after getting it in, you know? ^^;;
2018-12-04 15:01:15 <SpaceToast> (it still happens though, clearly)
2018-12-04 15:01:21 <_ikke_> right
2018-12-04 15:13:11 <TBB> Yeah, that's totally understandable, SpaceToast. So maybe that part of the process could be tweaked a bit too... perhaps by mailing the maintainer when the 6 month period is closing its end or so, possibly even offering the package for other maintainers in case the original one doesn't want to continue? (I admit I'm not aware of what the process is currently)
2018-12-04 15:13:29 <SpaceToast> TBB: I don't think there is a process atm ^^;;
2018-12-04 15:13:53 <_ikke_> There is a process, but it's not automated / formalized
2018-12-04 15:13:57 <SpaceToast> the way I'd do it is require testing/ packages to remain regularly updated for the usual time period (6 months) before they're allowed into community
2018-12-04 15:14:23 <_ikke_> Some packages do not get any upstream updates though
2018-12-04 15:14:30 <SpaceToast> if the maintainer has stuck around in testing/ for 6 months, and claims they're willing to keep doing so, and the package has been actually used for that long, it might make sense to throw it in
2018-12-04 15:15:02 <SpaceToast> _ikke_: well, if no updates have happened, and the maintainer is still there; technically they've remained updated to the latest version :)
2018-12-04 15:15:27 <SpaceToast> this is quite long, but it's a consequence of the length of the community/ support cycle, imo
2018-12-04 15:20:00 <clandmeter> the flow from testing to community/main should be properly documented.
2018-12-04 15:20:27 <TBB> for quality purposes some sort of a voting mechanism would be good, but naturally if upstream makes no changes or so then bothering the maintainer might be unnecessary... could an upstream update notification be automatized in any sensible way?
2018-12-04 18:00:42 <hiro> is there any downside to adding the /testing repo, or will it be guaranteed that there are no conflicts with other packages with the same name from community or main?
2018-12-04 18:00:55 <hiro> (adding to your local apk repos file)
2018-12-04 18:02:02 <hiro> i also wonder about how other people do this, like if they add /testing bec. many packages are only in there, they will never realize during apk search or apk add that something important they might want to rely on in the future is from /testing and not from main
2018-12-04 18:02:37 <hiro> so that would be about the visibility of this repository in the apk interface
2018-12-04 18:04:06 <hiro> TBB: perhaps you could have a thumbs-down button. people are much more likely to complain than to contribute something more useful.
2018-12-04 18:04:26 <hiro> TBB: so if a package has a lot of thumbs-down it might be good to manually check wtf is wrong with it :)
2018-12-04 18:05:46 <SpaceToast> hiro: and then, next time mozilla do something stupid, firefox gets 20 000 👎 ;)
2018-12-04 18:06:27 <hiro> SpaceToast: so as expected -> case in point i'd say
2018-12-04 18:06:37 <hiro> SpaceToast: also can serve as a WARNING to users who haven't updated yet
2018-12-04 18:06:48 <hiro> SpaceToast: like, it needn't be the maintainer's fault
2018-12-04 18:06:55 <SpaceToast> ah, encouraging using applications up until EOL sounds great~
2018-12-04 18:07:02 <TBB> hiro: ahh, brainstorming :D
2018-12-04 18:07:05 <SpaceToast> "don't run updates! I don't like a political move!"
2018-12-04 18:07:15 <SpaceToast> that's how these things usually go
2018-12-04 18:07:20 <hiro> SpaceToast: you mean creating incentive to keep some LTS versions around?
2018-12-04 18:07:43 <SpaceToast> such as firefox-esr, which is currently 52, and had some controversy as well? ;)
2018-12-04 18:07:53 <hiro> i didnt' know about any controversy
2018-12-04 18:08:04 <hiro> but personally i indeed always only used firefox-esr
2018-12-04 18:08:09 <hiro> s/only //
2018-12-04 18:08:12 <SpaceToast> there's a bunch of outrage with almost every firefox release, I didn't pick mozilla for no reason
2018-12-04 18:08:20 <hiro> right
2018-12-04 18:08:39 <SpaceToast> a whole "review our packages downstream" system seems wrong because it won't be used for what you're intending
2018-12-04 18:08:45 <hiro> i mean it's a browser after all, i could imagine there will be more breakage as most people are forced to use a browser or another nearly every day
2018-12-04 18:08:50 <SpaceToast> even "this package doesn't work for me" button won't do much
2018-12-04 18:09:05 <hiro> SpaceToast: for some big software, sure.
2018-12-04 18:09:06 <SpaceToast> because people will click it even if the problem is BCAK
2018-12-04 18:09:22 <SpaceToast> I'd be ok with a "this issue affects me too" button on bug reports
2018-12-04 18:09:27 <hiro> SpaceToast: but for 90% of the less-used stuff that gets nearly no users, this is a huge benefit over having no feedback at all from any user
2018-12-04 18:10:08 <hiro> like, i maintained some packages (though not on alpine), and i randomly found complaints about it, but i only once in 5 years got a PM from anybody affected by any problem
2018-12-04 18:10:28 <hiro> people like to complain, but they never try to talkk to the person responsible to fix it
2018-12-04 18:11:17 <hiro> if AT LEAST i had an easy way to see how unhappy they all are, i could try to investigate more deeply but also less often (as i get notified)
2018-12-04 18:11:19 <TBB> complaining in general tends to be therapeutic; talking to people in general not :)
2018-12-04 18:11:37 <hiro> haha
2018-12-04 18:12:32 <SpaceToast> erm... indeed, people usually do not (and should not) PM maintainers directly
2018-12-04 18:12:36 <SpaceToast> that's kind of a bad idea in general
2018-12-04 18:12:46 <hiro> yep
2018-12-04 18:12:50 <SpaceToast> you're supposed to open a bug in the bug tracker, explain the issue, and so on
2018-12-04 18:12:56 <hiro> those are the cases i thought one should try to cover
2018-12-04 18:12:58 <SpaceToast> so the fact that you got PMd at all is a problem with the person doing it
2018-12-04 18:13:03 <hiro> cause PM'ing maintainers is *already* possible
2018-12-04 18:13:11 <SpaceToast> and wrong, as most users likely know.
2018-12-04 18:13:22 <hiro> oh wait, i misread you
2018-12-04 18:13:34 <hiro> yeah, in my case PM was the best possible process, it was another project.
2018-12-04 18:13:44 <SpaceToast> so go fix that other project ;)
2018-12-04 18:13:51 <SpaceToast> Alpine has a bug tracker
2018-12-04 18:13:55 <hiro> nah, i don't care so much whether it is called PM or bug report
2018-12-04 18:13:56 <SpaceToast> (two even, because github has one built-in)
2018-12-04 18:14:00 <hiro> i mean whatever mechanism exists for that.
2018-12-04 18:14:35 <SpaceToast> well, alpine currently has roughly 1500 open bug reports
2018-12-04 18:14:42 <hiro> oh shit!
2018-12-04 18:14:48 <hiro> all for packages?
2018-12-04 18:15:06 <SpaceToast> hard to know ¯\_(ツ)_/¯
2018-12-04 18:15:13 <danieli> 1467 for the 'alpine linux' project in our bug tracker
2018-12-04 18:15:19 <SpaceToast> But technically, aren't most things in alpine a package?
2018-12-04 18:15:28 <SpaceToast> if the problem is, say, with apk, the bug regards the apk-tools package
2018-12-04 18:15:35 <hiro> right
2018-12-04 18:15:38 <SpaceToast> danieli: yes, as I said, roughly 1500 ;)
2018-12-04 18:15:46 <danieli> yup
2018-12-04 18:15:47 <hiro> you'd have to find the "core and core packages"
2018-12-04 18:15:59 <SpaceToast> and people still abuse the system as it is
2018-12-04 18:16:00 <mps> roughly speaking half of these bug reports aren't bugs
2018-12-04 18:16:15 <hiro> i see a race condition here, during danieli typing somebody could have opened a new bug report and then his number would be wrong
2018-12-04 18:16:16 <SpaceToast> for instance, one of the newest bug reports is "please update lighttpd to 1.4.52"
2018-12-04 18:16:19 <SpaceToast> it's marked as "urgent"
2018-12-04 18:16:31 <SpaceToast> I somehow suspect it isn't very urgent :)
2018-12-04 18:16:33 <danieli> hiro: and the relevance is?
2018-12-04 18:16:35 <_ikke_> But there is a difference with: "this package isn't packaged properly" and "the application is bugged"
2018-12-04 18:17:02 <SpaceToast> hiro: ultimately, point is, I do not see what advantage a 👎 button would give us, but I see a LOT of disadvantages.
2018-12-04 18:17:03 <hiro> danieli: it's a joke
2018-12-04 18:17:05 <danieli> aha
2018-12-04 18:17:30 <danieli> I'm not in a state of mind to take jokes very easily, been working too much today - my bad
2018-12-04 18:17:43 <hiro> danieli: sorry for that
2018-12-04 18:17:55 <hiro> danieli: get some fresh air, man ;)
2018-12-04 18:18:07 <danieli> anyway, those ~1500 issues cover documentation, different categories of packages, toolchain, security, ...
2018-12-04 18:18:14 <danieli> hiro: I'm about to get some sleep, actually
2018-12-04 18:18:32 <hiro> yeah, i cannot remember your timezones
2018-12-04 18:18:35 <danieli> woke up around 2-3 am, and it is now 7.18 pm
2018-12-04 18:18:35 <SpaceToast> danieli: is -doc not a package type? is gcc not one? is a security not bug a bug that affects the security of a package? and so on :)
2018-12-04 18:18:50 <hiro> danieli: yeah, go for it
2018-12-04 18:18:50 <SpaceToast> that's quite early to wake up, good night \o
2018-12-04 18:19:04 <danieli> SpaceToast: documentation can be more than just -doc, no need to nitpick to such an extreme level
2018-12-04 18:19:11 <hiro> danieli: your slave master will not be happy if your productivity goes down from sleeplessness
2018-12-04 18:19:18 <danieli> hiro: excuse me?
2018-12-04 18:19:50 <hiro> danieli: i mean have a good sleep :)
2018-12-04 18:20:14 <hiro> thought perhaps you need excuses, cause you were still here ;)
2018-12-04 18:20:34 <_ikke_> SpaceToast: a lot of infrastructure is not covered by packages
2018-12-04 18:21:05 <SpaceToast> _ikke_: fair enough, I just like taking things to their logical conclusions
2018-12-04 18:21:06 <danieli> hiro: I already have one, <Esc>:wq and `git commit && git push`
2018-12-04 18:21:11 <danieli> finishing off my stuff
2018-12-04 18:21:17 <_ikke_> SpaceToast: leaky abstractions :-)
2018-12-04 18:21:38 <SpaceToast> if an abstraction is perfect, need it remain an abstraction?~
2018-12-04 18:21:57 <hiro> danieli: hahaha
2018-12-04 18:22:33 <hiro> no || git force-push -> commendable
2018-12-04 18:22:44 <danieli> I'm not tired enough to dig myself into that hole
2018-12-04 18:24:22 <hiro> SpaceToast: haha
2018-12-04 19:37:08 <geosmin> so i tried importing the .dts files in boot from raspbian's latest OS image, A+ still won't boot
2018-12-04 19:37:36 <geosmin> not that that's unexpected, i'm not too sure what to try and can't find many resources
2018-12-04 19:39:14 <geosmin> anyone privy to previous work involved getting alpine up and running on previous new pi revisions?
2018-12-04 19:55:54 <mps> geosmin: .dts or .dtb file. kernel need .dtb (binary) and not .dts (source). Or you trying overlay dt file
2018-12-04 20:07:49 <geosmin> mps: sorry, i actually meant dtb
2018-12-04 20:08:21 <geosmin> i'd love to know /why/ exactly it isn't booting, i'd have more to work with
2018-12-04 20:08:30 <geosmin> as it is now i'm just a jackass with stupid questions
2018-12-04 20:58:40 <badloop> is the apk cdn having issues today?
2018-12-04 21:22:43 <_ikke_> badloop: Not that we're aware of
2018-12-04 21:22:56 <_ikke_> badloop: any specific issues?
2018-12-04 22:09:54 <brebs> fcolista: Hi, I submitted https://patchwork.alpinelinux.org/patch/4135/ (mate-screensaver: enable password lock) a month ago, is there any problem with applying it? Without it, mate-screensaver fails to actually lock the screen.
2018-12-04 22:40:34 <Greyztar> this is one very nice os! Much appreciated (,")
2018-12-05 00:11:20 <kpcyrd> is there anything special about alpine partprobe? fdisk lists my partitions, but I can't get them to show up in /dev
2018-12-05 00:35:09 <kpcyrd> I couldn't figure out if it's the kernels fault or busybox' fdisk, the paritions are there after a reboot
2018-12-05 01:47:42 <SpaceToast> kpcyrd: alpine uses mdev by default, and udev if you install and enable it
2018-12-05 01:48:08 <SpaceToast> sounds weird that fdisk would see them though
2018-12-05 01:48:14 <SpaceToast> next time, try triggering a coldboot and see what happens
2018-12-05 01:48:21 <SpaceToast> (if it helps, file a bug to the appropriate project :) )
2018-12-05 02:47:00 <kpcyrd> SpaceToast: I finished the install and ran into the same issue again (this time after a dd+partprobe)
2018-12-05 02:49:17 <kpcyrd> I happened to have dmesg -w open and noticed partprobe is causing a logline `sdc: sdc1 sdc2` which is also what `fdisk -l` says
2018-12-05 02:56:16 <FunWithFlags> Hey all. I've setup apline in virtualbox. All works fine. I'm trying to symlink a /mtn/outside to /var/www/localhost/htdocs. Host is windows. I've mounted fine, but symlinking messes up permissions. Is it possible?
2018-12-05 08:47:26 <hiro> nobody answered me yesterday if there's downside to adding the /testing repo, as in whether there will be packages with the same name as main or community that would create conflicts...
2018-12-05 08:48:14 <clandmeter> hiro, in general not
2018-12-05 08:48:45 <clandmeter> but that means you need to stay on edge only, not add testing to stable.
2018-12-05 08:49:36 <hiro> so edge/main+edge/community+edge/testing is all guaranteed to be compatible?
2018-12-05 08:49:51 <hiro> but what i describe could happen if you mix anything-edge with a stable release?
2018-12-05 08:50:09 <clandmeter> compatible yes
2018-12-05 08:50:38 <clandmeter> they live together, and the idea is to keep pkg names unique.
2018-12-05 08:50:52 <hiro> right
2018-12-05 08:51:01 <clandmeter> if you want something stable you should never mix branches
2018-12-05 08:52:27 <clandmeter> the repos you add should always be build against the branch you already have defined in repositories.
2018-12-05 08:54:13 <hiro> thank you
2018-12-05 08:54:23 <clandmeter> i hope that makes sense to you :)
2018-12-05 08:54:28 <hiro> yep
2018-12-05 09:37:43 <mps> hiro: I have stable, edge and testing in /etc/apk/repositories on more than one boxes
2018-12-05 09:38:20 <mps> sometimes I make some mess but usually could fix it
2018-12-05 09:38:54 <mps> of course, I tag them appropriately
2018-12-05 13:01:18 <hiro> sure
2018-12-05 13:01:23 <hiro> i'm too dumb for that! ;)
2018-12-05 13:08:25 <danieli> "stable, edge and testing"?
2018-12-05 13:15:05 <mps> danieli: yes ;)
2018-12-05 13:15:26 <danieli> doesn't sound good, but hey, you do you
2018-12-05 13:16:07 <mps> only 24 packages from testing
2018-12-05 13:16:25 <mps> 21 from edge
2018-12-05 13:17:20 <mps> 23 locally made
2018-12-05 13:18:11 <mps> total 440 packages installed
2018-12-05 16:25:55 <geosmin> so it seems the A+ is booting up and connecting to the network through wlan0
2018-12-05 16:26:07 <geosmin> that said i can't seem to get a console up or ssh access
2018-12-05 16:27:19 <AinNero> geosmin: boot console is crucial for debugging
2018-12-05 16:27:27 <AinNero> do you have serial or video adapter?
2018-12-05 16:27:39 <AinNero> for console?
2018-12-05 16:31:03 <geosmin> AinNero: http://ix.io/1viY
2018-12-05 16:32:13 <geosmin> for reference 3.8.1 wasn't booting at all on the A+, but i've since copied over the update bootloader, dtbs, etc. from the official rpi github and gotten this far
2018-12-05 16:33:07 <geosmin> really scratching my head as to why some portions of userland seem to be working and other's aren't
2018-12-05 16:33:39 <AinNero> this might be a fault in code i've written
2018-12-05 16:34:16 <AinNero> could you add 'debug_init' to the kernel command line?
2018-12-05 16:34:27 <AinNero> the initramfs will create quite much of output then
2018-12-05 16:34:27 <geosmin> sure thing
2018-12-05 16:35:08 <geosmin> i assume that's different from start_debug=1?
2018-12-05 16:36:37 <AinNero> i dont know. i want to trigger this code path: https://github.com/alpinelinux/mkinitfs/blob/cb013df1407dc75d033de2ce79aafc4edd510483/initramfs-init.in#L340
2018-12-05 16:37:45 <AinNero> ah wait
2018-12-05 16:38:42 <AinNero> you dont have 'quiet' on the command line, right? you dont see the "Alpine Init $VERSION" output, so the error is earlier..
2018-12-05 16:39:02 <AinNero> or all the output goes to tty1
2018-12-05 16:39:48 <geosmin> nope, quiet's out
2018-12-05 16:40:00 <AinNero> do you have a monitor connected?
2018-12-05 16:40:05 <AinNero> via hdmi?
2018-12-05 16:40:24 <geosmin> something's weird though, it almost seems like output stops for a dozen seconds or so before it fails
2018-12-05 16:40:36 <geosmin> no, but i can try
2018-12-05 16:40:49 <AinNero> i dont see failure
2018-12-05 16:40:59 <geosmin> lots of HDMI errors in that output though, we'll see if it works
2018-12-05 16:41:00 <AinNero> i see the output on a tty that nobody spawns a getty for
2018-12-05 16:41:56 <geosmin> that could be it. i've previously omitted setting up getty since it worked just fine without
2018-12-05 16:42:00 <geosmin> (on the B+)
2018-12-05 16:42:47 <AinNero> is it diskless boot?
2018-12-05 16:43:12 <geosmin> it's pretty much vanilla rpi aarch64
2018-12-05 16:43:13 <AinNero> ah yes, i can tell from the output.
2018-12-05 16:43:45 <AinNero> try removing console=tty1, so you should get the stdout from the init scripts
2018-12-05 16:45:13 <AinNero> when you boot diskless and dont get a getty for a device you specified via console=, its an bug in mkinitfs
2018-12-05 16:45:53 <AinNero> i rewrote that section of the initramfs to avoid these type of bugs, but thats only in edge, yet
2018-12-05 16:48:08 <geosmin> alright, tried HDMI and that was a lot more promising
2018-12-05 16:48:18 <geosmin> it... basically seems to boot fine.
2018-12-05 16:48:22 <AinNero> did you get an getty there?
2018-12-05 16:48:31 <geosmin> no, so there's no login prompt
2018-12-05 16:48:49 <geosmin> but initial installation and subsequent openrc seemed mostly fine
2018-12-05 16:49:02 <geosmin> saw a few errors about no space left on device in regards to some .txt files
2018-12-05 16:49:34 <AinNero> if you remove console=tty1 from the command line, you should see the same output via the serial console
2018-12-05 16:49:40 <AinNero> so you could paste that here
2018-12-05 16:49:55 <AinNero> if getty does not work for tty1, its likely some other system fault
2018-12-05 16:54:36 <geosmin> yup, seeing the same output from HDMI over serial now
2018-12-05 16:55:42 <AinNero> including the error messages im after?
2018-12-05 17:07:52 <geosmin> sorry for the delay, had to step away
2018-12-05 17:08:39 <geosmin> so i do get a login prompt if debug_init is set
2018-12-05 17:10:59 <AinNero> could you paste the error messages from startup?
2018-12-05 17:11:39 <AinNero> if there is not getty at all, there might be worse issues during startup
2018-12-05 17:11:48 <geosmin> http://ix.io/1vje
2018-12-05 17:12:55 <geosmin> init_debug seems to have solved the problem. ssh works fine
2018-12-05 17:14:13 <AinNero> > ERROR: vim-8.1.0115-r0: No space left on device
2018-12-05 17:14:38 <AinNero> it seems like the rootfs in ram runs out of space
2018-12-05 17:15:33 <AinNero> i think your apkovl has too many packages
2018-12-05 17:16:11 <geosmin> possible. IIRC apkvol is at~200MB and the board has 512M
2018-12-05 17:17:05 <AinNero> you are loading python3 and a full build toolchain (gcc) during boot, maybe you dont need that
2018-12-05 17:18:00 <geosmin> gcc no
2018-12-05 17:18:28 <AinNero> you could edit /etc/apk/world and remove things you dont need
2018-12-05 17:18:55 <AinNero> if you dont need gcc, you wont need the *-dev packages either
2018-12-05 17:19:18 <geosmin> ran 'apk del gcc', apk then installed vim. nice touch.
2018-12-05 17:20:22 <AinNero> think of 'apk del' like 'remove the line from /etc/apk/world, then run apk fix'
2018-12-05 17:20:52 <AinNero> that makes behavior like this much less surprising
2018-12-05 17:22:29 <geosmin> make sense, thanks
2018-12-05 17:23:01 <daggs1> Greetings, does kernel 4.18 and above can be found on alpine's repos?
2018-12-05 17:23:25 <AinNero> daggs1: its public searchable at https://pkgs.alpinelinux.org/packages
2018-12-05 17:23:39 <AinNero> answer is 'no'
2018-12-05 17:24:21 <daggs1> AinNero, I've asked because I didn't found, there is a 4.18 header so I've assumed I'm missing something
2018-12-05 17:25:45 <AinNero> daggs1: aports: main/linux-vanilla is at 4.14.85, so im not sure where you have seen 4.18
2018-12-05 17:26:21 <daggs1> AinNero, https://pkgs.alpinelinux.org/package/edge/main/x86_64/linux-headers
2018-12-05 17:27:05 <geosmin> AinNero: i assume debug_init making the difference between a succesful and anunsuccesful boot might be a bug?
2018-12-05 17:27:49 <AinNero> geosmin: debug_init only increases output verbosity
2018-12-05 17:28:02 <AinNero> the out-of-space errors are the elephant in the room
2018-12-05 17:28:27 <geosmin> only other change before it worked was removing console=tty1
2018-12-05 17:28:48 <AinNero> your system did still not boot successfully
2018-12-05 17:29:12 <geosmin> i've since added that bactrimmed apkvol down to 130M (from 200) and we're getting an unsuccesful boot again
2018-12-05 17:29:44 <geosmin> added console=tty1*
2018-12-05 17:29:56 <geosmin> and trimmed*
2018-12-05 17:32:58 <AinNero> does it boot without out-of-space errors?
2018-12-05 17:33:25 <geosmin> yeah, i was able to log in after the vim-related one
2018-12-05 17:34:00 <AinNero> didn't you just say you did not get an getty?
2018-12-05 17:35:39 <geosmin> after removing console=tty1 and adding debug_init i had the out of space error but was able to log in
2018-12-05 17:36:32 <AinNero> the correct solution would be using an root partition
2018-12-05 17:37:01 <geosmin> could also be related to gpu_mem, not sure what the default is
2018-12-05 17:37:29 <AinNero> operating under RAM exhaustion is definitely not supported.
2018-12-05 17:40:27 <AinNero> just for reference, i never had apkovl larger than a megabyte
2018-12-05 17:41:34 <AinNero> daggs1: there are newly introduced symbols
2018-12-05 17:41:59 <AinNero> it might be that 4.18 was added so compiled packages wont fail at build-time because of it
2018-12-05 17:44:42 <daggs1> AinNero, so 4.18 is in the pipeline I guess?
2018-12-05 17:45:12 <AinNero> i dont think so
2018-12-05 17:46:11 <AinNero> 4.14.85 is up-to-date, maybe the 3.10 will get 4.18
2018-12-05 17:46:48 <AinNero> oh, no, it wont, 4.18 is EOL
2018-12-05 17:47:20 <mps> 4.19 probably is next AL kernel
2018-12-05 17:47:40 <daggs1> AinNero, I don't get it, 4.19 is latest, 4.18 is newer than 4.14.x
2018-12-05 17:47:57 <AinNero> daggs1: you know of the branches?
2018-12-05 17:48:03 <AinNero> https://www.kernel.org/
2018-12-05 17:48:22 <AinNero> 4.18 is 10 days 'fresher' than 4.19
2018-12-05 17:48:39 <_ikke_> Wasn't 4.19 the next lts?
2018-12-05 17:49:02 <AinNero> _ikke_: probably, but 4.18 is EOL, so i was wrong above
2018-12-05 17:49:09 <mps> _ikke_: yes, because that AL will probably use it
2018-12-05 17:49:17 <_ikke_> yes
2018-12-05 17:49:22 <daggs1> AinNero, I'm new to alpine, I'm considering it as a os for my dvb server over a vm
2018-12-05 17:49:26 <_ikke_> ncopa already stated that
2018-12-05 17:49:56 <AinNero> daggs1: thats kernel versioning is for all linux distros
2018-12-05 17:50:04 <daggs1> but my usb dungle requires atlest kernel 4.17, currently I use debian with 4.18
2018-12-05 17:50:25 <_ikke_> daggs1: You probably have to wait for AL 3.9 then
2018-12-05 17:50:38 <daggs1> _ikke_, eta?
2018-12-05 17:50:44 <_ikke_> ASAP(tm)
2018-12-05 17:50:45 <mps> daggs1: I use AL with 4.18 for months on arm servers
2018-12-05 17:51:23 <daggs1> mps, custom kernel building?
2018-12-05 17:51:25 <_ikke_> mps: did you build it yourself/
2018-12-05 17:51:37 <AinNero> mps: not sure if that was just boasting
2018-12-05 17:51:52 <mpmc> 
ACTION hopes the build servers for armhf are working for 3.9.

2018-12-05 17:51:53 <mps> yes, custom for some hardware tweaks
2018-12-05 17:52:49 <mps> now, I'm just trying to build 4.19 for armv7 AL
2018-12-05 17:52:53 <AinNero> i could also say 'i run alpine with 4.14 with grsec patches' but thats also not helpful here
2018-12-05 17:53:42 <mps> AinNero: right, but it would take time to explain every change/tweak
2018-12-05 17:54:27 <AinNero> why explain?
2018-12-05 17:55:06 <mps> heh, then probably I misunderstood your comment
2018-12-05 17:55:30 <AinNero> i mean like, i could boast with it, but its not helping anyone here since the grsec patches are inaccessible to most of the users
2018-12-05 17:55:37 <AinNero> and most users are not able to build custom kernels
2018-12-05 17:55:39 <geosmin> AinNero: ah, so it seems the wrinkle was that a program i wrote is taking an absurdly long time to start, hanging the boot process
2018-12-05 17:55:44 <mps> just wanted to say that the AL works fine with 4.18, nothing more
2018-12-05 17:56:03 <geosmin> when we tried debug_init, etc.i'd happened to step away
2018-12-05 17:56:25 <geosmin> looking into the hang it seems urandom is the culprit
2018-12-05 17:57:03 <geosmin> many minutes ot generate the needed entropy, a process which was more or less instantaneous on the B+
2018-12-05 19:10:00 <joneskoo> I'm trying to figure out https://wiki.alpinelinux.org/wiki/Raspberry_Pi ; I want to create a headless SD card that will just launch my server. How can I prepare such an image without booting it on an actual raspberry?
2018-12-05 19:11:51 <joneskoo> I tried to run the alpine rpi distribution with qemu but I failed, just got a black screen. I'd like to try it with qemu since I only have one rpi and I have to bring down my service if I need to do things on an actual pi.
2018-12-05 19:12:20 <joneskoo> My goal is to be able to share instructions to re-create my SD card, without actually having to manually build it with an actual Pi.
2018-12-05 19:32:15 <joevandyk> Hi! I'm trying to get an HP printer working on a raspberry pi, but https://pkgs.alpinelinux.org/packages?name=hplip&branch=edge lists only packages for x86. What's the best way to go around getting that package available for arm?
2018-12-05 19:34:23 <danieli> joevandyk: seems like it's because sane is missing on some architectures
2018-12-05 19:34:58 <danieli> i have no clue why that is, the git repo for it doesn't say, for whatever reason
2018-12-05 19:35:42 <joevandyk> @daniel, sorry, i'm new. what is "sane"?
2018-12-05 19:36:56 <joevandyk> ah, the scanner support.
2018-12-05 19:37:18 <danieli> oh, sorry, you didn't highlight me properly so I didn't see your message
2018-12-05 19:37:22 <joevandyk> is it possible for hplip to not require sane? i just want to print, not scan :)
2018-12-05 19:37:31 <danieli> yes, "Scanner Access Now Easy" apparently
2018-12-05 19:37:33 <joevandyk> danieli: sorry, also new to irc :)
2018-12-05 19:37:55 <danieli> figured as much - protip, don't prefix usernames with "@" (it means channel operator/admin on IRC), and use tab completion :)
2018-12-05 19:38:02 <joevandyk> slack habit :(
2018-12-05 19:38:11 <danieli> good thing this isn't slack :P
2018-12-05 19:38:23 <danieli> anyway, you could try building it without sane yourself, considering the state of printer software, I doubt it'd work
2018-12-05 19:45:13 <joevandyk> danieli: the newest hplip don't compile on arm -- ones from six months ago do though. i was hoping someone else had fixed it. i was just downloading the hplip source code and installing manually. is there a better way to do that on alpine? never built an alpine package before.
2018-12-05 19:45:42 <danieli> joevandyk: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2018-12-05 19:45:50 <joevandyk> danieli: is there an owner of that hplip package? or arm building?
2018-12-05 19:46:32 <SpaceToast> joevandyk: you can look up packages over on https://pkgs.alpinelinux.org/, it'll tell you who the maintainer is
2018-12-05 19:46:34 <danieli> looks like there's no arm build (only x86 and x86_64), and yes, it does have an owner - although you should create an issue if there is an actual problem with the package
2018-12-05 19:47:45 <danieli> joevandyk: but as I said, you could try building it without sane support yourself
2018-12-05 19:52:51 <mps> joevandyk: month or two ago I built hplip (and what not) on armhf for someone who asked me on this channel to do that
2018-12-05 19:53:30 <danieli> PSA: be careful with trusting packages from third parties
2018-12-05 19:53:52 <mps> danieli: of course
2018-12-05 19:53:56 <joevandyk> does the apk build process support conditional dependencies, depending on architecture?
2018-12-05 19:54:55 <danieli> we don't really want an arm version of hplip in the official alpine repositories if it's half broken
2018-12-05 19:55:00 <danieli> if that's what you're thinking of
2018-12-05 19:55:03 <mps> danieli: I have to look if I didn't deleted all needed patches and can send APKBUILD file for someone to build it yourself
2018-12-05 19:55:23 <danieli> mps: tell joevandyk, not me :) I don't need hplip
2018-12-05 19:55:33 <joevandyk> mps: that would be awesome if you could, thanks. :)
2018-12-05 19:56:32 <joneskoo> so no raspberry users around? https://wiki.alpinelinux.org/wiki/Raspberry_Pi wondering if there's a way to prepare the SD card without running on an actual raspberry
2018-12-05 19:57:06 <mps> will try, but right now this machine is in phase of upgrade (kernel and AL to armv7)
2018-12-05 19:57:36 <joevandyk> danieli: to try to build my own package, i should checkout the aport git repo and modify the APKBUILD file for hplip?
2018-12-05 19:57:42 <SpaceToast> joneskoo: I am confused - the very link you point to tells you how to prepare the sdcard on "your workstation" (aka not the raspberry)
2018-12-05 19:57:53 <danieli> joevandyk: you don't need all of aports to build just one package
2018-12-05 19:58:02 <SpaceToast> sorry, joevandyk*
2018-12-05 19:58:03 <joneskoo> SpaceToast: yes but not how to include my custom binary on it to start headless
2018-12-05 19:58:32 <SpaceToast> have you tried
2018-12-05 19:58:37 <SpaceToast> putting your custom binary in there
2018-12-05 19:58:42 <SpaceToast> writing a service file
2018-12-05 19:58:45 <SpaceToast> and adding it to the runlevel?
2018-12-05 19:59:15 <danieli> SpaceToast: I think your nick highlight correction was a bit off there, buddy?
2018-12-05 19:59:28 <joneskoo> I don't see how to do that. the alpine-rpi-<version>-armhf.rpi.tar.gz includes directories "apks", "boot", "overlays", none of which sounds like something I could put stuff into
2018-12-05 19:59:59 <SpaceToast> danieli: I get confused when the first two letters are different, because that's how I normally differentiate nicks ^^;;
2018-12-05 20:00:26 <SpaceToast> oh right, armhf wants diskless mode
2018-12-05 20:00:32 <SpaceToast> I believe it does apkovl stuff, so you'd add it to that
2018-12-05 20:00:38 <danieli> yup
2018-12-05 20:00:40 <danieli> sounds about right
2018-12-05 20:00:45 <mps> joevandyk: here is the APKBUILD: http://tpaste.us/vbJO
2018-12-05 20:00:52 <joneskoo> diskless is what I want, yes
2018-12-05 20:01:11 <danieli> mps: but wasn't the issue that sane doesn't build on non-x86?
2018-12-05 20:01:53 <mps> and here is the needed patch: http://tpaste.us/Q1nQ
2018-12-05 20:02:09 <mps> danieli: this one I've built on armhf
2018-12-05 20:02:19 <mps> with these two files
2018-12-05 20:03:37 <mps> I must tell that I didn't tried it in work, I forgot the nickname of the man who asked me to make it and upload it
2018-12-05 20:03:55 <mps> but he told me that he downloaded it and tried
2018-12-05 20:04:09 <mps> I don't know if it worked
2018-12-05 20:04:42 <danieli> I remember when someone joined and talked about it, but I don't remember any of the details whatsoever
2018-12-05 20:05:34 <joneskoo> so is there a way to prepare the diskless overlay for raspberry without running on a raspberry (to enable building it on a different architecture)
2018-12-05 20:05:43 <mps> there are #alpine-linux logs somewhere on alpine site, could be searched probably
2018-12-05 20:06:12 <mps> I don't have logs locally
2018-12-05 20:06:16 <_ikke_> I do
2018-12-05 20:06:41 <mps> danieli: you have really good memory
2018-12-05 20:07:27 <_ikke_> What should I look for?
2018-12-05 20:07:39 <mps> _ikke_: I thought about local logs but never tried to set it. laziness :(
2018-12-05 20:07:49 <_ikke_> weechat logs automatically :)
2018-12-05 20:08:11 <SpaceToast> I have znc logs, but I'm technically at work :)
2018-12-05 20:08:21 <SpaceToast> joneskoo: look at the wiki regarding lbu and apkovl
2018-12-05 20:08:26 <mps> build hplip and ghostscript on armhf about month or two ago
2018-12-05 20:08:27 <SpaceToast> it's basically a tarball that gets extracted into /
2018-12-05 20:09:23 <_ikke_> myztic_webchat?
2018-12-05 20:10:00 <danieli> mps: totally. at times, feel like I have dementia.
2018-12-05 20:10:33 <mps> I'm lying, I have a logs, just forgot that
2018-12-05 20:11:28 <_ikke_> heh: 2018-10-10 22:59:11     mps     also, sane is built, now trying hplip
2018-12-05 20:11:43 <joneskoo> SpaceToast: from what I understand I'd still need to run the raspberry image. I failed to run it in qemu so far, as I said earlier I don't have a spare pi and I'd like to automate the build.
2018-12-05 20:12:07 <_ikke_> 2018-09-14 23:19:05     TBB     having fought with hplip for a year before giving up, I should remember... but I guess my mind is trying to protect itself by hiding everything hplip related from my consciousness
2018-12-05 20:13:06 <SpaceToast> joneskoo: I think there are some very specific detail to qemu when it emulates armhf
2018-12-05 20:13:16 <SpaceToast> you might need to give it a kernel and initrd directly, but that's just off my memory
2018-12-05 20:13:28 <_ikke_> 2018-09-14 23:07:34     Michitux        That "no change" part is only true for the first commit that affects sane, in the second commit for hplip there are actually changes for all arches. As I'm also writing in the pr, I have tested both on an actual armhf device (not just building).
2018-12-05 20:13:29 <SpaceToast> I don't quite see why you'd need to run the image to make a tarball though
2018-12-05 20:13:31 <mps> _ikke_: \o/
2018-12-05 20:16:22 <joneskoo> SpaceToast: how do I get a diskless on x86_64 so I can try this out, and carry that to raspberry?
2018-12-05 20:17:47 <joneskoo> run iso with qemu, maybe setup-bootable?
2018-12-05 20:17:49 <joevandyk> danieli: the package is building, thank you for your help!
2018-12-05 20:18:02 <danieli> joevandyk: huh? wasn't mps the person who helped you?
2018-12-05 20:18:28 <joevandyk> danieli: i haven't used his patch yet :)
2018-12-05 20:18:47 <danieli> just removed sane from the dependencies?
2018-12-05 20:19:02 <joevandyk> danieli: yeah, and a couple other modifications to the APKBUILD file
2018-12-05 20:19:06 <joevandyk> so far
2018-12-05 20:21:16 <SpaceToast> joneskoo: I'm still not sure why you need to *run* the image to create a tarball
2018-12-05 20:21:31 <SpaceToast> and how you expect x86_64 binaries to run on armhf
2018-12-05 20:22:11 <joneskoo> SpaceToast: now I'm just trying to see what lbu commit actually puts into the tarball - I guess you're saying it's just a tar that gets overlaid on the diskless image?
2018-12-05 20:22:19 <SpaceToast> yes
2018-12-05 20:22:24 <SpaceToast> it's really "that" simple :)
2018-12-05 20:22:46 <SpaceToast> you can look at the sources for mkinitfs (specifically initramfs-init, iirc?) to see how it gets used and found
2018-12-05 20:22:56 <SpaceToast> (though the finding mostly happens in nlplug-findfs)
2018-12-05 20:23:10 <joneskoo> ok. well then I should be able to easily produce the tar, so now I only need to test it on A diskless alpine, where I can run x86_64 which actually boots on qemu, I don't need to get the rpi image to run in qemu
2018-12-05 20:23:46 <joneskoo> faster to iterate if I don't need to deal with swapping SD cards and plugging in a monitor and dealing with a headless system
2018-12-05 20:23:48 <CompanionCube> SpaceToast: I mean you could use qemu-user for some binairies...
2018-12-05 20:24:04 <joneskoo> if someone has commandline for qemu to run the raspberry image, that would be much appreciated and best
2018-12-05 20:24:07 <SpaceToast> CompanionCube: they keep saying they have trouble running qemu with armhf
2018-12-05 20:24:39 <joneskoo> I think it's a matter of the commandline. I found one armhf commandline but it depended on downloading a different distro's older kernel version, and it was to run raspian
2018-12-05 20:25:06 <joneskoo> everything else I tried just gives me a black screen and I'm lost how to get forward from there as I get nothing at all
2018-12-05 20:26:42 <joneskoo> ok I made myself a diskless qemu x86_64 now, yay. I think this is correct.
2018-12-05 20:28:06 <mike309> hey all. i'm just looking into alpine, and noticed that the default apk repositories (in the docker image) are served over http, rather than https. anyone know why that is? http://dl-cdn.alpinelinux.org
2018-12-05 20:28:22 <danieli> mike309: apk checks signatures of the packages
2018-12-05 20:29:10 <danieli> most linux distributions use http for mirrors
2018-12-05 20:29:45 <danieli> if you really want SSL, you can use https://alpine.global.ssl.fastly.net/alpine/
2018-12-05 20:29:59 <danieli> functionally equivalent to dl-cdn
2018-12-05 20:30:09 <_ikke_> https only protects the connection from the mirror to your machine
2018-12-05 20:30:23 <_ikke_> https can perfectly serve you malware
2018-12-05 20:30:58 <danieli> checking signatures verifies that the data is the data you want, by the correct author - https only guarantees that it wasn't modified in transit
2018-12-05 20:30:59 <joneskoo> I guess a major factor is also that traditionally distros get mirrored by 3rd parties so that's another reason why they rely on untrusted http for distribution and signed packages
2018-12-05 20:31:14 <danieli> joneskoo: that too, but really, it's superfluous
2018-12-05 20:31:31 <_ikke_> joneskoo: signed packages are superior to https
2018-12-05 20:31:36 <SpaceToast> danieli, _ikke_: checking to make sure it wasn't modified in transit (and is the correct host) is still desireable though, isn't it? :)
2018-12-05 20:31:43 <_ikke_> (don't get me wrong, they complement eachother)
2018-12-05 20:31:47 <SpaceToast> yeah
2018-12-05 20:31:52 <SpaceToast> ideally I'd like us to have both
2018-12-05 20:31:53 <_ikke_> SpaceToast: The signature achieves that
2018-12-05 20:31:56 <danieli> ^
2018-12-05 20:32:34 <SpaceToast> _ikke_: erm, so a MITM that repalces the index, signature, and package with ones that all match up with one another will be solved by the signature of the real host?
2018-12-05 20:32:46 <danieli> either way, as joneskoo says, a lot of third party mirrors are beyond our control and will not implement https
2018-12-05 20:32:46 <_ikke_> SpaceToast: They won't validate
2018-12-05 20:32:59 <SpaceToast> why not? the index is valid, the signature matches the package and what's in the index
2018-12-05 20:33:01 <_ikke_> unless your local keys get compromised
2018-12-05 20:33:07 <_ikke_> they are signed with rsa keys
2018-12-05 20:33:16 <_ikke_> not just hashes
2018-12-05 20:33:18 <SpaceToast> right, I forgot about /etc/apk/keys
2018-12-05 20:33:26 <SpaceToast> any reason we use RSA specifically btw?
2018-12-05 20:33:29 <joneskoo> the only obvious added attack surface I can think of from plain http instead of https is a MITM attacker abusing a client vulnerability - OTOH arguably you could abuse vulnerabilities in the TLS stack which are probably as likely as ones in the client/package verification
2018-12-05 20:33:37 <SpaceToast> (curve25519 might be something to look into adding)
2018-12-05 20:33:43 <danieli> SpaceToast: it's common and considered secure
2018-12-05 20:34:34 <SpaceToast> 2048-bit keys are slowly approaching the breaking point (as is RSA in general, which is why everyone's scrambling to move away)
2018-12-05 20:34:43 <SpaceToast> we could use 4096-bit keys but that becomes somewhat computationally intensive at that point
2018-12-05 20:34:55 <danieli> changing it is a discussion for another time (or if you can keep it consise, at the very least for #alpine-devel)
2018-12-05 20:35:06 <SpaceToast> fair enough ¯\_(ツ)_/¯
2018-12-05 20:35:20 <SpaceToast> still, having https seems like a good idea (and some mirrors have it :D)
2018-12-05 20:37:24 <mike309> thanks for all the info. looks like fedora is using https, but not ubuntu
2018-12-05 20:37:51 <danieli> mike309: a lot of linux distributions only default to serving https on some mirrors
2018-12-05 20:38:14 <danieli> package managers check signatures either way
2018-12-05 20:39:02 <joneskoo> mike309: https://mirrors.fedoraproject.org/mirrorlist?repo=fedora-28&arch=x86_64 less than half https if I'm not mistaken
2018-12-05 20:39:49 <danieli> SpaceToast: there *are* reasons to use https, but there are also reasons not to
2018-12-05 20:39:55 <danieli> joneskoo: 29 mirrors out of a total 73 listed are https
2018-12-05 20:40:55 <SpaceToast> danieli: the reasons I can think of is badly configured certs (e.g having a wildcard for domain A but serving from domain B), and the computational cost (very low, since certs are typically 2048 RSA, which is deemed acceptable for package signatures right now) - am I missing anything?
2018-12-05 20:41:05 <danieli> let's discuss this in #alpine-devel
2018-12-05 21:10:48 <maldridge> good afternoon.  I'm considering running Alpine "on-the-metal" in AWS.  Is this a reasonable use?  I already use Alpine for containers, but I'd like to treat my base platform the same way with alpine VMs
2018-12-05 21:12:01 <SpaceToast> I use Alpine "on the metal" in Vultr, which I suspect is similar
2018-12-05 21:12:04 <SpaceToast> it works quite well :)
2018-12-05 21:12:51 <maldridge> what do you use to build your vultr images?
2018-12-05 21:14:17 <SpaceToast> ?
2018-12-05 21:14:26 <SpaceToast> the initial install is done with PXE and that's about it
2018-12-05 21:15:59 <maldridge> ah, ok
2018-12-05 21:16:32 <maldridge> I was asking how you put together your boot image that you run on vultr.  I build my containers with packer, but I don't think I can use it to build AMIs dissapointingly.
2018-12-05 21:16:53 <SpaceToast> I take the iso from the downloads page...
2018-12-05 21:16:58 <SpaceToast> I pop it into PXE...
2018-12-05 21:18:32 <maldridge> I see... I suspect we're talking past here as I don't see how you're possibly running PXE from your own computer to a machine residing somewhere else on the internet
2018-12-05 21:19:03 <SpaceToast> ???
2018-12-05 21:19:11 <SpaceToast> Vultr is based on KVM, and thus has PXE built-in
2018-12-05 21:19:22 <SpaceToast> you just give it your boot image, and it boots your vm with the boot image attached
2018-12-05 21:19:24 <maldridge> ah, so you're running on their pxe server
2018-12-05 21:19:27 <maldridge> ok
2018-12-05 21:19:39 <SpaceToast> yes, seeing as that's the only supported way of deploying vms
2018-12-05 21:19:42 <maldridge> initially it sounded like you were somehow booting a VM in their DC from a PXE server you ran somewhere else
2018-12-05 21:25:12 <Death_Syn> maldridge: my dayjob uses packer to build AMIs, but they're based on upstream centos AMIs, not sure if that's useful to you
2018-12-05 21:25:42 <Death_Syn> i've been meaning to try it out building my own images for void and alpine
2018-12-05 21:25:46 <maldridge> yeah, it doesn't seem there's a clean alpine AMI, and I think we'd like to build from plain anyway
2018-12-05 21:27:55 <danieli> there isn't an official AMI for sure, and it hasn't been a priority
2018-12-05 21:28:11 <danieli> it'd be nice though, I'll look into it soon enough
2018-12-05 21:28:15 <maldridge> nor would I ask it to be
2018-12-05 21:28:26 <maldridge> 
ACTION maintains Void's images, which also don't have official AMIs

2018-12-05 22:15:42 <Processus42> Hello. Just a quick question: What is the stance of Alpine maintainters regarding systemd ? Shoud I use it or not ?
2018-12-05 22:16:42 <maldridge> danieli: is the infrastructure for touch free install in place?  I'm reading the docs and don't see anything that seems like it can do this.
2018-12-05 22:17:11 <Processus42> (And: I wanted to read the ML archive, but lists.alpinelinux.org seems to be down :( )
2018-12-05 22:17:39 <SpaceToast> maldridge: depends on what you mean by "touch free"
2018-12-05 22:17:45 <SpaceToast> setup-alpine can take an answer file
2018-12-05 22:17:54 <SpaceToast> but you'd need to modify your image/rootfs to actually include and run it
2018-12-05 22:17:57 <maldridge> yeah, its invoking that that's the hard part
2018-12-05 22:18:09 <maldridge> hmm, I'll see if I can drive the login prompt programatically then
2018-12-05 22:19:02 <maldridge> or if there was a way I could get the image to start with ssh, I could make that work
2018-12-05 22:19:16 <SpaceToast> you'd still need to modify the image
2018-12-05 22:19:33 <maldridge> to start ssh I would need to modify the base image?
2018-12-05 22:19:38 <SpaceToast> but you could throw an apkovl into it that changes the runlevel to include openrc, and adds openssh to your world etc
2018-12-05 22:19:50 <SpaceToast> the base image does not come with an ssh daemon pre-installed
2018-12-05 22:19:59 <maldridge> right, but that still needs the system logged in and a way to run that
2018-12-05 22:20:01 <SpaceToast> so even if it's in the cache you'd still need to install and start it
2018-12-05 22:20:04 <SpaceToast> well, no
2018-12-05 22:20:17 <maldridge> it does if I am going to start with an iso
2018-12-05 22:20:33 <maldridge> if I skip the iso I might as well just grab the static tools and start at an empty loop device
2018-12-05 22:20:46 <SpaceToast> you can unpack the iso, add an apkovl that overrides the openrc runlevel and changes your world file to include an ssh package...
2018-12-05 22:20:50 <SpaceToast> and then repack the iso
2018-12-05 22:20:52 <SpaceToast> but it is quite complicated, yes
2018-12-05 22:21:08 <maldridge> right, and by the time I've done that, I'd rather just build from a blank loop instead
2018-12-05 22:21:14 <maldridge> since that saves an unpack/repack cycle
2018-12-05 22:21:45 <SpaceToast> :)
2018-12-05 22:30:47 <joneskoo> based on quick googling, https://www.gnu.org/software/xorriso/man_1_xorriso.html might be able to mutate an ISO image? never used it.
2018-12-05 22:31:11 <maldridge> yeah, I'ves used that functionality before, its not pretty
2018-12-05 22:31:37 <maldridge> I'm trying now to use packer's functionality to send arbitrary keypresses to the VM to just log it in and then download and run a script
2018-12-05 22:33:04 <joneskoo> doesn't sound pretty either. normal ssh isn't available?
2018-12-05 22:33:23 <maldridge> not in a clean ISO right after it boots
2018-12-05 22:36:56 <joneskoo> google "packer alpine" seems to find the way you're looking at already, boot + enter commands + HTTP answers file
2018-12-05 22:37:18 <maldridge> yep
2018-12-06 04:15:14 <cordel> Hello Group! I have run into a problem I have not been able to figure out. My working environment is unique as it's basically a dark lab. I recreate GIT aports, and Repositories on a local server. The repo is served out by lighttpd and I can view the index from a browser, update various devices and VM's. I have compared SHA256 sums of packages and everything looks beauty, though i get failures using mkimage specific
2018-12-06 04:15:14 <cordel> ally for rpi profile. Have checked through my build environment, made sure tar and gzip are truly installed, I can wget files from the repo, can apk fetch packages, yet my build bombs in the beginning: >>> mkimage-armhf: --> rpi_blobs; gzip: stdin: unexpected end of file. Could the lighttpd server be returning something I'm overlooking? curl shows header Content-Type: application/octet-stream and I don't see anythin
2018-12-06 04:15:14 <cordel> g unusual. Running edge on vbox VM with NFS mount to aports, and just copied aports to a local directory in the VM. Same error either way. Been trying to run this down for a couple weeks now, so hopefully someone can nudge me in the right direction since google is failing me.
2018-12-06 10:19:28 <iveqy> hi, I've booted alpine and run setup-alpine, I choose a diskless setup and then I want to store my configs at /media/sdb so I choose sdb for storing configs. However lbu commit fails since /dev/sdb is readonly. How can I fix this? Shouldn't this be fixed automatic?
2018-12-06 10:20:50 <AinNero> > /dev/sdb is readonly
2018-12-06 10:21:02 <AinNero> the block device is readonly or is it mounted readonly?
2018-12-06 10:21:12 <iveqy> yes, and mount - remount,rw /media/sdb doesn't change that
2018-12-06 10:22:04 <iveqy> AinNero: I don't know. It's a USB memory so it should be rw?
2018-12-06 10:22:18 <AinNero> what file system type is it?
2018-12-06 10:22:46 <AinNero> because iso9660 is read-only
2018-12-06 10:23:21 <iveqy> AinNero: I don't know. I just downloaded the alpine image. So what filesystem type is used?
2018-12-06 10:23:54 <iveqy> I sounds plausible that is it iso9660
2018-12-06 10:24:03 <AinNero> the alpine iso image is an iso file, so itself it wont be writable
2018-12-06 10:24:28 <iveqy> so I downloaded the iso file and then did a dd to my usb stick and booted from that
2018-12-06 10:25:01 <iveqy> just as in the guide https://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB
2018-12-06 10:25:59 <iveqy> so if I mount an other partition of the USB memory to alpine and save the configs there. How does alpine find them on the next boot?
2018-12-06 10:26:30 <AinNero> the initramfs does search devices for stored configs
2018-12-06 10:26:39 <AinNero> so it will work automagically
2018-12-06 10:27:16 <AinNero> hm, better would be to setup the disk with setup-bootable
2018-12-06 10:28:19 <iveqy> is that the same as install to disk  in setup-alpine?
2018-12-06 10:29:14 <AinNero> no, setup-bootable does an diskless boot per default, while install to disk will result in a root partition
2018-12-06 10:30:02 <iveqy> because setup-bootable wants a mounted directory and I guess it has to be a writeable directory and hence I can't do setup-bootable to the same usb stick that I've booted from
2018-12-06 10:31:16 <iveqy> hmm, I found a guide section about setup-bootable
2018-12-06 10:31:49 <iveqy> I'll try that. It seems like this should not be needed, when reading the first part of that page about newer alpine iso images already containing MBR
2018-12-06 10:35:28 <AinNero> iveqy: i did that once, adding another partition after the iso file system
2018-12-06 10:36:16 <AinNero> but the iso is special because it can also be mounted without mbr, as sdb instead of sdb1
2018-12-06 10:36:31 <AinNero> and that happened, and while sdb is mounted, i can't as well mount sdb2
2018-12-06 10:37:45 <iveqy> AinNero: that's exactly the problem I'm facing now
2018-12-06 10:38:33 <iveqy> so perhaps booting from one USB memory and then inserting an other usb memory, running setup-bootable to the second usb memory and then just remove the first usb memory and reboot
2018-12-06 10:38:59 <AinNero> uhm, there is a way to detach a diskless boot medium so you can do it on the same usb stick
2018-12-06 10:39:45 <AinNero> boot into diskless, copy the modloop into /, umount the modloop from the usb, mount the modloop from /
2018-12-06 10:39:50 <AinNero> then you can umount the usb
2018-12-06 10:48:24 <iveqy> I wish there was a way to just use the ISO and add some configuration to it
2018-12-06 10:48:37 <iveqy> it would be mush simplier to create new usb sticks then
2018-12-06 11:00:22 <AinNero> maybe there should be an /boot tarball like there is for the raspi
2018-12-06 12:23:59 <AinNero> whats the difference between gcompat and libc6-compat?
2018-12-06 12:25:18 <AinNero> is gcompat installable at all, since it both depends on and replaces libc6-compat?
2018-12-06 12:46:56 <TBB> it is; I haven't checked but we use it in our work project
2018-12-06 13:06:29 <AinNero> aports/pull/5661 seems to be okay, but im considering rewriting that part of the code to support arbitrary interpreter paths
2018-12-06 13:06:36 <AinNero> because glibc interpreter paths are somewhat arbitrary
2018-12-06 13:07:26 <AinNero> TBB: did something happen regarding the mkinitfs thing?
2018-12-06 15:34:33 <geosmin> i assume an edited /etc/init.d/hostname will get replaced after an update. is this correct?
2018-12-06 15:37:57 <geosmin> if yes, is the same true for /etc/conf.d/hostname?
2018-12-06 15:38:25 <_ikke_> geosmin: afaik, they are considered configuration files, so they won't be overwritten
2018-12-06 15:49:50 <geosmin> _ikke_: both or just the ones in conf.d?
2018-12-06 15:50:22 <_ikke_> afaik both
2018-12-06 15:51:22 <geosmin> alright, thanks
2018-12-06 16:16:40 <AinNero> afaik the init.d ones are per default excluded from lbu
2018-12-06 17:22:16 <geosmin> hmm
2018-12-06 17:22:39 <geosmin> it seems /etc/conf.d/hostname defines hostname="localhost"
2018-12-06 17:23:33 <geosmin> this appears to be unused by /etc/init.d/hostname, which just checks if /etc/hostname exists else sets it explicitly to "localhost", not the variable defined in conf.d
2018-12-06 18:27:15 <maldridge> does alpine have a mechanism for network login?
2018-12-06 18:32:29 <danieli> define network login
2018-12-06 18:32:48 <maldridge> I have a central source of account data and I use that to log on to the platform
2018-12-06 18:32:59 <maldridge> for musl I believe you would need musl-nslcd
2018-12-06 18:33:03 <jwh> more verbosity
2018-12-06 18:33:15 <danieli> I'd say it depends on the source of data
2018-12-06 18:33:17 <jwh> so ldap
2018-12-06 18:33:25 <maldridge> sure, we can use ldap in this example
2018-12-06 18:33:39 <jwh> well thats what nslcd is...
2018-12-06 18:33:55 <maldridge> strictly, nslcd is a protocol, and you can plug in other modules that understand how to speak it
2018-12-06 18:34:03 <jwh> yeah, but traditionally
2018-12-06 18:34:10 <maldridge> s/is a/impliments a/
2018-12-06 18:34:11 <jwh> but be less vague and maybe someone will have a suggestion
2018-12-06 18:34:27 <danieli> yeah, it's a bit vague for me
2018-12-06 18:34:36 <maldridge> my vagueness is because if the base infrastructure is there, I'll port over the packages I want to use the authentication mechanisms I want
2018-12-06 18:35:13 <maldridge> if its possible to get ldap to work with alpine, then I can get another one to work too, but I know musl is tricky with feeding in group and passwd data from non /etc/* sources
2018-12-06 18:35:29 <maldridge> and looking at the package tree, I don't see the shim packages I think need to be there to support this
2018-12-06 18:36:00 <danieli> I don't know if network login is possible on alpine, and I haven't tried - I haven't used alpine in enterprise systems myself
2018-12-06 18:36:02 <danieli> someone else might have
2018-12-06 18:36:11 <maldridge> ok
2018-12-06 18:43:23 <maldridge> hmm, looks like I'd need musl-nscd.  I can look into porting this if I can get it working in a lab env
2018-12-06 22:16:36 <g1eb> hii alpine! I'm working on a docker container trying to apk add lapack/openblas
2018-12-06 22:16:55 <g1eb> from what i understand these are only available on the community packages repo
2018-12-06 22:18:12 <g1eb> in the end i want install scipy with pip, scipy requires those libs to be installed, and it fails during the build step
2018-12-06 22:18:46 <g1eb> can somebody help me installe these community packages in docker? i can share my Dockerfile :)
2018-12-06 22:21:30 <g1eb> this is what i have now ^ http://vpaste.net/WVdWE
2018-12-06 22:27:02 <SpaceToast> g1eb: it'd help to see the logs of "what failed"
2018-12-06 22:27:09 <SpaceToast> as well as those magical "./" contents
2018-12-06 22:28:01 <SpaceToast> but, a few things to note:
2018-12-06 22:28:05 <g1eb> ohh I can provide logs of the docker build command, they fail with numpy intall missing those dependent libs
2018-12-06 22:28:07 <SpaceToast> 1. the community repo is included by default
2018-12-06 22:28:19 <SpaceToast> 2. don't use --update, as per docker recommendations, only install new things you need
2018-12-06 22:28:29 <SpaceToast> 3. don't rm /var/apk/cache - just run your apk commands with --no-cache
2018-12-06 22:28:43 <SpaceToast> 4. try basing your image on something newer than 3.6
2018-12-06 22:28:45 <SpaceToast> 5. py-pip exists
2018-12-06 22:29:27 <SpaceToast> ok, so can I have your repo, so I can try and fix the giant mess that your Dockerfile is?
2018-12-06 22:30:36 <g1eb> SpaceToast, thanks for the help! Let me see what I can change quickly and post the logs
2018-12-06 22:30:42 <SpaceToast> ...
2018-12-06 22:32:11 <SpaceToast> also there is no reason to have multiple apk runs
2018-12-06 22:33:18 <g1eb> yeah, i was trying to copy what jfloff said in his reply here: https://github.com/jfloff/alpine-python/issues/24
2018-12-06 22:33:48 <g1eb> i think it's a similar issue, here are my build logs http://vpaste.net/4i2BG
2018-12-06 22:34:17 <g1eb> there really isnt anything magical in the ./ directory, as far as requirements go I need numpy and scipy
2018-12-06 22:35:02 <SpaceToast> I cannot reproduce
2018-12-06 22:35:04 <SpaceToast> your code
2018-12-06 22:35:07 <SpaceToast> without the ./ directory
2018-12-06 22:35:12 <SpaceToast> because I have nothing in it
2018-12-06 22:35:16 <SpaceToast> anything in it is thus magical.
2018-12-06 22:36:14 <g1eb> let me see what i can do about that, get a smaller reproduction case :)
2018-12-06 22:40:03 <g1eb> sorry that Dockerfile had a lot of legacy stuff in there, im just cleaning it up
2018-12-06 22:40:46 <SpaceToast> by the way
2018-12-06 22:40:51 <SpaceToast> py3-numpy and py3-scipy are also packages ;)
2018-12-06 22:41:05 <SpaceToast> (though sure, if you have a "real repo" that might be awkward)
2018-12-06 22:41:13 <SpaceToast> (you may want to take advantage of a build stage though)
2018-12-06 22:41:25 <g1eb> you said the community repo is included by default right?
2018-12-06 22:41:35 <SpaceToast> yes.
2018-12-06 22:41:50 <SpaceToast> no need for fancy tagging and stuff like that
2018-12-06 22:42:01 <SpaceToast> to try and reproduce I'm currently trying to get a dockerfile that just pip installs numpy and scipy
2018-12-06 22:42:06 <SpaceToast> building takes time due to laptop though
2018-12-06 22:42:06 <g1eb> i can remove the community repo url and the @community suffixes then
2018-12-06 22:42:34 <SpaceToast> it seems to build just fine here though
2018-12-06 22:42:39 <g1eb> yeah, im doing the same, getting an empty directory with just requirements.txt
2018-12-06 22:42:43 <SpaceToast> so if/when it installs, I'll show you how horrible it can be ;)
2018-12-06 22:43:55 <SpaceToast> numpy passed just fine
2018-12-06 22:45:18 <g1eb> im running into docker not finding the requirements file haha
2018-12-06 22:46:10 <g1eb> ohh i see, i had to put it into that funny directory
2018-12-06 22:47:11 <Greyztar> how can i prevent resolv.conf from changing at each reboot?Neither seems to work
2018-12-06 22:48:19 <SpaceToast> Greyztar: by not using things that overwrite it
2018-12-06 22:48:32 <SpaceToast> g1eb: also consider throwing your WORKDIR into the start, and not doing weird bash cd things
2018-12-06 22:50:37 <Greyztar> SpaceToast: im running it as an lxc image,i think its the udhcpd is doing it,which ties to the netowrk script no?
2018-12-06 22:50:52 <g1eb> SpaceToast so glad you are helping, this whole thing is full of legacy code i wonder how it even worked so well till now
2018-12-06 22:51:38 <Greyztar> SpaceToast: i tried using static ip as per wiki no luck either
2018-12-06 22:52:43 <SpaceToast> Greyztar: tl;dr figure out what's overwriting it (it may be lxc itself, for example) and fix that :)
2018-12-06 22:52:56 <SpaceToast> for instance, I'm not sure exactly how busybox ifupdown actually summons dhcp
2018-12-06 22:53:08 <SpaceToast> you might be able to pass it NO_DNS=my_iface
2018-12-06 22:54:25 <Greyztar> SpaceToast: been battling it all day long now hehe,for most the other systemd containers its just remove the symlink,also i can set static ip,didwith alpine aswell but still gets overwritten,its the udhcp thing which i cant stop seems
2018-12-06 22:55:24 <Greyztar> SpaceToast: ill check with ifupdown
2018-12-06 22:59:07 <SpaceToast> ping g1eb
2018-12-06 22:59:24 <g1eb> SpaceToast I've got a repo to reproduce this issue
2018-12-06 22:59:30 <SpaceToast> g1eb: http://sprunge.us/4tsy5v
2018-12-06 22:59:33 <SpaceToast> I see no issue
2018-12-06 22:59:37 <SpaceToast> this builds successfully
2018-12-06 22:59:43 <SpaceToast> enjoy :)
2018-12-06 23:00:21 <g1eb> oooh lemme see what the difference is with my dockerfile, thank you so much sapcetoast!
2018-12-06 23:01:05 <g1eb> it's kinda sad it takes 10-15 minutes to compile this guy and in the end it pukes for smth that failed 12 minutes ago haha
2018-12-06 23:01:15 <Processus42> Hello there. I (sys) installed alpine 3.8 on a VirtualBox VM from a sandard iso and (mostly) default options. The VM has only one core. When I tried to ssh on it, it went on kernel panic. I had to enable NX feature to make it works. Why is it even booting when the VM doesn't supports NX ? What are the options used to compile the kernel ?
2018-12-06 23:01:50 <SpaceToast> Processus42: that can be seen here: https://git.alpinelinux.org/cgit/aports/tree/main/linux-vanilla
2018-12-06 23:02:03 <SpaceToast> however, if you're installing to a virtual machine, you should probably be using the virtual kernel and image.
2018-12-06 23:04:06 <Processus42> I tried, but it couldn't recognize the virtual network card :/
2018-12-06 23:04:50 <SpaceToast> that's strange, given that I've ran that exact setup before ;)
2018-12-06 23:05:25 <Processus42> The network card wasn't showing up in ip
2018-12-06 23:05:33 <Processus42> I'll give it a try once more
2018-12-06 23:06:46 <Processus42> Now it can recognize one of the two inserted card
2018-12-06 23:09:57 <Processus42> Okay, it can recognize the HostOnlyIf
2018-12-06 23:10:27 <Processus42> But not the NAT one (I started with only a NAT the first time, that's why it didn't shown up in ip link )
2018-12-06 23:11:36 <SpaceToast> how do you determine if it "shows up" or not?
2018-12-06 23:11:47 <Processus42> I just type ip link
2018-12-06 23:11:57 <SpaceToast> hmm, ok, it should be showing up
2018-12-06 23:12:09 <SpaceToast> very strange, since I've definitely had a NAT card working
2018-12-06 23:12:11 <Processus42> (Oh, and I'm solving the IP with dhcpcd to determine which one it is)
2018-12-06 23:12:20 <SpaceToast> I'm doing other things right now
2018-12-06 23:12:21 <Processus42> Can you give it a try now ?
2018-12-06 23:12:24 <Processus42> Okay
2018-12-06 23:12:26 <SpaceToast> but I'll give it a try soon
2018-12-06 23:12:28 <SpaceToast> :)
2018-12-06 23:12:33 <Processus42> Thanks :)
2018-12-06 23:12:37 <SpaceToast> if you stick around here I'll ping you once I've done that
2018-12-06 23:12:43 <Processus42> Thanks
2018-12-06 23:12:55 <Processus42> I'll try to find what's the issue as well ;)
2018-12-06 23:18:04 <palate> hello :)
2018-12-06 23:18:40 <palate> Trying to build a C++ module including <asm/termbits.h>, I get: "No such file or directory"
2018-12-06 23:18:50 <Processus42> SpaceToast: It's on my side: VM misconfiguration (My NAT card is configured as Am79C873 and the HostOnly is 82540EM). There is drivers for the 82540EM but not for the Am79C873. I'll just change the Am79C873 to the Intel one.
2018-12-06 23:19:00 <palate> I have no clue why that could be, so I don't know what to look for :/
2018-12-06 23:19:09 <SpaceToast> Processus42: there we are :D
2018-12-06 23:19:25 <SpaceToast> palate: can you find that file and print the path here?
2018-12-06 23:19:45 <Processus42> Now I'll re-install Alpine with virt :)
2018-12-06 23:19:53 <Processus42> *for virt
2018-12-06 23:19:58 <palate> SpaceToast: this file is not on my alpine linux, but it is in Ubuntu, for instance
2018-12-06 23:20:12 <SpaceToast> have you considered getting that file? ;)
2018-12-06 23:20:23 <SpaceToast> let's see...
2018-12-06 23:20:34 <SpaceToast> looks like it's in linux-headers
2018-12-06 23:20:50 <SpaceToast> so tl;dr - install your dependencies
2018-12-06 23:20:51 <palate> SpaceToast: I'm always confused with linux-headers... is that something I should install?
2018-12-06 23:20:56 <SpaceToast> yes, it is a package
2018-12-06 23:21:03 <SpaceToast> you install it by running `apk add linux-headers`
2018-12-06 23:21:14 <SpaceToast> they are the development headers for the linux kernel, which your C++ module apparently uses
2018-12-06 23:21:15 <palate> SpaceToast: why is it not installed by default? Sounds important... xD
2018-12-06 23:21:33 <SpaceToast> do you want the development packages for the linux header in a system that will never compile code?
2018-12-06 23:21:36 <SpaceToast> no, of course not
2018-12-06 23:21:47 <SpaceToast> if you're going to be installing linux-headers, you also want to install gcc by default
2018-12-06 23:21:55 <SpaceToast> and surely, you wouldn't argue gcc should be installed by default too ;)
2018-12-06 23:22:13 <palate> SpaceToast: no, but gcc comes with build-base, doesn't it?
2018-12-06 23:22:19 <SpaceToast> sure
2018-12-06 23:22:25 <SpaceToast> and not everything actually uses linux-headers
2018-12-06 23:22:35 <palate> :)
2018-12-06 23:22:51 <SpaceToast> so it shouldn't be in build-base, should it?
2018-12-06 23:22:56 <SpaceToast> since it's not a basic build requirement :D
2018-12-06 23:23:02 <palate> I guess I got confused by the name... "linux-headers"  sounds pretty important on a linux-based system :D
2018-12-06 23:23:19 <palate> let me try to build with that, then
2018-12-06 23:23:59 <palate> ok, i'm probably missing other things: error: 'strndupa' was not declared in this scope
2018-12-06 23:24:16 <SpaceToast> try and see what that module you're trying to compile claims it depends on
2018-12-06 23:24:19 <palate> It's actually a good thing: alpine is very minimal, so I get to learn the real dependencies :D
2018-12-06 23:24:19 <SpaceToast> and then install those things ;)
2018-12-06 23:24:34 <SpaceToast> most good packages will straight up tell you (e.g in the README) what you need to have to compile them
2018-12-06 23:25:08 <palate> SpaceToast: that module is developed on Ubuntu, they get many things by default :D
2018-12-06 23:25:35 <palate> that's the one: https://github.com/intel/mavlink-router
2018-12-06 23:25:36 <SpaceToast> so what you're telling me is that the module's authors are not very good at that whole "best practices" stuff ;)
2018-12-06 23:25:41 <SpaceToast> oh, intel
2018-12-06 23:25:45 <SpaceToast> so that IS what you're saying :)
2018-12-06 23:26:11 <palate> haha
2018-12-06 23:26:59 <SpaceToast> intel make some really cool things in hardware (especially the networking segment)
2018-12-06 23:27:07 <SpaceToast> but their software skills could use some erm... brushing up
2018-12-06 23:27:18 <SpaceToast> (for instance, iwd, for the longest time, straight up would cause failures)
2018-12-06 23:27:26 <SpaceToast> (and iwd is one of their pieces of software I like the most!)
2018-12-06 23:28:24 <palate> oh I did not know it
2018-12-06 23:28:30 <palate> I know wpa_supplicant, though
2018-12-06 23:28:40 <SpaceToast> yeah, iwd is a wpa_supplicant + iwconfig replacement
2018-12-06 23:28:50 <SpaceToast> it's quite nice, when it's not full of horrifying eldritch bugs
2018-12-06 23:28:52 <SpaceToast> anyway, you can look up where each symbol is (use google, it is your friend :) )
2018-12-06 23:29:02 <SpaceToast> once you've figured out all of the dependencies, you might try sending them a bug report / PR
2018-12-06 23:29:15 <SpaceToast> basically going "list of deps: you need these; please add to README"
2018-12-06 23:30:24 <palate> right
2018-12-06 23:30:37 <palate> yeah, it's not so easy to find which package has which symbol
2018-12-06 23:31:14 <SpaceToast> if you can identify what file it's supposed to be in, you can use this to look for it: https://pkgs.alpinelinux.org/contents
2018-12-06 23:31:25 <SpaceToast> and you can use google to identify the library it should be in
2018-12-06 23:31:29 <SpaceToast> past that it's mostly putting 2 and 2 together :)
2018-12-06 23:31:40 <SpaceToast> (I've done this a lot when I was trying to do LFS on Cygwin for fun)
2018-12-06 23:32:37 <palate> SpaceToast: seems like it's supposed to be inside string.h... doesn't help much
2018-12-06 23:33:14 <SpaceToast> palate: that'd be musl-dev
2018-12-06 23:33:17 <palate> When it comes to things that seem very basic in a linux system, I apparently quickly get confused. Say if a linux system doesn't have the same string.h as another one, I feel like I'm just screwed
2018-12-06 23:33:20 <SpaceToast> https://pkgs.alpinelinux.org/contents?file=string.h&path=%2Fusr%2Finclude&name=&branch=edge&arch=x86_64 - this is the search I used
2018-12-06 23:33:32 <palate> SpaceToast: I installed musl-dev, but I still have the same error
2018-12-06 23:33:37 <SpaceToast> as I said, if you figure out the file it's in, you can use https://pkgs.alpinelinux.org/contents to search
2018-12-06 23:33:41 <SpaceToast> then it is not in string.h :)
2018-12-06 23:33:46 <palate> haha
2018-12-06 23:34:07 <palate> I need to practice my "looking-for-symbols" skills :D
2018-12-06 23:34:31 <SpaceToast> ah, looks like that's a musl thing
2018-12-06 23:34:33 <SpaceToast> you might need to patch
2018-12-06 23:34:46 <SpaceToast> intel is intel so they assumed glibc
2018-12-06 23:34:50 <SpaceToast> because of course they did
2018-12-06 23:35:15 <SpaceToast> you can define it as so : #define strndupa(x, t) strncpy(alloca(strlen(x)+1),x,t)
2018-12-06 23:36:14 <palate> SpaceToast: right, so Alpine is not using glibc
2018-12-06 23:36:20 <SpaceToast> no, we use musl
2018-12-06 23:36:27 <palate> oh, that's what musl it :)
2018-12-06 23:36:29 <SpaceToast> glibc is too non-standard, buggy, heavy, and broken :)
2018-12-06 23:36:34 <palate> s/it/is
2018-12-06 23:37:03 <palate> I like the Alpine mentality. And I don't like that most developers use Ubuntu and don't care about other distros
2018-12-06 23:37:57 <palate> SpaceToast: where would I define that? =/
2018-12-06 23:38:07 <SpaceToast> in a header it's used in :D
2018-12-06 23:38:17 <SpaceToast> assuming Intel aren't insane, wherever they #include <string.h>
2018-12-06 23:38:21 <SpaceToast> unfortunately, intel is intel
2018-12-06 23:38:26 <palate> :D
2018-12-06 23:38:50 <SpaceToast> so there are no guarantees they aren't insane
2018-12-06 23:38:52 <SpaceToast> (if anything...)
2018-12-06 23:42:18 <palate> actually, since I make a docker container just for this project... would it be super bad to patch string.h?
2018-12-06 23:43:10 <SpaceToast> it would make Rich Felker very very sad
2018-12-06 23:43:31 <SpaceToast> oh also you're meant to use strnlen, not strlen
2018-12-06 23:43:33 <SpaceToast> blech
2018-12-06 23:43:42 <SpaceToast> then there's macro arguments, etc :/
2018-12-06 23:44:07 <palate> Now I know who Rich Felker is :D
2018-12-06 23:44:23 <palate> I did not get your point about strnlen? You mean the intel project uses that?
2018-12-06 23:44:32 <SpaceToast> no, that's the correct strndupa implementation
2018-12-06 23:44:34 <SpaceToast> hold on
2018-12-06 23:44:44 <SpaceToast> https://cgit.gentoo.org/proj/musl.git/tree/sys-libs/efivar/files/27-strndupa.patch?id=0d259ad72bb04c5a9bc1c83d35b1cd327f4240e6
2018-12-06 23:45:03 <SpaceToast> that stuff in green is what you can patch string.h is, but I recommend patching the actual project
2018-12-06 23:45:06 <SpaceToast> rather than system files
2018-12-06 23:45:09 <SpaceToast> (and then sending the patch in :) )
2018-12-06 23:45:38 <palate> oh, gentoo uses musl as well
2018-12-06 23:46:02 <palate> Well, the project includes <string.h> in 6 places: https://github.com/intel/mavlink-router/search?q=string.h&unscoped_q=string.h
2018-12-06 23:46:24 <SpaceToast> gentoo has a musl stage, and supports it as a libc
2018-12-06 23:46:33 <SpaceToast> I was part of making that happen, back when I actually worked with it :)
2018-12-06 23:47:11 <palate> So I'm not sure what I would do there. Make them all include a common header like "common.h", and have this common.h include string.h and define strndupa?
2018-12-06 23:47:28 <palate> Oh that's cool :)
2018-12-06 23:47:31 <SpaceToast> palate: looks like putting that into common/util.h would help
2018-12-06 23:47:34 <palate> So you're one of the creators of Alpine?
2018-12-06 23:47:37 <SpaceToast> lol no
2018-12-06 23:47:42 <SpaceToast> I'm a recent arrival here
2018-12-06 23:47:50 <palate> but you were working on gentoo?
2018-12-06 23:48:00 <SpaceToast> I contributed to Gentoo, a while ago, yeah
2018-12-06 23:48:04 <SpaceToast> mostly llvm / musl stuff
2018-12-06 23:48:07 <palate> right
2018-12-06 23:48:24 <palate> I'm on Arch Linux, actually. Always wonder if I should try gentoo at some point
2018-12-06 23:48:30 <SpaceToast> anyway, put the stuff in that gentoo patch (ifndef -> endif) into common/util.h
2018-12-06 23:48:36 <SpaceToast> and file a bug to that intel project
2018-12-06 23:48:41 <palate> I try to use alpine for my docker containers, but as you see I sometimes struggle with compilation :)
2018-12-06 23:48:43 <SpaceToast> yelling "YOU GUYS ARE BAD AND NOT MUSL COMPATIBLE"
2018-12-06 23:48:50 <SpaceToast> and shove that in their faces \o/
2018-12-06 23:48:55 <palate> :D
2018-12-06 23:49:06 <palate> Let me try that
2018-12-06 23:49:08 <SpaceToast> (don't actually yell, please be polite)
2018-12-06 23:49:13 <palate> hahahahaha
2018-12-06 23:49:26 <SpaceToast> (even if it's.... ew.... intel)
2018-12-06 23:54:23 <palate> arf. That goes one step further, but I fear I will never manage to make it musl compatible
2018-12-06 23:54:35 <palate> "TCSETS2 was not declared" now
2018-12-06 23:55:02 <maldridge> don't lose hope!
2018-12-06 23:55:03 <palate> I remember having seen that in another project, that was coming from intel people as well, and I had to patch the source code to remove it
2018-12-06 23:55:16 <SpaceToast> ah, intel..
2018-12-06 23:55:32 <palate> And it turned out that it was not used anywhere -_-
2018-12-06 23:55:39 <SpaceToast> :D
2018-12-06 23:55:41 <SpaceToast> intel! \o/
2018-12-07 00:01:04 <palate> interestingly, TCSETS2 is in /usr/include/asm-generic/ioctls.h, on my alpine...
2018-12-07 00:01:41 <SpaceToast> sounds like they're not including it then~
2018-12-07 00:02:13 <palate> yes
2018-12-07 00:02:21 <palate> included it, and it compiles now \o/
2018-12-07 00:31:41 <palate> Great, it builds! I'll test it tomorrow and will open a PR for Intel then
2018-12-07 00:31:59 <SpaceToast> nice :)
2018-12-07 00:32:01 <palate> Thanks a lot for your help SpaceToast! Very helpful!
2018-12-07 00:32:02 <SpaceToast> good luck!
2018-12-07 00:32:24 <palate> Going to bet now, it's 01:30 am here ^^
2018-12-07 00:32:28 <palate> good night ;-)
2018-12-07 00:33:37 <SpaceToast> \o nite
2018-12-07 12:32:59 <mikroskeem> hello! i'd like to know that what license do all aports' apkbuild files use?
2018-12-07 12:33:08 <mikroskeem> i tried looking around but i'm not sure
2018-12-07 12:33:17 <mikroskeem> ..where i could find that information
2018-12-07 12:33:54 <danieli> I'd assume MIT, but now that I'm looking at it, there's no LICENSE file or license defined in the readme
2018-12-07 12:33:56 <danieli> ncopa: ?
2018-12-07 12:34:22 <TBB> I seem to remember a discussion about this earlier this year
2018-12-07 12:34:45 <danieli> I'm senile ¯\_(ツ)_/¯
2018-12-07 12:36:08 <mikroskeem> well thing is that some of the PKGBUILDs i've created for my distro are derived from APKBUILDs
2018-12-07 12:36:21 <mikroskeem> and i'd like to clarify that what license they use
2018-12-07 12:36:32 <mikroskeem> so that's bit.. tricky
2018-12-07 13:48:31 <TBB> okay, there was a conversation about the topic almost exactly one year ago, there were suggestions that APKBUILDs should be MIT licensed, some said APKBUILDs should follow the license of the software they package. a lot of heated discussion ensued, a lawyer had some insightful comments to add, and in the end it was found to be a complicated mess and that's pretty much it
2018-12-07 13:54:22 <danieli> so apkbuilds are unlicensed
2018-12-07 14:00:31 <jwh> licensed apkbuilds? ugh
2018-12-07 14:00:40 <jwh> what insanity is this
2018-12-07 14:04:18 <jn__> well, they're arguably nontrivial
2018-12-07 14:04:29 <_ikke_> Aren't the APKBUILDS copyrighted to the authors by default?
2018-12-07 14:04:41 <jn__> and i'd be better to have clarity on the allowed usage
2018-12-07 14:05:15 <SpaceToast> _ikke_: yeah, afaik the default is a copyright to the author
2018-12-07 14:06:00 <SpaceToast> as I believe I've mentioned before, any code *I*'ve written may be considered MIT, and any non-code I've written may be considered CC-BY-SA, and I'd suggest the same for alpine as a whole, but that's not my decision to make
2018-12-07 14:12:29 <mps> OT: soon, every word will have attached license :(
2018-12-07 14:19:37 <jn__> even a statement of triviality for the whole repo would probably be better than having no license at all
2018-12-07 14:31:42 <danieli> _ikke_: bingo
2018-12-07 14:32:06 <danieli> SpaceToast: unless explicitly defined, i don't think it matters
2018-12-07 14:52:58 <SpaceToast> danieli: well, implicit is copyright to authors, which means derivative works are technically forbidden
2018-12-07 14:53:08 <danieli> yup
2018-12-07 14:54:02 <SpaceToast> which has a couple of interesting side-effects
2018-12-07 14:54:29 <SpaceToast> for instance, the original author for any given APKBUILD can, at any time, sue any contributor making a patch
2018-12-07 15:25:24 <mps> SpaceToast: you wrote that terrific words just when I'm in changing APKBUILD to fix it :)
2018-12-07 15:25:47 <SpaceToast> mps: hey I've submitted patches too (one of them even made it in)
2018-12-07 15:26:09 <SpaceToast> my point being though - it can have (large) effects, and thus does matter
2018-12-07 15:26:35 <mps> SpaceToast: I'm joking about all that IP nonsense
2018-12-07 15:28:47 <mps> should we all ask permissions (and paying) descendants of old priests to use any writing system ;)
2018-12-07 15:29:37 <SpaceToast> if they're alive, have legal right to sue, and are willing to do so...
2018-12-07 15:29:39 <SpaceToast> probably, yes ;)
2018-12-07 15:30:08 <SpaceToast> they don't have legal right to sue, mostly because of how copyright law works
2018-12-07 15:30:21 <SpaceToast> (it expires unless explicitly renewed, but it takes quite a while)
2018-12-07 15:30:33 <mps> have to look at my ancestry line. maybe I could be rich overnight :)
2018-12-07 15:30:37 <SpaceToast> still, that's a false equivalency
2018-12-07 15:31:00 <mikroskeem> ugh
2018-12-07 15:31:20 <mikroskeem> so, basically i'm not allowed to use them in the end? :D
2018-12-07 15:31:38 <SpaceToast> mikroskeem: you can, but the original authors can also sue you over your usage
2018-12-07 15:31:45 <SpaceToast> you can ask them individually to obtain individual licenses
2018-12-07 15:31:53 <SpaceToast> (I suspect they won't mind)
2018-12-07 15:32:34 <mps> it is old marketing trick: Fear, Uncertainty and Doubt
2018-12-07 15:32:43 <mikroskeem> right now i have `# Based on: https://git.alpinelinux.org/cgit/aports/tree/main/x/APKBUILD` in the beginning of my PKGBUILDs
2018-12-07 15:32:57 <SpaceToast> attribution != license
2018-12-07 15:33:00 <mikroskeem> and same goes for other sources
2018-12-07 15:51:01 <jomat> gajim 1.0.3-r0 (edge/testing) working for anyone?
2018-12-07 15:52:29 <jomat> Error relocating /usr/lib/python3.6/site-packages/cryptography/hazmat/bindings/_openssl.abi3.so: SSL_CTX_set_psk_client_callback: symbol not found
2018-12-07 15:55:39 <_ikke_> Does it need to be recompiled?
2018-12-07 16:18:04 <jomat> _ikke_: I don't know, perhaps py3-cryptography as _openssl.abi3.so is from there
2018-12-07 17:17:40 <geosmin> hoping someone can help with this; i'd like to set a static ip based on hostname
2018-12-07 17:18:24 <_ikke_> can you elaborate?
2018-12-07 17:18:48 <geosmin> one thing that comes to mind is creating a service that tweaks /etc/network/interfaces before networking goes up
2018-12-07 17:18:50 <geosmin> right
2018-12-07 17:19:08 <geosmin> building an embedded solution. i'm aiming for it ot be agnostic
2018-12-07 17:19:41 <geosmin> we've got a number of pis that each do similar but different tasks
2018-12-07 17:20:08 <geosmin> ideally, i'd like to maintain a single alpine image
2018-12-07 17:20:36 <geosmin> how i'm setting things up now, they determine what their role is based on a gpio bridge
2018-12-07 17:21:27 <geosmin> ie. at boot it checks if pins x and y are connected, if yes set hostname to foo, set static ip to bar, and the service in question for our use-case changes what it does accordingly
2018-12-07 17:22:55 <geosmin> i suspect this isn't the best implementation, but currently a modified /etc/conf.d/hostname does the gpio checks and changes the hostname before unexporting the pins, subsequent processes now simply reference the hostname to know what to do
2018-12-07 17:23:41 <geosmin> anyway, for maitanence i'd like to, at boot, set astatic ip based on either the gpio pins or the hostname
2018-12-07 17:24:02 <geosmin> easier to ssh in, but maybe there's a better solution
2018-12-07 17:25:07 <geosmin> (unfortunately we can't set static ips to DNS gateway-side)
2018-12-07 17:25:15 <geosmin> static IPs to mac addresses*
2018-12-07 17:25:36 <geosmin> ah, and in any case that wouldn't work since it's no longer hardware-agnostic
2018-12-07 17:25:53 <geosmin> tl;dr: i want to set a static ip based on bridged gpio pins
2018-12-07 17:26:06 <geosmin> ...sorry if all that was to verbose :)
2018-12-07 17:27:35 <_ikke_> I'm eating atm
2018-12-07 17:32:02 <danieli> yeah, that was both pretty verbose (15 messages) and it seems pretty pi-specific
2018-12-07 17:48:32 <geosmin> my thinking right now is a seperate service between hostname and bringing the network up that seds /etc/network/interfaces based on the hostname
2018-12-07 17:48:55 <geosmin> does that sound reasonable?
2018-12-07 18:17:05 <squarebracket> is there a way to add CA certificates to busybox's wget?
2018-12-07 18:19:52 <squarebracket> i used to be able to add them with update-ca-certificates
2018-12-07 18:38:37 <jwh> hmm, so... efi alpine?
2018-12-07 18:40:00 <SpaceToast> squarebracket: update-ca-certificates(1) is a part of the ca-certificates package
2018-12-07 18:40:04 <SpaceToast> please make sure you have it installed
2018-12-07 19:22:19 <jwh> so uh, because I suck
2018-12-07 19:22:42 <jwh> can mkinitfs make a "fallback" initramfs? (a la the arch one?)
2018-12-07 19:22:55 <SpaceToast> jwh: what would the difference be?
2018-12-07 19:24:58 <jwh> moar modules
2018-12-07 19:25:00 <jwh> basically
2018-12-07 19:25:37 <jwh> rather than autodetection and such
2018-12-07 19:25:59 <SpaceToast> jwh: alpine doesn't do module autodetection atm
2018-12-07 19:26:08 <SpaceToast> they are statically configured in /etc/mkinitfs/mkinitfs.conf
2018-12-07 19:26:17 <jwh> oh hm
2018-12-07 19:26:17 <SpaceToast> as such, I continue to fail to see why you'd want two :)
2018-12-07 19:26:27 <jwh> it has a limited module selection though right?
2018-12-07 19:26:36 <SpaceToast> the modules available are in features.d
2018-12-07 19:26:41 <jwh> well, I don't know exactly what my target device might need
2018-12-07 19:26:43 <SpaceToast> if you want more, you can submit a patch to mkinitfs
2018-12-07 19:26:56 <SpaceToast> I would suggest you start by figuring that out then :)
2018-12-07 19:27:29 <jwh> thats why I was asking if there was a way to generate a more inclusive initramfs :D
2018-12-07 19:27:39 <SpaceToast> yes - by adding more features to mkinitfs.conf
2018-12-07 19:27:45 <jwh> but it might be moot anyway, can't find a keyboard
2018-12-07 19:27:48 <SpaceToast> that'll just be your main one, rather than "backup"
2018-12-07 19:28:10 <jwh> hm
2018-12-07 19:28:19 <jwh> yeah its only to boot once I guess
2018-12-07 20:05:24 <squarebracket> SpaceToast: that's correct, that's what i use to update the certs. it does not work any more.
2018-12-07 20:05:50 <SpaceToast> can you define "does not work"?
2018-12-07 20:06:06 <SpaceToast> "my car doesn't work" is not very helpful to a mechanic, for instance
2018-12-07 20:06:23 <squarebracket> you're right, sorry. i mean wget borks with an ssl error
2018-12-07 20:06:57 <SpaceToast> so the ca-certificates package is installed, up to date, and you can run update-ca-certificates without errors?
2018-12-07 20:07:45 <squarebracket> that is correct, let me make a couple gists so i can show what's going on without flooding the channel
2018-12-07 20:10:21 <SpaceToast> does the error mention "ssl_helper"?
2018-12-07 20:14:12 <squarebracket> it mentions ssl_client. so this is what i was happily doing in alpine until recently, which now only works in alpine 3.4: https://gist.github.com/squarebracket/9d9d080080187a7fe98b58cc31086255
2018-12-07 20:14:30 <squarebracket> here is it not working in latest alpine: https://gist.github.com/squarebracket/e719069522436873bc6f13efb359cac9
2018-12-07 20:14:59 <squarebracket> i then found this bug, which sounds like it's making busybox compile against libressl instead of openssl: https://bugs.alpinelinux.org/issues/8917
2018-12-07 20:15:17 <squarebracket> however, installing it instead of openssl still doesn't work: https://gist.github.com/squarebracket/342ccdbb4fd9d43b964898bdca8cdd5a
2018-12-07 20:16:01 <SpaceToast> let me take a look on my system
2018-12-07 20:18:25 <SpaceToast> hmm, I'm getting `bad address 'intermediate-ca.vantrix.com'`
2018-12-07 20:19:09 <SpaceToast> my host has that too
2018-12-07 20:19:19 <SpaceToast> and all my VPS
2018-12-07 20:19:24 <SpaceToast> is this a private DNS entry?
2018-12-07 20:19:34 <squarebracket> yes
2018-12-07 20:19:40 <squarebracket> it's our internal corporate CA
2018-12-07 20:19:43 <SpaceToast> erm
2018-12-07 20:19:50 <SpaceToast> I can't exactly try to reproduce the issue then, can I?
2018-12-07 20:20:13 <squarebracket> i mean, you could generate your own self-signed cert
2018-12-07 20:20:35 <squarebracket> but unfortunately no, you can't reproduce with my exact cert
2018-12-07 20:20:35 <SpaceToast> eh
2018-12-07 20:21:11 <SpaceToast> so you're telling me to set up my own X.509 infrastructure and a webserver to debug your issue? ;)
2018-12-07 20:22:13 <SpaceToast> still, I suspect the issue is more with ssl_helper / busybox-wget than anything else
2018-12-07 20:22:15 <squarebracket> no, just generate some self-signed cert with a bogus DNS entry in your /etc/hosts. should be effectively the same
2018-12-07 20:22:37 <squarebracket> right.
2018-12-07 20:22:56 <squarebracket> but... busybox-wget should be getting its certs from somewhere, no?
2018-12-07 20:23:04 <SpaceToast> it uses ssl_helper
2018-12-07 20:24:15 <squarebracket> i'm not familiar with ssl_helper, is that part of libressl or something?
2018-12-07 20:25:06 <SpaceToast> no, it's a part of busybox
2018-12-07 20:25:16 <squarebracket> i see...
2018-12-07 20:25:19 <SpaceToast> it's linked against an ssl library and provides the tls layer for things like wget, nc, etc
2018-12-07 20:25:26 <squarebracket> ah, right
2018-12-07 20:25:39 <squarebracket> so then whatever ssl library it's linking against should be managing / validating the certs, yes?
2018-12-07 20:26:01 <SpaceToast> depends ¯\_(ツ)_/¯
2018-12-07 20:26:09 <SpaceToast> have you tried my suggestion yet?
2018-12-07 20:26:33 <SpaceToast> alternatively, wanna write a dockerfile I can just load-up that reproduces your issue, so I don't have to fumble around to even reproduce it to begin with?
2018-12-07 20:26:34 <squarebracket> maybe i missed it, what was your suggestion?
2018-12-07 20:27:02 <SpaceToast> update whole system (including busybox - there's an update waiting there), install both libressl and openssl (alongside ca-certificates)
2018-12-07 20:29:43 <kh405_tv_> Hey, anyone is running deluge on alpine? I'm getting some errors when starting the daemon
2018-12-07 20:30:22 <kh405_tv_> seedbox:/opt$ deluged seedbox:/opt$ [ERROR   ] 15:09:34 main:248 Error relocating /usr/lib/python2.7/site-packages/cryptography/hazmat/bindings/_openssl.so: SSL_CTX_set_psk_client_callback: symbol not found Traceback (most recent call last):   File "/usr/lib/python2.7/site-packages/deluge/main.py", line 240, in start_daemon     from deluge.core.daemon import Daemon   File "/usr/lib/python2.7/site-packages/deluge/core/daemon.
2018-12-07 20:30:31 <kh405_tv_> This is what I get
2018-12-07 20:30:42 <SpaceToast> kh405_tv_: looks like a result of our migration from libressl to openssl
2018-12-07 20:31:06 <SpaceToast> I think I recall something along these lines being disgussed a while ago
2018-12-07 20:31:16 <SpaceToast> either way, that's a semi-known issue, you can poke in #alpine-devel if you'd like
2018-12-07 20:31:26 <squarebracket> SpaceToast: don't you mean the opposite? from openssl to libressl?
2018-12-07 20:31:33 <SpaceToast> (mention py-cryptography and and openssl.so)
2018-12-07 20:31:44 <SpaceToast> squarebracket: no, libressl to openssl
2018-12-07 20:31:51 <SpaceToast> libressl keeps breaking ABI and removing useful features
2018-12-07 20:32:23 <squarebracket> i see.... i understood it was the reverse from https://bugs.alpinelinux.org/issues/8917
2018-12-07 20:32:33 <squarebracket> installing both didn't work btw
2018-12-07 20:32:38 <kh405_tv_> Umm, i'll go check it out, thanks!
2018-12-07 20:32:56 <SpaceToast> squarebracket: that's 7 months ago
2018-12-07 20:32:56 <squarebracket> it seems that only the apk wget is linked against the system ssl package
2018-12-07 20:33:03 <SpaceToast> the migration is... still happening now, to some degree
2018-12-07 20:33:19 <squarebracket> i see...
2018-12-07 20:33:29 <kh405_tv_> Why is it taking so long :O ?
2018-12-07 20:33:33 <SpaceToast> squarebracket: well, I can fetch https files with my busybox wget after installing libressl (since it hasn't been moved yet)
2018-12-07 20:33:41 <SpaceToast> kh405_tv_: mostly lack of manpower
2018-12-07 20:34:00 <SpaceToast> kh405_tv_: but again, the issue in question has nothing to do with the migration
2018-12-07 20:34:02 <kh405_tv_> Damn :S
2018-12-07 20:34:05 <SpaceToast> it just happens to mention libressl
2018-12-07 20:35:16 <kh405_tv_> So how should I go to fix my issue then ?
2018-12-07 20:36:11 <squarebracket> yeah. and the wget apk, which _does_ use the certs that ca-certificates touches, also links against libressl, not openssl. so... i don't know wtf. i guess i'll have to just see what's actually happening when requesting https in the alpine busybox wget
2018-12-07 20:37:23 <SpaceToast> kh405_tv_: you're affected by https://github.com/pyca/cryptography/issues/4588
2018-12-07 20:37:28 <SpaceToast> the current git version seems to be fixed
2018-12-07 20:37:38 <SpaceToast> so you want to update py-cryptography (the affected package)
2018-12-07 20:37:56 <SpaceToast> you might need to get it from edge, but don't take my word for it, seeing as I'm at work, and distracted
2018-12-07 20:38:08 <SpaceToast> my suggestion was to poke in #alpine-devel and mention your issue + py-cryptography
2018-12-07 20:38:11 <kh405_tv_> I'm from edge, and still having the issue
2018-12-07 20:41:42 <squarebracket> by the way, is there someone i can poke about the bug tracker? i tried registering so i could file a bug, but it's been a couple hours and i still haven't received an activation email. i wonder if the `+` i provided in the email is causing errors
2018-12-07 20:44:44 <kh405_tv_> I've got everything up to date on edge, and still having the issue, are you sure it has been fixed ?
2018-12-07 20:45:12 <SpaceToast> do you have libressl *and* openssl installed? (I'm not sure what the state of the transition is)
2018-12-07 20:45:40 <kh405_tv_> I might not have both
2018-12-07 20:49:11 <kh405_tv_> Which one should I try first?
2018-12-07 20:53:42 <joneskoo> what volumes are searched to find apkovl file?
2018-12-07 21:07:02 <kh405_tv_> tryed with both libressl and openssl still nada
2018-12-07 21:15:28 <joneskoo> https://github.com/alpinelinux/mkinitfs/blob/master/initramfs-init.in#L530 ok so nlplug-findfs finds it from all volumes, first match wins
2018-12-07 21:50:28 <johnnyfive> Howdy howdy! I'm running into this error when trying to install new software: "ERROR: libcrypto1.1-1.1.1a-r0: trying to overwrite etc/ssl/openssl.cnf owned by libressl-2.5.5-r2". What's the idiomatic way to solve these conflicts in alpine?
2018-12-07 21:56:36 <johnnyfive> Looks like it solved itself by changing the order I installed my dependencies, but I don't like that. I'd like to learn more about this error, if anyone has any insights. Thanks!
2018-12-08 03:30:16 <danieli> johnnyfive: tried apk upgrade / apk fix?
2018-12-08 03:30:21 <danieli> we don't use libressl anymore
2018-12-08 03:48:14 <johnnyfive> roger that danieli, i'm knee deep in the build system trying to resolve this. ty for the tip
2018-12-08 03:48:45 <danieli> the issue most likely isn't in the build system, but now you know where to look, at least
2018-12-08 04:50:46 <maldridge> curiousity, why did alpine switch off of libressl?
2018-12-08 04:51:04 <danieli> openssl has fixed many of the issues we had with it, and libressl isn't keeping up with new development
2018-12-08 04:52:04 <maldridge> interesting
2018-12-08 04:52:21 <danieli> there's a really long thread about it in alpine-devel
2018-12-08 11:49:19 <mpmc> To the sub to the mailing lists I just send a blank message right?
2018-12-08 15:53:01 <jomat> _ikke_: just fyi, recompiling py-cryptography didn't help and recompiling gajim doesn't work because of the missing symbol…
2018-12-08 15:54:14 <_ikke_> jomat: hmm
2018-12-08 16:13:26 <Processus42> Hello. I would want to know what package provides mkimage.sh
2018-12-08 16:21:21 <jomat> there's only a /usr/bin/mkimage (without .sh) in uboot-tools: https://pkgs.alpinelinux.org/contents?file=mkimage*
2018-12-08 16:21:35 <danieli> https://git.alpinelinux.org/cgit/aports/tree/scripts/mkimage.sh
2018-12-08 16:21:38 <danieli> probably looking for this
2018-12-08 16:32:25 <palate> hello :)
2018-12-08 16:34:46 <palate> do you guys know manylinux?
2018-12-08 16:35:04 <palate> (just curious, I don't have a specific question)
2018-12-08 16:35:50 <_ikke_> haven't heard of it
2018-12-08 16:38:14 <FreeBDSM> hello
2018-12-08 16:38:29 <FreeBDSM> could someone, please, backport wkhtmltopdf package from edge to at least 3.8?
2018-12-08 16:39:48 <FreeBDSM> ncopa: you are the maintainer of wkhtmltopdf packages for edge branch, could you backport them to 3.8?
2018-12-08 16:41:05 <_ikke_> FreeBDSM: The focus is mainly to get 3.9 out atm, would it work if it's part of 3.9?
2018-12-08 16:42:49 <FreeBDSM> _ikke_: well, I need that package in production, I can't rely on 'edge' branch as it is probably too breaky
2018-12-08 16:43:42 <FreeBDSM> _ikke_: how soon is 3.9 going to release?
2018-12-08 16:45:03 <_ikke_> ncopa can probably tell that better than I can
2018-12-08 16:47:23 <_ikke_> You can also build the package yourself against 3.8
2018-12-08 16:53:27 <danieli> ^
2018-12-08 16:53:57 <danieli> _ikke_: he said he's gonna be absent most of the upcoming week
2018-12-08 16:56:29 <_ikke_> so probably not this week :-P
2018-12-08 16:56:46 <danieli> I couldn't if I tried, no access
2018-12-08 16:56:58 <danieli> oh right, when 3.9 is releasing
2018-12-08 16:57:00 <danieli> "ASAP" ™️
2018-12-08 16:57:38 <_ikke_> Hey, it's trademarked to me :P
2018-12-08 16:58:01 <danieli> ©️ Natanael Copa
2018-12-08 17:05:34 <FreeBDSM> also, it will probably take some time after the release until it gets uploaded to docker's registry :(
2018-12-08 17:06:06 <danieli> I can make a patch for the docker repo and pass it to someone with access, hopefully ncopa keeps a little eye on Alpine while he's gone
2018-12-08 17:06:38 <danieli> I mean, I'm 99.99999% sure it won't be out until after he's back, but eh, my logic isn't always water proof
2018-12-08 17:07:10 <FreeBDSM> oh, god, wkhtmltopdf doesn't install even on edge!
2018-12-08 17:07:17 <danieli> broken?
2018-12-08 17:07:21 <FreeBDSM> `ERROR: unsatisfiable constraints:   wkhtmltopdf (missing):     required by: world[wkhtmltopdf]`
2018-12-08 17:07:41 <danieli> lol
2018-12-08 17:07:42 <danieli> the package is wkhtmltopdf
2018-12-08 17:07:46 <danieli> wait, what?
2018-12-08 17:07:56 <danieli> oh right, it's in testing - my bad, I'm tired
2018-12-08 17:08:05 <danieli> FreeBDSM: add the testing repository to /etc/apk/repositories
2018-12-08 17:08:19 <FreeBDSM> how?
2018-12-08 17:08:42 <_ikke_> add a mirror with /edge/testing to that file
2018-12-08 17:08:48 <FreeBDSM> thanks, got it
2018-12-08 17:09:12 <_ikke_> So it should first be moved from testing to community before it can even be part of a stable release
2018-12-08 17:09:49 <FreeBDSM> :(
2018-12-08 17:10:03 <danieli> indeed
2018-12-08 17:10:27 <_ikke_> FreeBDSM: Testing the package and providing feedback certainly would help
2018-12-08 17:10:54 <FreeBDSM> what can I do about it? I'm a CI/deploy engineer, so to speak, I don't want to switch my docker images to another OS because of that, but I do need wkhtmltopdf for production
2018-12-08 17:11:56 <_ikke_> FreeBDSM: If you need it now, I would build the package myself against 3.8
2018-12-08 17:12:03 <FreeBDSM> how?
2018-12-08 17:12:15 <danieli> I would like it to be available in stable sooner or later, it's a pretty commonly used tool
2018-12-08 17:12:40 <_ikke_> sure
2018-12-08 17:13:11 <FreeBDSM> https://github.com/wkhtmltopdf/wkhtmltopdf no build steps?
2018-12-08 17:13:14 <_ikke_> https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2018-12-08 17:13:39 <_ikke_> FreeBDSM: https://git.alpinelinux.org/cgit/aports/tree/testing/wkhtmltopdf/APKBUILD
2018-12-08 17:13:52 <FreeBDSM> oh, right
2018-12-08 17:14:08 <_ikke_> FreeBDSM: I have a Dockerfile to quickly setup a build container
2018-12-08 17:14:45 <_ikke_> FreeBDSM: http://tpaste.us/J6xv
2018-12-08 17:14:59 <_ikke_> instead of FROM alpine:edge, you use alpine:3.8
2018-12-08 17:15:26 <_ikke_> download the APKBUILD, put it in a directory, and run abuild -r
2018-12-08 17:16:02 <FreeBDSM> the result will be a .pkg archive?
2018-12-08 17:16:23 <_ikke_> .apk, yes
2018-12-08 17:16:33 <_ikke_> it will be placed in a directory in the home dir
2018-12-08 17:17:15 <_ikke_> The only thing you need to do is serve that directory from a webserver and you have a repository
2018-12-08 17:19:22 <FreeBDSM> _ikke_: why not automate it a bit more? add CMD/ENTRYPOINT so that it would just accept an url to APKBUILD as an arg, wget it, build by it and just spit an .apk?
2018-12-08 17:19:54 <_ikke_> FreeBDSM: Because I usaully use this to develop my APKBUILDS as well
2018-12-08 17:20:00 <FreeBDSM> that image could then uploaded to docker hub so that anyone could use it to quickly build a package for older branch if the package builds?
2018-12-08 17:20:17 <_ikke_> Feel free to change it to do that
2018-12-08 17:20:18 <danieli> that idea doesn't really scale at all
2018-12-08 17:20:24 <danieli> or so it seems in my head
2018-12-08 17:20:33 <FreeBDSM> doesn't scale?
2018-12-08 17:20:43 <danieli> I probably misread somewhere along the line, let's see
2018-12-08 17:21:14 <danieli> right, gotcha
2018-12-08 17:21:30 <FreeBDSM> ?
2018-12-08 17:21:39 <danieli> nevermind what I said, I misunderstood
2018-12-08 17:21:46 <FreeBDSM> oh, okay
2018-12-08 17:21:59 <danieli> either way, it's open source software, you can do anything you want
2018-12-08 17:22:19 <FreeBDSM> well, if I do it for myself - I'll solve mine problem, but there are others
2018-12-08 17:22:43 <danieli> a pull request and subsequent review could solve that part, if it's feasible :)
2018-12-08 17:25:26 <FreeBDSM> pull request to where? I don't know if it'd be okay by you, but I'm thinking that it'd be optimal, if said Dockerfile would be shared publicly: either advertised by pkgs.alpinelinux.org or even uploaded to hub.docker.com as an image by your official account, with some guide on how to use it
2018-12-08 17:26:35 <FreeBDSM> anyways, thank you for the info and Dockerfile!
2018-12-08 17:27:29 <FreeBDSM> _ikke_: is adding user that necessary?
2018-12-08 17:27:40 <danieli> we'll see, not everyone uses docker - but it could probably be written a wiki page on down the road, if it's created and accepted
2018-12-08 17:27:45 <_ikke_> FreeBDSM: yes, abuild by default refuses to run as root
2018-12-08 17:27:50 <danieli> also, the abuild repository @ github / git.alpinelinux.org
2018-12-08 17:28:25 <FreeBDSM> danieli: come on, alpine is made for docker :)
2018-12-08 17:28:32 <danieli> it isn't :P
2018-12-08 17:28:37 <_ikke_> alpine already existed before docker
2018-12-08 17:28:38 <danieli> it is popular for containers though
2018-12-08 17:28:46 <danieli> but that doesn't really change much
2018-12-08 17:31:10 <FreeBDSM> https://git.alpinelinux.org/cgit/aports/tree/testing/wkhtmltopdf/APKBUILD explain the last line, please
2018-12-08 17:31:20 <danieli> should be in the wiki page _ikke_ linked
2018-12-08 17:31:33 <FreeBDSM> why there are spaces and archive name after the hash?
2018-12-08 17:31:52 <danieli> uh?
2018-12-08 17:32:08 <FreeBDSM> sha512sums="3f982a6964085b36f62988b618d88d4339f03ad8d2e733c1fcba5e2e7cb545ccc99313ce35474605df818f91c289a87ddd952f21974fdbe846f990bf03fc2fdb  wkhtmltopdf-0.12.5.tar.gz"
2018-12-08 17:32:18 <danieli> that's the checksum for the source tarball
2018-12-08 17:32:33 <FreeBDSM> checksum is the part before the space
2018-12-08 17:32:45 <danieli> yes?
2018-12-08 17:32:59 <FreeBDSM> I mean there's more than checksum in that var, why so?
2018-12-08 17:33:14 <danieli> sha512sums="hash filename"
2018-12-08 17:33:21 <danieli> there's only one hash
2018-12-08 17:34:09 <FreeBDSM> oh, okay, so it's just that the format is so
2018-12-08 17:34:25 <danieli> it's the same format sha512sum outputs
2018-12-08 17:34:29 <danieli> I don't understand the problem?
2018-12-08 17:34:47 <FreeBDSM> why `wkhtmltopdf-0.12.5.tar.gz` instead of `$pkgname-$pkgver.tar.gz`?
2018-12-08 17:35:01 <danieli> because $pkgname is a variable, substituted for 'wkhtmltopdf' in this case
2018-12-08 17:35:03 <danieli> it's just a shell script
2018-12-08 17:36:01 <_ikke_> FreeBDSM: abuild checksum generates it anyway
2018-12-08 17:36:10 <FreeBDSM> yeah, so the result would be the same, but next time the maintainer would only need to change the left part, ah, nvm, got it
2018-12-08 17:36:11 <danieli> yup, you don't do that part manually
2018-12-08 17:36:17 <danieli> it's all in the wiki page
2018-12-08 17:36:21 <danieli> https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2018-12-08 17:36:34 <_ikke_> FreeBDSM: note that there can be more than one source file
2018-12-08 17:36:47 <FreeBDSM> abuild checksum
2018-12-08 17:36:51 <FreeBDSM> yeah, got it, sorry
2018-12-08 20:18:20 <FreeBDSM> wellp, another package is outdated: my application throws an error: `ImageMagick version 7.0.7-32 has known bugs that prevent it from working` and it is the latest available in the 3.8 branch of the repo :(
2018-12-08 20:19:05 <_ikke_> What version doesn't have these bugs?
2018-12-08 20:19:41 <jaroslaw> hi there
2018-12-08 20:19:51 <jaroslaw> aer there any java IDE on alpine ?
2018-12-08 20:20:12 <FreeBDSM> _ikke_: I guess latest release would be a good guess: 7.0.8-15
2018-12-08 20:20:15 <jaroslaw> like netbeans, eclipse or inetllij
2018-12-08 20:22:02 <_ikke_> jaroslaw: don't see anything packaged
2018-12-08 20:22:33 <danieli> probably not, but you're welcome to package it
2018-12-08 20:23:04 <jaroslaw> :]
2018-12-08 21:14:40 <FreeBDSM> does alpine repo store older versions of packages?
2018-12-08 21:15:02 <_ikke_> nope
2018-12-08 22:17:02 <joneskoo> Is there a way to create apk cache for raspberry aarch64 without running on a raspberry? I'm looking to automate build of SD card image.
2018-12-08 22:30:59 <maldridge> joneskoo: I'm not sure if there's a more idiomatic way on alpine, but qemu static will allow you to run the aarch64 binaries on an x86 platform
2018-12-08 23:06:08 <sphrak> hey, im on the 3.8 docker image. trying to install musl and musl-dev v1.1.20-r2 but when I specify the --repository to edge/main it still pulls the 1.1.19-r10 version.. what might I be doing wrong?
2018-12-08 23:06:50 <danieli> you shouldn't mix edge and stable, for what it's worth
2018-12-08 23:06:50 <sphrak> I have checked the mirror im using aswell and it lists the correct version
2018-12-08 23:07:37 <danieli> if you mix edge and stable, your system will most likely break, unless you know what you're doing
2018-12-08 23:07:42 <danieli> anyway, paste your /etc/apk/repositories please?
2018-12-08 23:09:30 <sphrak> hmm right, im doing it to get access to pg11. one moment its on this other machine
2018-12-08 23:12:00 <sphrak> danieli: its just these two, https://dpaste.de/h0gC
2018-12-08 23:12:28 <sphrak> but I run the installation command with '--repository https://dl-4.alpinelinux.org/alpine/edge/main/'
2018-12-08 23:13:36 <danieli> try package pinning, or just adding it to /etc/apk/repositories
2018-12-08 23:14:02 <danieli> if you have to mix edge and stable (NOT recommended), package pinning is the way to go
2018-12-08 23:15:20 <sphrak> danieli: that is not to be confused with repository pinning described in the wiki right?
2018-12-08 23:15:34 <danieli> that should be the one I'm aiming at, why?
2018-12-08 23:17:56 <sphrak> Am I not already doing that by saying apk add <pkg> --repository no?
2018-12-08 23:19:19 <danieli> try apk add -U --repository X
2018-12-08 23:19:43 <danieli> if that works, it's odd, but not impossible
2018-12-08 23:20:47 <sphrak> if -U is upgrade (--help doesnt tell) then it does not work. I will try repo pinning
2018-12-08 23:20:53 <danieli> -U is ignore cache
2018-12-08 23:21:06 <danieli> or to update it, rather
2018-12-08 23:22:48 <sphrak> -U did not work, will give repo pinning a go
2018-12-08 23:23:22 <danieli> I can't remember the specific differences between --repository and package pinning, I'm running low on sleep and my head is filled with math problems
2018-12-08 23:24:34 <sphrak> mate you should get some sleep then, gotta service that brain
2018-12-08 23:33:36 <sphrak> danieli: Even via repo pinning it pulls the old package so I assume its just some compat issue. The pinning stuff was nice tho, reduced alot of lines actually so thanks for that and sleep well :-)
2018-12-08 23:33:42 <danieli> hm..
2018-12-08 23:34:11 <danieli> I'm not sure why it's pulling the old package (from 3.8) rather than from edge
2018-12-08 23:34:17 <danieli> it's easier to just use edge
2018-12-08 23:35:29 <sphrak> The reason im even trying this is to test if it solves a problem with python 3.6.7 described here https://github.com/docker/compose/issues/6292 (last post)
2018-12-08 23:35:46 <sphrak> but I can run 3.6.6 meanwhile so no biggie, just wanted to know if it actually fixes it
2018-12-08 23:35:57 <danieli> try alpine:edge, see if it works there
2018-12-08 23:40:32 <sphrak> danieli: that does pull the correct version but that would require a bit of rewriting of the docker image im using
2018-12-08 23:41:44 <sphrak> must compat issue tho, since I can pull postgres 11 from edge/main no problem
2018-12-08 23:41:52 <sphrak> must be a* sorry
2018-12-09 11:42:50 <TwistedFate> hello all
2018-12-09 11:43:07 <_ikke_> zo/
2018-12-09 11:43:42 <TwistedFate> i am looking into alpine linux, can someone tell me what are the main differences between busybox vs coreutils and musl vs glibc? in few short lines :)
2018-12-09 11:44:45 <jn__> busybox doesn't implement as many features as the corresponding parts of coreutils
2018-12-09 11:45:28 <jn__> musl is much smaller in code size, and thus more readable than glibc, and also stricter wrt. POSIX compliance
2018-12-09 11:46:26 <jn__> by code size i meant source code size, but it probably also applies to the size of the compiled libc.so
2018-12-09 11:50:51 <mps> jn__: on Debian stretch: -rwxr-xr-x 1 root root 1.7M Jan 14  2018 lib/x86_64-linux-gnu/libc-2.24.so
2018-12-09 11:51:30 <mps> -rwxr-xr-x 1 root root 647K Mar  9  2017 /lib/x86_64-linux-musl/libc.so
2018-12-09 11:51:51 <jn__> 
ACTION nods

2018-12-09 11:52:47 <mps> musl is just 30% of the gnu libc
2018-12-09 11:54:41 <mps> never benchmarked them but my feeling the ratio in the speed is the same
2018-12-09 12:00:17 <TwistedFate> can i play video games on alpine like i can on say debian?
2018-12-09 12:00:27 <TwistedFate> or is musl/busybox less supported?
2018-12-09 12:01:57 <AinNero> TwistedFate: less supported
2018-12-09 12:02:09 <AinNero> there are some compatibility stuff that works for things like OpenArena
2018-12-09 12:02:37 <jn__> first distinction: open source games or proprietary games?
2018-12-09 12:03:13 <jn__> on the open source side, i'm pretty sure that debian has more games packaged
2018-12-09 12:03:19 <TwistedFate> proprietary ofc
2018-12-09 12:03:30 <TwistedFate> there are almost no good open source games unfortunately :/
2018-12-09 12:03:31 <jn__> proprietary games expect glibc
2018-12-09 12:04:10 <jn__> there is a compatibility shim called gcompat that makes a musl-based system like alpine look as if it had glibc
2018-12-09 12:05:45 <TwistedFate> hmm so alpine is best suited for workstations and enterprise :/
2018-12-09 12:07:45 <jn__> it is somewhat server/container focused, yes
2018-12-09 12:08:01 <mps> imo, Alpine is best suited for servers and appliances although some of us use it on workstations
2018-12-09 12:08:45 <jn__> then again, there are the postmarketOS people who (try to) use alpine plus additional packages on smart phones
2018-12-09 12:09:24 <jn__> (but postmarketOS is a research project right now)
2018-12-09 15:54:47 <mps> does musl support GB2312 encoding
2018-12-09 17:02:30 <kupi> hi
2018-12-09 17:02:44 <kupi>  /dev/dri folder is missing on my alpine
2018-12-09 17:02:55 <kupi> I have ThinkPad A21p
2018-12-09 19:50:11 <Hamilton> So I installed alpine on VBox. (yaay?!)...I'm kinda n00b. Should I continue with root name?
2018-12-09 20:01:25 <Hamilton> If not, why it is not mentioned in Wiki that I should change my user to non-root?
2018-12-09 20:07:31 <Nicop06> Hamilton: What is root name? what step of the install are you in?
2018-12-09 20:07:48 <Hamilton> Nicop06, I already installed
2018-12-09 20:08:03 <Hamilton> I mean username is root. people suggest that I should change it
2018-12-09 20:08:15 <Nicop06> Yes it's true
2018-12-09 20:08:35 <Nicop06> you can use adduser
2018-12-09 20:08:37 <Hamilton> useradd is not avaible
2018-12-09 20:08:58 <Hamilton> oh it's adduser
2018-12-09 20:09:17 <Nicop06> yes alpine is using busybox so no adduser
2018-12-09 20:09:28 <Nicop06> useradd is the busybox replacement
2018-12-09 20:10:35 <Nicop06> might need to look at `adduser --help` before as I am unsure if it creates the home directory and prompt for password
2018-12-09 20:11:49 <Hamilton> So I'm in VBox which my main is Windows 7 .I'm gonna mainly use Alpine for programming C/Rust/C++ maybe...I'm kinda familiar with vim. I'm happy about the performance and keyboard latency in the current setup. If I go Xorg and i3wm, would that hurt the current CPU/RAM usage much?
2018-12-09 20:14:55 <Nicop06> I think this will increase your memory consuption by a little more than 100MB so it should be fine
2018-12-09 20:15:35 <brebs> Hamilton: Define "much", i.e. what's significant to you, what's your expectation. Alpine is lightweight.
2018-12-09 20:15:39 <Nicop06> As for the CPU it shouldn't be a problem if you have a CPU with hardware virtualization acceleration (needs to be activated in the BIOS on most PC)
2018-12-09 20:16:28 <Hamilton> brebs, Well I have manjaro+Xfce and there is noticeable latency in inputting words in terminal
2018-12-09 20:16:43 <Hamilton> under Vbox also
2018-12-09 20:19:16 <Nicop06> Unless you swap, this might be due to the GPU
2018-12-09 20:19:32 <Hamilton> What is you guys setup if you use Alpine as daily driver?
2018-12-09 20:19:49 <Nicop06> For the CPU just make sure hardware virtualization extensions are activated
2018-12-09 20:20:09 <Nicop06> I use Alpine only on servers, sorry
2018-12-09 20:20:58 <Hamilton> Nicop06, :)
2018-12-09 20:21:51 <brebs> Hamilton: Try the Manjaro forums (if running Manjaro as the host OS). My first suspicion would be e.g. a kernel module tweak required due to specific weird hardware involved.
2018-12-09 20:23:17 <Hamilton> I'm gonna try alpine for a few weeks. So is it weird if I use tmux+vim in TTY mode and only use it in TTY mode?
2018-12-09 20:23:54 <Hamilton> And what is Alpine's terminal?
2018-12-09 20:24:03 <Hamilton> Terminal emulator app I mean
2018-12-09 20:26:13 <brebs> In my opinion, sticking to TTY mode is a bit weird, why not use all the advantages of a GUI?
2018-12-09 20:26:31 <Hamilton> I can't reboot in my account name
2018-12-09 20:26:38 <Hamilton> brebs, sudo is not available?
2018-12-09 20:27:15 <brebs> The terminal emulator app, as in other general-purpose linuxes, depends on the desktop environment used, plus the usual desktop-agnostic ones.
2018-12-09 20:27:52 <pardis> If you want sudo, install sudo
2018-12-09 20:28:06 <brebs> apk add sudo, it's that simple ;)
2018-12-09 20:28:19 <Hamilton> But how can I reboot?
2018-12-09 20:28:43 <Hamilton> I should first install sudo then `sudo reboot`?
2018-12-09 20:28:53 <pardis> Or just su to root and then reboot
2018-12-09 20:28:55 <brebs> Use the "reboot" command - am I missing something? Do you mean from the GUI?
2018-12-09 20:28:56 <pardis> It's up to you
2018-12-09 20:29:11 <Hamilton> brebs, I can't reboot in my non-root user from TTY
2018-12-09 20:29:18 <Hamilton> It's weird
2018-12-09 20:30:00 <brebs> That's by design, usually only the root user can reboot.
2018-12-09 20:30:39 <brebs> You can set up sudo, so that your non-root user is allowed to reboot. For example.
2018-12-09 20:46:54 <Hamilton> So I ran setup-xorg and apk add i3wm... still i3 does not start
2018-12-09 20:51:40 <brebs> It's not meant to start, automatically. You have to set it up to start. E.g. set up ~/.xinitrc
2018-12-09 20:52:51 <Hamilton> brebs, https://wiki.alpinelinux.org/wiki/VirtualBox_guest_additions
2018-12-09 20:53:06 <Hamilton> I added testing repo to apk. But I'm still getting the error mentioned there
2018-12-09 20:53:19 <Hamilton> I assume I need oracleVBox video drivers for i3
2018-12-09 20:55:37 <brebs> If you're trying to start xorg and it's not starting, then look in /var/log/Xorg.0.log for diagnostics.
2018-12-09 20:58:45 <Hamilton> brebs, Shouldn't I instal a "video package" first? https://wiki.alpinelinux.org/wiki/XFCE_Setup
2018-12-09 21:18:02 <brebs> Sounds right. However, going back to the start, if you have a weird latency delay when running Manjaro in Virtualbox, then there's no guarantee that running Alpine instead will magically fix that delay. Maybe worth a try anyway.
2018-12-09 21:24:43 <Hamilton> https://unsee.cc/f50e877c/
2018-12-09 21:24:51 <Hamilton> startx does not start Xrg
2018-12-09 21:25:17 <Hamilton> I followed a combination of XFCE and awsome-wm Wiki on Alpine
2018-12-09 21:32:07 <Hamilton> http://termbin.com/nsey
2018-12-09 21:45:15 <danieli> this might be a dumb question, but did you install any drivers?
2018-12-09 21:47:53 <Hamilton> danieli, vmware drivers for Video
2018-12-09 21:58:30 <danieli> Hamilton: xf86-video-vmware?
2018-12-09 21:58:37 <Hamilton> danieli, yes
2018-12-09 21:58:51 <danieli> hm, i'm not seeing anything about it in the xorg log, at all
2018-12-09 21:59:14 <Hamilton> I just did `apk add xf86-video-vmware` as suggested here
2018-12-09 21:59:20 <Hamilton> https://wiki.alpinelinux.org/wiki/XFCE_Setup
2018-12-09 21:59:42 <Hamilton> But after that not working, I also added the sis one. Still doesn't work (after reboot)
2018-12-09 22:00:04 <danieli> hm, I'm not sure, I haven't messed too much with alpine and X myself
2018-12-09 22:01:22 <Hamilton> VESA worked!
2018-12-09 22:07:24 <Hamilton> i3wm says something is messing. exiting with 127...also mouse is slow :(
2018-12-09 22:07:31 <Hamilton> I should give it a go another time
2018-12-09 22:08:59 <danieli> Hamilton: does it say exit code 127 in the status bar?
2018-12-09 22:09:02 <danieli> or does i3 just not start?
2018-12-09 22:09:08 <Hamilton> status bar
2018-12-09 22:09:18 <danieli> what version of i3status?
2018-12-09 22:09:31 <danieli> 127 means status_command not found
2018-12-09 22:10:07 <Hamilton> danieli, yeah "status_command not found"
2018-12-09 22:10:14 <danieli> did you install i3status?
2018-12-09 22:10:25 <danieli> if so, which version?
2018-12-09 22:10:30 <Hamilton> no i did not
2018-12-09 22:10:45 <danieli> then it's no wonder it doesn't find it :)
2018-12-09 22:10:54 <Hamilton> Why wasn't it a dependency of i3wm in the firtst place?
2018-12-09 22:11:07 <danieli> well, i3 supports different status bars and status bar data providers
2018-12-09 22:11:16 <danieli> i use i3blocks myself, moving over to polybar soon enough
2018-12-09 22:11:32 <danieli> (i3bar with i3blocks, and soon replacing i3bar entirely with polybar)
2018-12-09 22:11:54 <Hamilton> Shoudn't Alt+Enter bringup terminal?
2018-12-09 22:12:07 <danieli> Hamilton: https://build.i3wm.org/docs/i3-sensible-terminal.html
2018-12-09 22:12:25 <Hamilton> Also Win+Enter changes the cursor to wait icon
2018-12-09 22:12:36 <Hamilton> indefinitly it seems
2018-12-09 22:12:39 <danieli> do you have one of the terminals listed in that page installed?
2018-12-09 22:12:50 <Hamilton> oh no
2018-12-09 22:13:01 <Hamilton> So alpine doesn't come with a terminal
2018-12-09 22:13:07 <danieli> it doesn't come with a gui by default
2018-12-09 22:13:08 <Hamilton> The one I was usign was TTY
2018-12-09 22:13:17 <danieli> just curious, have you used i3 before?
2018-12-09 22:13:23 <Hamilton> no :) Xfce
2018-12-09 22:13:30 <danieli> xfce + i3?
2018-12-09 22:13:34 <Hamilton> no Xubuntu
2018-12-09 22:13:45 <danieli> aha
2018-12-09 22:13:50 <Hamilton> Imma go hardcore with this one
2018-12-09 22:13:54 <Hamilton> Alpine+i3 under VBox
2018-12-09 22:13:59 <danieli> well, alpine *is* made for superusers, and doesn't install any unnecessary deps, you'll have to add them yourself
2018-12-09 22:14:45 <danieli> I'm mostly on Arch + i3, occasionally OpenBSD + cwm or i3
2018-12-09 22:14:53 <danieli> done a lot of different combos through the years
2018-12-09 22:15:13 <danieli> anyway, the golden rule still applies, google is your best friend
2018-12-09 22:15:17 <danieli> https://pkgs.alpinelinux.org/packages is also very helpful
2018-12-09 22:15:37 <Hamilton> danieli, The mouse seem to be laggy
2018-12-09 22:15:47 <Hamilton> Is it because of VESA?
2018-12-09 22:16:02 <danieli> I don't know
2018-12-09 22:16:22 <danieli> it could be a number of things
2018-12-09 22:47:18 <Galaxor> Hi.  I'm having problems installing gdal-dev from the testing repo.  I suspect a bug in the package.  Is there a bugtracker for packages that I can look at?
2018-12-09 22:58:39 <danieli> Galaxor: bugs.alpinelinux.org
2018-12-09 22:58:42 <danieli> what's the issue?
2018-12-10 00:07:32 <Galaxor> danieli: apk add --no-cache --update    --repository http://dl-cdn.alpinelinux.org/alpine/edge
2018-12-10 00:07:38 <Galaxor> gdal-dev
2018-12-10 00:08:07 <Galaxor> leads to unsatisfiable constraints: so:libcrypto.so.1.1 (missing)
2018-12-10 00:08:20 <Galaxor> It looks like only libcrypto1.0 is available in the repos.
2018-12-10 00:08:37 <danieli> right, we moved from libressl to openssl 1.1
2018-12-10 00:09:04 <Galaxor> So, does the gdal package need to be informed of this somehow?
2018-12-10 00:09:07 <danieli> mixing edge and stable is not supported, and things are likely to break
2018-12-10 00:09:48 <Galaxor> danieli: Ohh, I see.  So I'll have to install a system that's all edge in order for this to work.
2018-12-10 00:09:57 <Galaxor> danieli: gdal doesn't seem to be in stable at all.
2018-12-10 00:10:05 <danieli> the testing repository is only in edge
2018-12-10 00:10:32 <danieli> you *could* pull the gdal APKBUILD, adapt it to use libressl again (undo the openssl change) and build it against 3.8
2018-12-10 00:11:41 <Galaxor> danieli: I was building a docker image based on python:3-alpine, but I wanted to have gdal on there.  I could redo my Dockerfile to be based on alpine:edge instead.
2018-12-10 00:12:03 <danieli> edge is the (rolling) development branch, so beware
2018-12-10 00:12:28 <Galaxor> danieli: Yeah, it might be better to continue to use python:3-alpine, but build gdal from source.  I dunno.
2018-12-10 00:12:54 <Galaxor> danieli: So, the problem seems to be that gdal isn't in the stable repo.  Will it be there at some point in the future?
2018-12-10 00:13:39 <danieli> iirc when it's confirmed to work and not break horribly, the maintainer could request it
2018-12-10 00:14:01 <Galaxor> Is there anything I could do to speed the process along?  Bug the maintainer?
2018-12-10 00:14:16 <danieli> you could make an issue @ bugs.alpinelinux.org
2018-12-10 00:14:53 <Galaxor> Aha!  I was planning on doing that anyway, but now that I've talked to you, my bug will make sense and contain good info!
2018-12-10 00:24:07 <danieli> issues get posted in our development channel too
2018-12-10 13:07:54 <pax_rhos> hello
2018-12-10 13:25:11 <pax_rhos> php7-imagick package depends on imagemagick. and, as you know, imagemagick version available in alpine is broken. what should alpine users do about that?
2018-12-10 13:29:50 <clandmeter> broken as in?
2018-12-10 13:32:12 <pax_rhos> it doesn't generate previews
2018-12-10 13:32:37 <pax_rhos> anyways, I've built imagemagick 7.0.8 and installed it above existing imagemagick
2018-12-10 13:33:22 <pax_rhos> now `apk list | imagemagick-7` lists 2 packages (7.0.8.15-r1 as installed and 7.0.7.32-r1 without '[installed]')
2018-12-10 13:37:35 <pax_rhos> nvm, figured that out
2018-12-10 14:40:39 <pax_rhos> well, not really
2018-12-10 14:40:57 <pax_rhos> the problem is that now I get a PHP Warning: `Version warning: Imagick was compiled against Image Magick version 1799 but version 1800 is loaded. Imagick will run but may behave surprisingly in Unknown on line 0`
2018-12-10 14:41:15 <pax_rhos> how to resolve this?
2018-12-10 15:20:24 <Mo0O> hi there
2018-12-10 15:20:47 <Mo0O> do you know how to use ~/.ashrc ?
2018-12-10 15:20:57 <Mo0O> it looks like it's not sourced
2018-12-10 15:25:55 <_ikke_> Mo0O: You should usr .profile afaik
2018-12-10 15:27:51 <sylvblck> greetings o/
2018-12-10 15:31:43 <sylvblck> I'm trying to install alpine standard on an APU2, which only has a serial connection.
2018-12-10 15:32:44 <sylvblck> Is there a iso where ttyS0 console is enabled or do i have to build it myself?
2018-12-10 15:39:30 <pax_rhos> someone who knows how to build packages, please, help
2018-12-10 15:41:17 <kahiru> sylvblck: it seems the virt image has serial enabled by default, but that is probably not the right one for a real hw
2018-12-10 15:42:12 <kahiru> sylvblck: however rebuilding isos with alpine is quite easy
2018-12-10 15:42:18 <sylvblck> i noticed it too but the ethernet driver is missing
2018-12-10 15:43:22 <Mo0O> _ikke_: looks like ~/.profile and /etc/profile are not sourced on no login shell, I've tryied to set ENV in /etc/profile but it still doesn't do the tricks
2018-12-10 15:43:25 <Mo0O> still digging
2018-12-10 15:44:03 <sylvblck> kahiru: i will give it a second try.
2018-12-10 15:44:15 <sylvblck> is it possible in docker ?
2018-12-10 15:44:23 <kahiru> it should be
2018-12-10 15:44:43 <Mo0O> s/no login/non-login/
2018-12-10 15:46:00 <sylvblck> i already tried it by modifying the standard aport but it end in a rescue shell :/
2018-12-10 15:46:41 <sylvblck> kahiru thanks for the help
2018-12-10 15:47:24 <kahiru> sylvblck: what were the instructions you used to build a custom iso?
2018-12-10 15:48:33 <sylvblck> kahiru: https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage
2018-12-10 15:49:11 <sylvblck> i took the standard entry and apended the cmdline
2018-12-10 15:49:30 <sylvblck> console=ttyS0,115200
2018-12-10 15:50:07 <Kanarok> has NE1 managed to get usb otg running on alpine?
2018-12-10 15:50:25 <Kanarok> with the raspi zero w, that is
2018-12-10 15:57:50 <pax_rhos> whom to whine to about updating a package for alpine repos?
2018-12-10 15:59:12 <kahiru> the issue tracker probably
2018-12-10 16:00:59 <pax_rhos> it has the smallest 'register' link evar
2018-12-10 16:04:25 <kahiru> sylvblck: failed with grep /sysroot/etc/inittab ?
2018-12-10 16:04:45 <sylvblck> exactly
2018-12-10 16:04:55 <sylvblck> you know the reason ?
2018-12-10 16:05:04 <kahiru> nah, just tried building and booting an image
2018-12-10 16:05:25 <sylvblck> so you have the same problem i guess
2018-12-10 16:06:11 <sylvblck> maybe it works with a second console entry
2018-12-10 16:06:53 <kahiru> I'd be surprised
2018-12-10 16:13:11 <blaman> will any base alpine stuff (libc, etc) care if I put an fqdn in /etc/hostname, as opposed to a hostname?
2018-12-10 16:15:49 <kahiru> anyone, if I rebuild an alpine standard image, it fails to boot (fails with /sbin/init not found in new root). is this a known issue?
2018-12-10 16:18:43 <kahiru> brb in 20
2018-12-10 16:23:16 <pax_rhos> http://www.imagemagick.org/download/releases/ 7.0.7-32 isn't even listed! please, update it to at least 7.0.7-39!
2018-12-10 16:54:38 <Mo0O> none of the variable exported from /etc/profile are available for non-login shell
2018-12-10 16:54:48 <Mo0O> wierd not? :D
2018-12-10 16:55:54 <Mo0O> which means that even after setting ENV it's not available when invoking a non-login shell
2018-12-10 16:56:30 <Mo0O> that's not the behavior describ in ash man page: https://linux.die.net/man/1/ash
2018-12-10 16:56:42 <Mo0O> described*
2018-12-10 16:57:14 <Mo0O> "If the environment variable ENV is set on entry to a shell, or is set in the .profile of a login shell, the shell next reads commands from the file named in ENV."
2018-12-10 17:58:10 <misterunknown> Hi guys. My icinga2 fails to start, because /var/cache/icinga2 is only read- and writable by root. I can't change that, because it gets overridden by something. It's not /usr/lib/icinga2/prepare-dirs, because I fixed it there, too. Any ideas?
2018-12-10 18:34:11 <quentinsf> Apologies for newby question.  Trying to set up alpine on a separate /dev/sdb disk. Either setup-alpine or setup-disk fail after partitioning with ERROR: unsatisfiable constraints:
2018-12-10 18:34:12 <quentinsf>   .setup-apkrepos (missing):
2018-12-10 18:34:12 <quentinsf>     required by: world[.setup-apkrepos]
2018-12-10 18:34:27 <geosmin> why does /etc/hostname get created at boot despite it not being in lbu include, nor hostname (the command) creating it when set by /etc/init.d/hostname
2018-12-10 18:34:30 <quentinsf> I can see references to this problem online, but no guidance or solutions
2018-12-10 19:25:01 <misterunknown> Is anyone running an icinga instance on alpine edge?
2018-12-10 19:26:02 <geosmin> ncopa: regarding the hostname variable bug from last week, i hadn't tested through an actual reboot. though both the init.d script and conf.d script work as expected this is made moot by /etc/hostname being included and populated at boot by alpine-baselayout
2018-12-10 22:52:46 <Keij0> Hi, how do I spoof mac address permanently, can I do that in /etc/network/interfaces?
2018-12-10 23:06:10 <nwe> has someone try https://hub.docker.com/r/syncomm/spotify/ under alpine-linux and got it to work?
2018-12-11 00:02:22 <maldridge> I'm missing something obvious, apk add shadow fails wiht this being missing, but I can see it present in the web search for 3.8, is there something I'm missing with installing this package?
2018-12-11 00:05:18 <KH405_TV> I know I asked before, but i'll reask cause i never found the answer, I'm trying to start Deluged daemon and getting this error
2018-12-11 00:05:38 <KH405_TV> main:248 Error relocating /usr/lib/python2.7/site-packages/cryptography/hazmat/bindings/_openssl.so: SSL_CTX_set_psk_client_callback: symbol not found
2018-12-11 00:05:38 <KH405_TV> Traceback (most recent call last):
2018-12-11 00:05:38 <KH405_TV>   File "/usr/lib/python2.7/site-packages/deluge/main.py", line 240, in start_daemon
2018-12-11 00:05:38 <KH405_TV>     from deluge.core.daemon import Daemon
2018-12-11 00:05:45 <KH405_TV> And a buch of other modules failling
2018-12-11 00:05:53 <KH405_TV> Anyone has any clue where to look for ?
2018-12-11 00:15:40 <maldridge> KH405_TV: offhand your python bindings seem to be built against a version of openssl you don't have
2018-12-11 00:16:35 <KH405_TV> Anyway to fix that? Not familiar with Python too much
2018-12-11 00:19:18 <maldridge> I can't say for alpine
2018-12-11 00:19:24 <maldridge> that's beyond my knowledge right now
2018-12-11 00:21:55 <KH405_TV> Damn, i'll try to keep looking, if you've got any ideas, shoot them my way, ill try them
2018-12-11 10:28:23 <Kanarok> anyone playing with alpine and raspberry pi otg stuff?
2018-12-11 14:12:44 <essence> when contributing with new packages, is there any disadvantage in sending the patches to the mailing list instead of using Github PRs?
2018-12-11 14:15:41 <essence> idk if there's any preference from the Alpine Linux maintainers, I'm asking that because I sent to the mailing list a new package (e2tools) but idk where to see any automated build/test logs, or if there aren't any feedbacks when sending them to the mailing list
2018-12-11 14:17:43 <clandmeter> essence, there is no preference in general.
2018-12-11 20:23:40 <Hamilton> So about my problem of 2 days ago: I have Alpine on Vbox but mouse lags and VMware graphics doesn't seem to work
2018-12-11 20:23:50 <Hamilton> I have these video drivers installed http://termbin.com/jsmo
2018-12-11 20:24:09 <Hamilton> Until I installed VESA, X wasn't even starting up
2018-12-11 20:32:33 <Hamilton> So here are my video drivers : http://termbin.com/6to0 . none of them were working until I installed VESA
2018-12-11 20:33:22 <Hamilton> And here are all my current packages FWIW: termbin.com/al2q
2018-12-11 20:43:31 <Keij0> Is this chat alive?
2018-12-11 20:43:43 <_ikke_> Keij0: yes, but atm it's quiet
2018-12-11 20:43:50 <danieli> yes, but it can be slow
2018-12-11 20:44:04 <Keij0> https://wiki.alpinelinux.org/wiki/Wireless_AP_with_udhcpd_and_NAT I'm trying to use this
2018-12-11 20:44:05 <_ikke_> danieli: You need to up your game :P
2018-12-11 20:44:13 <danieli> 7 seconds late, darn it
2018-12-11 20:44:40 <danieli> Keij0: and what's the issue...?
2018-12-11 20:44:47 <Keij0> I have 3 NICs, one is an entry point, the other ones have to be routed
2018-12-11 20:45:25 <Keij0> I got SSH connection with a static IP from the ethernet NIC that is supposed to be routed
2018-12-11 20:45:34 <Keij0> and I could ping it from wifi that's supposed to be routed
2018-12-11 20:45:41 <Keij0> but those iptables don't work
2018-12-11 20:46:00 <Keij0> I can access the internet from the host that is supposed to route everything, but nothing goes through
2018-12-11 20:46:09 <Keij0> had the same on OpenBSD and nobody helped me lmao
2018-12-11 20:46:57 <Keij0> it is Fujitsu Futro A300 with VIA Eden ULV onboard + ath9k NIC + tp-link ethernet NIC, the internet comes in through the onboard VIA NIC with a spoofed mac
2018-12-11 20:48:07 <danieli> i mean, there's always ##networking if you don't get help, this is a fairly OS-agnostic issue
2018-12-11 20:48:42 <Keij0> those iptables should work tho, at least for the wifi :/
2018-12-11 20:48:58 <danieli> anyway, you mind elaborating/rephrasing a bit? it might just be me being tired, but I don't really understand what you meant
2018-12-11 20:49:48 <danieli> specifically "SSH connection with a static IP from the ethernet NIC" and "access the internet from the host that is supposed to route everything, but nothing goes through"
2018-12-11 20:50:59 <Keij0> I couldn't set up dhcp so I just set it up statically for the moment, the guide recommends udhcpd which doesn't I couldn't find in packages
2018-12-11 20:51:32 <Keij0> the host I described has internet access and the hosts in internal subnets don't get routed outside
2018-12-11 20:51:50 <Keij0> with the iptables rules that are in that guide
2018-12-11 20:51:53 <danieli> udhcpd is part of busybox
2018-12-11 20:53:09 <Keij0> Isn't busybox shipped by default?
2018-12-11 20:53:27 <danieli> it is
2018-12-11 20:53:32 <Keij0> either way, it should work without dhcp too
2018-12-11 20:53:53 <danieli> yup, assuming all hosts are in the same segment, with an IP in the same subnet
2018-12-11 20:55:11 <Keij0> I made wifi and the second ethernet different subnets because I thought that if I replicated an IP on both interfaces it would cause conflicts
2018-12-11 20:55:20 <brebs> Keij0: tcpdump is a great way to see the traffic flowing, for debugging.
2018-12-11 20:58:08 <Keij0> I'll probably do it next weekend because my dormmate needed that cable I want to use as an entrypoint eventually, just idk how to debug it when the NIC that pushes the traffic further is spammed with random garbage from that network
2018-12-11 20:58:26 <danieli> ignore that interface
2018-12-11 20:59:18 <Keij0> which should I scan, the ones I want to be routed? if it doesn't reply it's noting interesting probably
2018-12-11 20:59:37 <Hamilton> Another problem is I can't install the guest addition: http://termbin.com/4sdo
2018-12-11 20:59:42 <Hamilton> https://wiki.alpinelinux.org/wiki/VirtualBox_guest_additions
2018-12-11 20:59:46 <Keij0> on OpenBSD they were coming out to the network, coming back but they didn't get passed on to the target networks even though the rules were right
2018-12-11 20:59:52 <danieli> Hamilton: we don't use grsec (linux-hardened) anymore
2018-12-11 21:00:02 <danieli> you should switch to vanilla
2018-12-11 21:00:19 <Hamilton> danieli, Wiki is not updated?
2018-12-11 21:00:38 <Hamilton> And what is vanilla package's name?
2018-12-11 21:01:12 <danieli> looks like that page hasn't been edited for over a year
2018-12-11 21:01:37 <danieli> it's virtualbox-guest-additions-$flavor
2018-12-11 21:01:47 <danieli> https://pkgs.alpinelinux.org/packages?name=*virtualbox*&branch=edge&arch=x86_64
2018-12-11 21:03:04 <Hamilton> danieli, So out of these 4, I should only have the vanilla one?http://termbin.com/6to0
2018-12-11 21:03:12 <danieli> if you run the vanilla kernel, yes
2018-12-11 21:03:16 <Hamilton> meaning I should remove all of them?
2018-12-11 21:03:23 <danieli> ... you have several installed?
2018-12-11 21:03:35 <danieli> you're only supposed to install the one matching your kernel flavor
2018-12-11 21:04:06 <Hamilton> danieli, My kernel is virtual flavoured. (~60 MiB). But I need to have X to have i3
2018-12-11 21:04:17 <Hamilton> So I installed video drivers to be able to start X
2018-12-11 21:05:11 <danieli> virtualbox-guest-modules-virt is probably what you want then
2018-12-11 21:06:21 <Hamilton> danieli, So I delete these 2 video-drivers? xf86-video-modesetting
2018-12-11 21:06:22 <Hamilton> xf86-video-vmware ?
2018-12-11 21:06:32 <Hamilton> I Installed them out of frustration to make X work
2018-12-11 21:06:38 <danieli> I haven't used alpine in vmware so I cannot say
2018-12-11 21:07:17 <Hamilton> danieli, The link you mentioned, what is the difference between them?
2018-12-11 21:07:28 <Hamilton> virt, vanilla and virtualbox-guest-additions
2018-12-11 21:07:31 <danieli> they're built for different kernel flavors
2018-12-11 21:07:47 <Hamilton> It's not obvious that I needed virt  one
2018-12-11 21:14:09 <Hamilton> danieli, I followed this wiki : https://wiki.alpinelinux.org/wiki/XFCE_Setup  but I still can't start X http://termbin.com/nm8rp
2018-12-11 21:40:14 <maldridge> I may be missing something obvious here, but I'm observing openrc tell me that a service was already started, but when I ssh to the machine and run service <service> status it shows clearly it is stopped
2018-12-11 21:40:23 <maldridge> is there some case under which this can happen?
2018-12-11 22:20:43 <maldridge> ok, I've got a system now where I can reliably run a vagrant provisioner which starts a service, the service says "ok" for start status, but then when I ssh in it hasn't been started
2018-12-11 22:21:02 <maldridge> is this some byzantine failure in openrc, or some other misfeature I've run into?
2018-12-11 22:23:33 <brebs> maldridge: Yes, that's possible, it's mostly by pidfile, and pidfiles are not a perfectly reliable method.
2018-12-11 22:24:25 <maldridge> I see
2018-12-11 22:24:36 <maldridge> I'm using the supervise-daemon system if that matters
2018-12-11 22:24:46 <maldridge> I guess I could rm the pidfile before trying to start up again
2018-12-11 22:25:51 <brebs> Or use e.g. pgrep
2018-12-11 22:27:57 <maldridge> I think I may just solve this problem by using a vagrant plugin to bounce the cluster after the initial provisioning step
2018-12-11 23:31:04 <dsetchell> Hello -- just looking for some docs. When installing a fresh alpine from iso via `sys` install how do I add a package to the system?
2018-12-11 23:31:42 <dsetchell> Meaning how do I, after running `setup-alpine`, add say, `bc`, to the set of packages on the system before rebooting into the actual new install?
2018-12-12 01:49:03 <johnnyfive> I just finished creating an application+container that creates and signs repository index/metadata files for RPM, DEB, and APK, all thanks to alpine. W00t! (just had to share my joy!)
2018-12-12 09:51:54 <ulili> Hi, i am trying to use alpine more and more for my docker container, maybe going to use for next homeserver also.
2018-12-12 09:52:40 <ulili> I am having issue to run a perl script which needs libconfig but there is no apk package for it.
2018-12-12 09:53:11 <ulili> What is the way to go for using alpine to install the dependency?
2018-12-12 09:54:38 <kahiru> no package for libconfig?
2018-12-12 09:56:00 <AinNero> libconfig is in main
2018-12-12 09:58:08 <_ikke_> https://pkgs.alpinelinux.org/package/edge/main/x86_64/libconfig
2018-12-12 09:59:34 <mps> OP probably talking about perl Conf::Libconfig
2018-12-12 10:00:32 <ulili> Right mps
2018-12-12 10:01:35 <mps> it is not in alpine but it is easy to build it for yourself
2018-12-12 10:01:35 <ulili> Anyway of cause i installed libconfig package already it is needed for further compile anyhow.
2018-12-12 10:04:45 <mps> ulili: there is apkbuild-cpan in Alpine which create APKBUILD file for perl packages (modules) easily
2018-12-12 10:05:58 <mps> but looks like it doesn't work now because cpan.org doesn't serve queries anymore (all is moved to metacpan.org)
2018-12-12 10:09:02 <ulili> HM OK. Guess i'll take then the "easy" way using a Debian container until there is conf::libconfig Installation possible. Cause at first issue when it comes to perl i would fail to continue AS i do not know that language.
2018-12-12 10:12:07 <mps> ulili: of course, you can do whatever you find to work for your case.
2018-12-12 10:14:29 <mps> but, I was using Debian for over twenty years and have a experience with it but switched to Alpine because of the Debain famous 'dependency of hell'
2018-12-12 10:15:40 <mps> and I don't want to offend Debian, it is quite good OS distribution, to be clear
2018-12-12 10:54:24 <ulili> mps: i am not a fan of Debian but AS it is widely used it gets its Jobs eine quite easy. Most containers use alpine for me AS it is simple small and i like that AS my homrserver has limited RAM also.
2018-12-12 10:55:39 <ulili> So AS soom AS i stumble over solution to install conf::libconfig in arch i willbtry it again that way, but for now it is just easier to use Debian.
2018-12-12 10:56:08 <_ikke_> ulili: AS is written just normally as 'as' :-)
2018-12-12 11:05:08 <Arahael> (It's not an acronym)
2018-12-12 11:10:19 <ulili> _ikke_: it is stupid auto correct on Android.
2018-12-12 11:24:49 <TomJepp> AS stands for Autonomous System though
2018-12-12 11:31:26 <_ikke_> But not in this context
2018-12-12 18:54:29 <jomat> is gajim (1.0.3-r0, edge/testing) working for anyone here?
2018-12-12 23:47:52 <arvidkahl> Hello everyone. If something like "Executing fontconfig-2.13.0-r0.trigger
2018-12-12 23:47:53 <arvidkahl> Error relocating /usr/lib/libexpat.so.1: getrandom: symbol not found" comes my way, what would be a good thing to check? My google research shows nothing about this package in particular, and the build was working as-is 5 days ago.
2018-12-12 23:52:20 <FreeBDSM> how to get `locale` onto alpine?
2018-12-12 23:52:33 <danieli> arvidkahl: have you updated the system recently? are you mixing edge and stable (e.g. 3.8)?
2018-12-12 23:53:03 <arvidkahl> yes, I think i am mixing edge and stable.
2018-12-12 23:53:19 <arvidkahl> i guess "it worked until now" is not a good defence..
2018-12-12 23:53:33 <danieli> it is prone to breaking. is this in a disposable container, on a 'real host'?
2018-12-12 23:53:42 <arvidkahl> container
2018-12-12 23:53:58 <danieli> what's the /etc/apk/repositores you're using?
2018-12-12 23:54:16 <danieli> if you're using edge repos on a 3.8 system, it's unsupported and will most likely break at some point
2018-12-12 23:54:34 <arvidkahl> edge/main
2018-12-12 23:54:41 <danieli> using alpine:edge?
2018-12-12 23:54:55 <arvidkahl> so i update the dockerfile  to FROM alpine:edge  (from alpine:3.8), i guess
2018-12-12 23:55:13 <arvidkahl> oh boy this will be fun :)
2018-12-12 23:55:18 <danieli> you could stick wholly to 3.8 (with 3.8 repos) or use edge
2018-12-12 23:55:32 <danieli> see if the former works, i'd prefer stable if i was you (assuming it's possible)
2018-12-12 23:55:54 <danieli> one sec, i think i'm having a breakthrough with elixir tests breaking on an arch
2018-12-12 23:56:39 <arvidkahl> ah good times, my container is an elixir application as well
2018-12-12 23:56:52 <arvidkahl> the ONE package i really need only lives in edge
2018-12-12 23:57:38 <danieli> it'll be in 3.9 (it'll be out soon™️), i'll see what i can do about getting it into 3.8
2018-12-12 23:58:00 <danieli> also, neat-o! i just mentioned this to someone in #alpine-devel, but i've been an erlanger for some years, just recently falling in love with elixir
2018-12-12 23:58:35 <arvidkahl> it's wonderful as a production system that is actually easy to maintain. just the deployment is a bit quirky, as demonstrated by my cry for help
2018-12-12 23:59:22 <danieli> elixir and erlang have been really nice to package compared with many other languages (cough cough julia, mercury, swi-prolog, java, haskell, ++)
2018-12-13 00:01:01 <arvidkahl> oh yes, I think the BEAM, as hard as it is to get multiple of them together, is very easily containerized. i guess we have erlang releases to thank for that
2018-12-13 00:01:35 <arvidkahl> and the thing is, the erlang/otp and elixir parts never were an issue. the thing i struggle with here is wkhtmltopdf and the font-related libs it depends on
2018-12-13 00:01:58 <danieli> yeah, it's not elixir/erlang itself that's the problem :P just the libraries they're built against
2018-12-13 00:02:01 <danieli> that's often how it is
2018-12-13 00:02:34 <arvidkahl> okay, looks like the FROM alpine:edge  did the trick. i wonder for how long :) thank you so much for being here for me immediately
2018-12-13 00:03:07 <danieli> thank my broken sleep schedule :)
2018-12-13 00:04:58 <arvidkahl> not sure if i should applaud that :)
2018-12-13 00:28:43 <FreeBDSM> how to get `locale` onto alpine?
2018-12-13 00:29:29 <danieli> FreeBDSM: considering it's musl and not glibc, see https://wiki.musl-libc.org/functional-differences-from-glibc.html#The-C/POSIX-locale
2018-12-13 00:40:40 <FreeBDSM> danieli: thanks. Honestly, I hardly understand the locales (all that musl/glibc/iconv). We use alpine as a base for docker-image with php-fpm. We want to use wkhtmltopdf, but for some reason applying a template (I also don't know the details) - we get broken pdfs (looks like all non-latin chars are invisible), I'm trying to find a solution.
2018-12-13 00:40:59 <danieli> tried C.UTF-8?
2018-12-13 00:41:09 <FreeBDSM> no
2018-12-13 00:41:22 <FreeBDSM> how to try it?
2018-12-13 00:41:50 <KH405> https://wiki.alpinelinux.org/wiki/Nextcloud#PostgreSQL I'm trying to do that, but when I install nextcloud-pgsql, this doesen't exist /etc/init.d/postgresql is it normal ?
2018-12-13 00:42:20 <FreeBDSM> danieli: it looks like I need to set up some env-vars, right?
2018-12-13 00:42:26 <danieli> FreeBDSM: correct
2018-12-13 00:43:18 <FreeBDSM> danieli: how to get current values? `locale` is absent and `env` doesn't list LANG/LC_*
2018-12-13 00:43:55 <danieli> KH405: I don't use nextcloud myself, but it looks like the nextcloud-pgsql just provides postgres support for nextcloud, you'll most likely have to install postgres yourself
2018-12-13 00:45:25 <danieli> FreeBDSM: try adding LANG="C.UTF-8" to /etc/profile or whichever profile / shell rc file
2018-12-13 00:46:30 <FreeBDSM> currently, there is `export CHARSET=UTF-8`
2018-12-13 00:46:38 <danieli> did you set it?
2018-12-13 00:47:01 <KH405> danieli: Makes sense!
2018-12-13 00:47:09 <KH405> Will give that a try
2018-12-13 00:47:09 <FreeBDSM> danieli:  where? in the output of `env`? no. in the /etc/profile? yes.
2018-12-13 00:47:30 <danieli> FreeBDSM: I mean, did you add CHARSET=UTF-8 to /etc/profile?
2018-12-13 00:47:39 <FreeBDSM> no
2018-12-13 00:47:52 <danieli> hm, it does look like it's set by default
2018-12-13 00:49:05 <KH405> denieli: so I should only need apk add postgresql ?
2018-12-13 00:49:09 <FreeBDSM> maybe some fonts are missing then?
2018-12-13 00:49:18 <KH405> Will I need the dev, client etc ?
2018-12-13 00:49:44 <danieli> KH405: -dev is development libraries, and -client is fairly self explanatory
2018-12-13 00:50:07 <danieli> FreeBDSM: it has been a pain in the butt in the past, I don't know what you have to do off the top of my head - I haven't touched wkhtmltopdf for quite some time. someone else might know better though
2018-12-13 00:50:16 <KH405> Yeah, I know what they are, just wondering if i'm gonna need them :P, never really played with SQL at all, so i'm jsut starting
2018-12-13 00:50:27 <KH405> wondering *
2018-12-13 00:50:44 <danieli> KH405: I don't use nextcloud so I don't really know, I would think you only need the server, but the client can be handy
2018-12-13 00:51:54 <FreeBDSM> :(
2018-12-13 00:53:22 <FreeBDSM> YAYY
2018-12-13 00:53:25 <FreeBDSM> solved the issue!
2018-12-13 00:53:29 <danieli> what was it?
2018-12-13 00:53:32 <FreeBDSM> it was actually about missing fonts!
2018-12-13 00:53:37 <danieli> bah, typical
2018-12-13 00:53:46 <danieli> always something non-trivial when it comes to musl and locales
2018-12-13 00:53:47 <FreeBDSM> `apk install ttf-opensans` fixed the issue
2018-12-13 00:55:32 <danieli> assuming you meant apk add, but yeah, good to know :)
2018-12-13 00:55:39 <FreeBDSM> yeah, `apk add`
2018-12-13 00:55:45 <FreeBDSM> it's time to start a wiki :)
2018-12-13 00:55:50 <jwh> musl and locales makes me sad
2018-12-13 00:55:56 <danieli> me too jwh, me too
2018-12-13 00:56:09 <jwh> the default php config emits warnings coz iconv may as well not eist
2018-12-13 00:56:10 <danieli> glibc is kind of a massive mess when it comes to that, but musl is a bit of a pain
2018-12-13 00:56:11 <jwh> exist
2018-12-13 00:56:12 <jwh> heh
2018-12-13 00:56:35 <jwh> no TRANSLIT either
2018-12-13 01:15:14 <dalias> is there a recommended way on alpine to configure either touchpad toggle via fn key on laptop, or automatic touchpad disable when external mouse is connected?
2018-12-13 01:15:54 <danieli> you should be able to do that with udev
2018-12-13 01:16:05 <danieli> and synclient
2018-12-13 01:16:30 <dalias> i did it manually with xinput
2018-12-13 01:16:36 <dalias> but you have to know the input device id #
2018-12-13 01:16:43 <dalias> (and i didn't script it)
2018-12-13 01:16:57 <danieli> something along the lines of this? https://wiki.archlinux.org/index.php/Touchpad_Synaptics#Disable_touchpad_on_mouse_detection
2018-12-13 01:17:11 <danieli> mouse[0-9]*
2018-12-13 01:17:35 <dalias> maybe, will try. thanks
2018-12-13 01:17:43 <danieli> looks like it should work
2018-12-13 01:17:56 <danieli> not sure about the function key, probably something along the same lines
2018-12-13 01:18:42 <danieli> you could probably find the keycode for fn+disable touchpad and trap that somewhere
2018-12-13 01:33:14 <dalias> yeah i might add both
2018-12-13 01:33:28 <dalias> default disable touchpad if mouse is plugged in but allow toggle on
2018-12-13 08:47:38 <mps> dalias: xinput allows set paramaters using device name not just id
2018-12-13 09:48:04 <PureTryOut[m]> Why is `rust` only available for x86_64? 🤔
2018-12-13 10:02:05 <AinNero> PureTryOut[m]: issues with bootstrapping
2018-12-13 10:02:33 <AinNero> not specific to alpine afaik, but musl in general
2018-12-13 10:04:18 <PureTryOut[m]> ah damn. are there fixes being made?
2018-12-13 10:08:32 <clandmeter> i think there was some work done on rust in latest release.
2018-12-13 10:08:44 <clandmeter> https://github.com/rust-lang/rust/blob/master/RELEASES.md
2018-12-13 16:44:12 <arvidkahl> my upgrade from 3.8 to edge snowballed into an explosion of dependency upgrades which left no survivors. i figured out how to get it all reigned in using 3.8, with just one package using edge/community. I learned quite a bit, and have made base image for my container to be rid of the risk of alpine packages being moved around. thanks again @danieli for your assistance last night
2018-12-13 16:55:15 <danieli> arvidkahl: glad you got it sorted, and yeah, here be dragons
2018-12-13 16:56:06 <arvidkahl> you truly ever appreciate the stability and ease of 3.8 if you have ventured into edge :)
2018-12-13 16:56:27 <arvidkahl> makes me appreciate the care that is being put into making 3.9 just as stable
2018-12-13 16:57:03 <arvidkahl> i was biting my nails because i wanted to get the latest erlang/OTP going, but not at the price of loss of sleep
2018-12-13 17:05:44 <wej> is edge really that problematic? i guess i was pretty lucky then. i switched at some point from 3.6 to edge and never had any issues
2018-12-13 17:06:07 <_ikke_> wej: depends on when you update
2018-12-13 17:06:39 <wej> i see. is there any way to tell when i should avoid updating?
2018-12-13 17:11:21 <danieli> if you need stability, edge *can* be a bad choice
2018-12-13 17:11:26 <danieli> it's the development branch
2018-12-13 17:11:57 <wej> i know :) i wouldn't complain if anything breaks
2018-12-13 17:12:35 <danieli> a poke about broken stuff is nice
2018-12-13 17:13:33 <wej> sure, what i meant was more like, i wouldn't want you guys to fix my mess in case anything breaks because i use edge ;)
2018-12-13 17:14:10 <danieli> if you use edge normally and without any funny business (using edge repos on non-edge), a poke about broken stuff is welcome
2018-12-13 17:14:26 <wej> ok, i'll keep that in mind :)
2018-12-13 17:14:46 <jn__> it's especially useful because current edge will end up in a stable release in the future
2018-12-13 17:14:51 <jn__> (more or less)
2018-12-13 17:16:04 <wej> another thing, is there a package for stubby (the dns-over-tls resolver) in alpine? i know on some distributions it is called differently or part of another package, so maybe i was looking for the wrong thing
2018-12-13 17:16:09 <danieli> releases are essentially freeze releases with some cherry picked stuff
2018-12-13 19:42:20 <danieli> any clue what font the alpine linux logo is?
2018-12-13 19:42:24 <danieli> cc clandmeter _ikke_
2018-12-13 19:42:36 <clandmeter> nope
2018-12-13 19:42:39 <danieli> darn it
2018-12-13 19:42:48 <clandmeter> if anyone knows, its ncopa.
2018-12-13 19:42:53 <clandmeter> he knows the author
2018-12-13 19:44:52 <danieli> was thinking one of you might have looked into it at some point
2018-12-13 19:45:04 <danieli> i tried asking a long time ago but never got a response about the font
2018-12-13 19:45:12 <danieli> or rather, i never got an answer
2018-12-13 20:34:22 <jedahan> @danieli maybe `toilet -f wideterm alpine` :)
2018-12-13 20:34:59 <jedahan> or with circle: ⓐ ⓛ ⓟ ⓘ ⓝ ⓔ
2018-12-13 20:35:42 <danieli> haha
2018-12-13 20:37:24 <caryatid> Hello! I'm looking for the best way to have some stuff run just after first boot. similar to rc.local ( runs all boot ) is ok, but seems the `local` service isn't enabled by default. Basically this is for vm config. Imagine I have a disk image just after alpine install. I want to mount that image and place a shell file in it that will be executed on first boot to do some config ( install packages, set default services, etc ). 
2018-12-13 20:39:42 <mps> caryatid: rc-update add local
2018-12-13 20:39:58 <jedahan> I'm getting `missing libcrypto`, which is used by cargo
2018-12-13 20:40:03 <jedahan> on an apk update
2018-12-13 20:40:07 <danieli> you could probably add a service that disables itself after running (after first boot), not sure if that's super clean though
2018-12-13 20:40:07 <jedahan> which I havent done in a few months
2018-12-13 20:40:10 <mps> local is not enabled by default
2018-12-13 20:40:10 <caryatid> mps: right .. but how do I run that from a mounted image -
2018-12-13 20:40:24 <jedahan> any information on how to resolve that missing dependency?
2018-12-13 20:40:32 <danieli> jedahan: try `apk fix`
2018-12-13 20:40:32 <mps> mounted? care to explain
2018-12-13 20:40:39 <danieli> if you've upgraded to edge, that'll happen
2018-12-13 20:40:46 <caryatid> i.e. can I edit a file to effecively set local to default?
2018-12-13 20:40:53 <jedahan> @danieli I'm not on edge, and apk fix doesn't fix it
2018-12-13 20:41:03 <_ikke_> caryatid: should be a symlink in the apropriate directory
2018-12-13 20:41:08 <danieli> jedahan: do you have any edge repositories added?
2018-12-13 20:41:11 <jedahan> ERROR: unsatisfiable constraints:  so:libcrypto.so.1.1 (missing):, and so:libssl.so.1.1 (missing):
2018-12-13 20:41:13 <jedahan> checking now
2018-12-13 20:41:19 <danieli> we moved from libressl to openssl
2018-12-13 20:41:40 <_ikke_> caryatid: ln -s /etc/init.d/local /etc/runlevels/default
2018-12-13 20:41:46 <caryatid> mps: situation: alpine installed to a qemu image to point that setup-alpine says "reboot". qemu gets halted. I have that image to mount and modify before next boot
2018-12-13 20:41:58 <caryatid> mps: _ikke_: that's the way. Thank you!
2018-12-13 20:42:00 <mps> /etc/runlevels/default/local -> /etc/init.d/local
2018-12-13 20:42:13 <mps> 'ln -s' of course
2018-12-13 20:42:22 <jedahan> oh weird, i did have an edge entry for community
2018-12-13 20:42:23 <jedahan> strange
2018-12-13 20:42:48 <jedahan> @danieli interesting, is there a good read on the decision to do that?
2018-12-13 20:42:55 <_ikke_> jedahan: mailing list
2018-12-13 20:43:23 <danieli> you should either use a 3.8-only install, or upgrade to edge entirely
2018-12-13 20:43:48 <_ikke_> jedahan: https://lists.alpinelinux.org/alpine-devel/6308.html
2018-12-13 20:43:58 <jedahan> thanks!
2018-12-13 20:44:01 <jedahan> brb
2018-12-13 20:44:06 <jedahan> yeah i think it was another user who added that :)
2018-12-13 20:44:10 <jedahan> everyone on my box has sudo permissions
2018-12-13 23:18:35 <cljm> I'm experiencing 100% cpu usage of docker-containerd just after installation & not running any containers - anyone else know why this might happen?
2018-12-13 23:19:06 <cljm> https://pastebin.com/raw/NSHzJXF4 <- is some of my system info
2018-12-13 23:19:41 <cljm> I basically just do "apk add docker" and "service docker start" on a fresh install to make this happen
2018-12-13 23:37:14 <danieli> hm, that is interesting
2018-12-13 23:37:21 <danieli> is it running on bare metal? if not, which hypervisor?
2018-12-13 23:37:34 <danieli> you know what, would you mind creating an issue at bugs.alpinelinux.org ?
2018-12-13 23:44:15 <cljm> danieli: sure thing - I believe this to be more of a docker issue, but it's also my first experience with alpine.
2018-12-13 23:44:25 <danieli> it's possible, but i haven't heard of that before
2018-12-13 23:44:37 <danieli> it sounds like an edge case *somewhere*, in some magical land of code
2018-12-13 23:44:39 <cljm> The docker guys aren't really responsive, so I thought it couldn't hurt to ask here. The system works great otherwise... it's running under vmd on OpenBSD
2018-12-13 23:44:44 <danieli> ncopa works at docker
2018-12-13 23:44:53 <danieli> he also happens to be the project BDFL
2018-12-13 23:45:01 <cljm> neato :)
2018-12-13 23:45:09 <cljm> I'm a huge fan of alpine at work
2018-12-13 23:45:13 <danieli> he's gone most of this week though, at a conference in paris
2018-12-13 23:45:19 <danieli> hah, i hear that every once in a while
2018-12-13 23:45:26 <danieli> it has its warts, but it's slowly getting better
2018-12-13 23:45:40 <cljm> eh - at least it doesn't appear to have systemd ;)
2018-12-13 23:45:54 <danieli> hah, or policykit
2018-12-13 23:46:00 <danieli> nah, I'm gonna refrain from the jabs
2018-12-13 23:46:10 <danieli> there are pros and cons :)
2018-12-13 23:46:49 <cljm> I'll get a proper bug report together and post it with reproducable steps sometime tonight or tomorrow. No worries if you guys end up closing it... I know I'm in the minority with my system
2018-12-13 23:47:29 <danieli> no worries, i'll have a look for sure, it sounds interesting
2018-12-13 23:47:40 <danieli> i'll have to reproduce it first though
2018-12-13 23:52:43 <cljm> danieli: I may actually try to reproduce this on another machine before filing a bug, so give me some extra time :)
2018-12-13 23:52:53 <danieli> no worries buddy
2018-12-13 23:52:54 <cljm> if I can, I'll post the OpenBSD steps as well
2018-12-13 23:53:06 <danieli> using vmd/vmm?
2018-12-13 23:53:12 <cljm> yup
2018-12-13 23:53:22 <danieli> then i could repro it, i have some openbsd boxen sitting around
2018-12-13 23:53:32 <cljm> \o/
2018-12-13 23:53:53 <danieli> s/could/could try to/
2018-12-14 14:16:30 <beli3ver> hi
2018-12-14 14:16:45 <beli3ver> I am new at alpine and I hope somebody could help me with a little problem
2018-12-14 14:18:36 <jn__> beli3ver: go ahead and find out
2018-12-14 14:19:11 <beli3ver> ok, thanks. When I tried to start I get always errors about sandbox and firefox is not starting how could I fix this
2018-12-14 14:20:54 <beli3ver> tried to start firefox
2018-12-14 15:04:38 <cljm> danieli: so I wasn't able to reproduce the issue on a t420, so I've opened a ticket with OpenBSD. The newer intel CPUs probably have a magic bit somewhere that vmm/vmd isn't using... thanks for your help yesterday
2018-12-14 15:05:55 <weiti> hi, just a quick (newbie) question, for the diskless install mode, is it possible to use another filesystem than vfat? CF card with ext2/3/4?
2018-12-14 15:44:58 <clandmeter> weiti, yes you can
2018-12-14 15:48:15 <weiti> clandmeter: ok, thanks. will test it ...
2018-12-14 16:33:03 <danieli> cljm: it wouldn't surprise me if it's a generic linux on vmm/vmd issue, it's pretty fresh tech
2018-12-14 16:34:12 <cljm> definitely - I was just thrown for a loop due to everything else working so seamlessly
2018-12-14 16:34:21 <cljm> docker is a crazy piece of tech, though
2018-12-14 16:36:16 <danieli> while you're struggling with that, i finally wiped my laptop and got alpine booted on it
2018-12-14 16:36:29 <danieli> lvm2 on luks with grub isn't easy
2018-12-14 16:36:41 <danieli> i got it set up and the system bootstrapped and all, but my system refuses to boot grub
2018-12-14 16:37:09 <cljm> -_-
2018-12-14 16:37:23 <danieli> oh well, by doing this, i can see plenty of "easy" things to fix to improve quality of life
2018-12-14 16:37:48 <cljm> what are you planning to use this box for? I'm new to alpine, and it's very intriguing
2018-12-14 16:38:31 <danieli> this one, my laptop
2018-12-14 16:38:36 <danieli> er, my primary laptop
2018-12-14 16:39:05 <cljm> nice. I'll admit that it seems customizable as hell
2018-12-14 16:39:20 <cljm> I absolutely hate all the crap that comes on ubuntu, etc these days
2018-12-14 16:39:37 <danieli> yeah, but it's 100x harder to install than arch because it has only a fraction of the convenience script and documentation
2018-12-14 16:39:56 <cljm> I'll agree with that :)
2018-12-14 16:40:20 <cljm> I can't speak for everyone, but my experience comes from using it almost exclusively within docker (absolutely perfect for that)
2018-12-14 19:08:26 <danieli> and when trying LVM without LUKS, it complaines about multiple memory leaks
2018-12-14 19:08:27 <danieli> god damn it
2018-12-14 19:10:41 <danieli> holy hell, it actually worked
2018-12-14 19:12:23 <danieli> doesn't boot though, darn it
2018-12-14 19:15:49 <danieli> ... openrc just froze for some reason
2018-12-14 19:23:23 <Processus42> Hello there. What is the CLIBC environement variable I can find in some APKBUILD files, and how is it set ?
2018-12-14 19:24:24 <danieli> we only use musl now (and not any other libcs like uclibc), so it shouldn't really matter
2018-12-14 19:24:30 <danieli> i can look around real quick though
2018-12-14 19:25:25 <Processus42> Don't bother yourself too much :) , I could look as well, but I hoped someone had the answer already
2018-12-14 19:25:32 <danieli> try #alpine-devel
2018-12-14 19:25:36 <danieli> i don't know off the top of my head to be honest
2018-12-14 19:25:55 <Processus42> Thanks :)
2018-12-14 20:51:40 <muhaha> Any idea why in alpine:3.8 does not work dig ANY google.com?    (-t ANY)
2018-12-14 20:58:26 <_ikke_> muhaha: works for me
2018-12-14 21:03:59 <muhaha> docker run --rm -it alpine:3.8 sh -c "apk update > /dev/null && apk add --no-cache bind-tools > /dev/null && dig any +short google.com"
2018-12-14 21:09:18 <programmerq> muhaha▸ as an aside, with that command you can omit the 'apk update' command entirely. it's a bit strange to see 'apk update && apk add --no-cache' together. the apk update builds the cache. --no-cache tells 'apk add' to query the repository and then not cache it locally, but you've just finished a command that caches it locally.
2018-12-14 21:13:54 <programmerq> muhaha▸ in my own testing, it looks like that behavior happens inside a docker container that isn't an alpine container. "dig -t any" seems to not be happy with docker's built-in DNS server, so this isn't an alpine question. Come on over to #docker and I'll take a further look at it.
2018-12-14 21:14:52 <muhaha> programmerq: seems that its working only with local cache.
2018-12-14 21:15:02 <muhaha> dig ANY +short google.com @1.1.1.1 does not work
2018-12-14 21:23:03 <programmerq> muhaha▸ I tried it in my docker-for-mac, and it just hangs. if I specify my home network's router, it works. If I specify 1.1.1.1 or 8.8.8.8 it doesn't. If I specify 129.168.65.1 (which is the DNS server as seen inside the docker-for-mac VM) then it hangs. If I try it on a linux host directly on my network, it works from inside the container through Docker's 127.0.0.11 internal DNS server.
2018-12-14 21:23:13 <programmerq> where are you running it? insido a docker-for-mac VM?
2018-12-14 21:23:18 <programmerq> inside*
2018-12-14 21:23:56 <programmerq> (btw, I see the same behavior whether I use the alpine dig binary or a centos dig binary)
2018-12-14 21:25:32 <programmerq> my home router runs a pretty vanilla dnsmasq with 8.8.8.8/8.8.4.4 set as upstreams
2018-12-14 22:02:33 <danieli> muhaha: it works for me, but (understandably) doesn't return any data
2018-12-14 22:02:38 <danieli> and that's on arch
2018-12-14 22:18:26 <abond_> Hello.. Could someone provide details on why my apk add is failing with "unsatisfiable constraints"?
2018-12-14 22:18:53 <abond_> I'm building on 3.8.1 and I'm trying to run "RUN apk add --update ca-certificates && \     rm /var/cache/apk/*"
2018-12-14 22:40:37 <danieli> abond_: which repos are you using? why not `apk add --no-cache ca-certificates`?
2018-12-14 22:41:53 <abond_> danieli: this would be on the gliderlabs/alpine
2018-12-14 22:42:07 <danieli> uh?
2018-12-14 22:42:20 <danieli> oh right, that one
2018-12-14 22:42:33 <abond_> docker run -it gliderlabs/alpine apk add --no-cache ca-certificates
2018-12-14 22:43:04 <abond_> Fails also. I actually found this issue: https://github.com/gliderlabs/docker-alpine/issues/464
2018-12-15 06:01:08 <npx> so... how can I resolve " so:libavcodec.so.57 (missing):"
2018-12-15 06:02:48 <npx> sorry, was tired ;)
2018-12-15 06:50:20 <npx> hahah wow so community/kodi requires essentially every piece of software in the history of software as a dependency
2018-12-15 06:50:44 <npx> (build time at least)
2018-12-15 07:12:47 <clandmeter> npx, yes and then they include a bunch of other libs with the tarball to make it even nicer.
2018-12-15 07:18:55 <jwh> use non-bundled libraries at your peril
2018-12-15 07:18:56 <jwh> etc
2018-12-15 07:19:03 <jwh> theres a reason they're bunded
2018-12-15 07:19:05 <jwh> bundled*
2018-12-15 09:21:38 <Processus42> danieli: CLIBC is given a default value by abuild, which determines it by looking at gcc -dumpmachine output
2018-12-15 14:53:38 <palate> hello :)
2018-12-15 14:54:23 <palate> I'm trying to build a project on alpine linux, with cmake. The project requires eigen3, which usually comes with `Eigen3ConfigVersion.cmake`. On Alpine, though, it doesn't
2018-12-15 14:54:55 <palate> So cmake finds and Eigen3Config.cmake file, but ignores it because it doesn't contain the right version
2018-12-15 14:55:07 <palate> Do you guys know about that?
2018-12-15 22:13:01 <jaroslaw> hey
2018-12-15 22:13:15 <jaroslaw> firefox 53 ESR is the only one in 3.8.1 ?
2018-12-15 22:13:31 <jaroslaw> or even 52 ESR
2018-12-15 22:30:57 <danieli> looks like firefox-esr is 52.9 in both edge and 3.8
2018-12-16 00:24:38 <kupi> hi
2018-12-16 00:36:53 <danieli> hi
2018-12-16 02:27:37 <jaroslaw> danieli: but 3.8.1 has only 52 ESR
2018-12-16 02:32:18 <danieli> what exactly is the problem? I'm not sure I understand
2018-12-16 09:40:39 <KH405_TV> Hey, could someone help me make a RC-Script?
2018-12-16 09:40:58 <KH405_TV> Is this https://dev.deluge-torrent.org/wiki/UserGuide/Service/FreeBSD Suppose to work ?
2018-12-16 09:41:00 <KH405_TV> The first one
2018-12-16 09:41:27 <KH405_TV> I'm getting /etc/init.d/deluged: .: line 19: can't open '/etc/rc.subr': No such file or directory
2018-12-16 09:41:41 <_ikke_> Doesn't look like an openrc init script
2018-12-16 09:41:44 <_ikke_> no
2018-12-16 09:41:56 <KH405_TV> I get that /etc/rc.subr dosen't exist, but what should it point too?
2018-12-16 09:42:45 <_ikke_> openrc init scripts look a bit different
2018-12-16 09:43:10 <KH405_TV> _ikke_: You've got any documentation for a noob for me on how to write scripts that only start daemon on startup?
2018-12-16 09:43:20 <_ikke_> KH405_TV: https://github.com/OpenRC/openrc/blob/master/service-script-guide.md
2018-12-16 09:43:21 <KH405_TV> Weird, cause it says it's an Rc script :s
2018-12-16 09:43:24 <_ikke_> Was looking for it
2018-12-16 09:43:35 <_ikke_> rc is common term
2018-12-16 09:45:06 <KH405_TV> Thansk, gonna read that before going to bed!
2018-12-16 09:46:18 <_ikke_> KH405_TV: If the deluged is a sane, you probably only have define a few options
2018-12-16 09:46:35 <KH405_TV> ''sane'' ?
2018-12-16 09:47:17 <_ikke_> forking / pid-file behaviour
2018-12-16 09:47:41 <KH405_TV> Ok
2018-12-16 09:52:22 <KH405_TV> Anything I could base my script on?
2018-12-16 10:22:44 <_ikke_> KH405_TV: look at existing scripts in /etc/init.d
2018-12-16 20:01:58 <feuerteufel> Hi
2018-12-16 20:03:24 <feuerteufel> Is someone here who is using "fcron"
2018-12-16 20:04:11 <feuerteufel> On my site it is crashing when I start the service
2018-12-16 20:05:28 <danieli> OS version? can you run it manually without crashes?
2018-12-16 20:05:38 <feuerteufel> After that, there is no way (I know) to stop it (only with reboot)
2018-12-16 20:05:54 <feuerteufel> Alpine 3.8.1
2018-12-16 20:06:39 <feuerteufel> It worked before I upgraded to 3.8.1
2018-12-16 20:07:31 <danieli> but can you run it manually (outside openrc)?
2018-12-16 20:08:47 <feuerteufel> Thats what I tryed and found aut that there ist the same issue when it is started I can't stop it
2018-12-16 20:10:11 <danieli> no output whatsoever?
2018-12-16 20:10:16 <feuerteufel> I need the "@reboot" option, that's why i'm using fcron
2018-12-16 20:10:58 <feuerteufel> On start he say's is started, on  Stop I get this:
2018-12-16 20:11:15 <feuerteufel> service fcron stop
2018-12-16 20:11:15 <feuerteufel>  * Stopping fcron ...                                                                         [ !! ]
2018-12-16 20:11:15 <feuerteufel>  * ERROR: fcron failed to stop
2018-12-16 20:11:40 <danieli> what's the output when you run it manually (in the shell) though?
2018-12-16 20:13:36 <feuerteufel> Oh, you mean "/usr/sbin/fcron" like this?
2018-12-16 20:14:56 <feuerteufel> What I also found out, is that zhere ist no pid to find
2018-12-16 20:19:12 <feuerteufel> In /var/log/messages I only find something like that:
2018-12-16 20:19:16 <feuerteufel> xen-cel /etc/init.d/fcron[8891]: WARNING: fcron has already been started
2018-12-16 20:19:34 <feuerteufel> xen-cel /etc/init.d/fcron[8671]: /etc/init.d/fcron uses runscript, please convert to openrc-run.
2018-12-16 20:19:54 <danieli> interesting, looks like it has a bad init script in that case
2018-12-16 20:20:20 <feuerteufel> I know, I have to change the init script which I allready did
2018-12-16 20:21:28 <feuerteufel> Yes, the init script ist not up to date in the package
2018-12-16 20:21:59 <danieli> it hasn't been touched in almost 5 years
2018-12-16 20:22:11 <feuerteufel> But the prob is, I can't stop the fcron
2018-12-16 20:22:23 <danieli> create an issue at bugs.alpinelinux.org
2018-12-16 20:23:06 <feuerteufel> I always have to reboot
2018-12-16 20:44:07 <feuerteufel> danieli: I think, I find, at least one, problem ...
2018-12-16 20:44:45 <feuerteufel> Here I found:
2018-12-16 20:44:51 <feuerteufel> http://fcron.free.fr/doc/en/install.html
2018-12-16 20:45:02 <feuerteufel>  fcron requires a fcron user and a fcron group. If they don't exist already, make install will propose to create them for you.
2018-12-16 20:45:37 <feuerteufel> On my sys is none of them
2018-12-16 20:46:01 <danieli> did you create an issue about fcron at bugs.alpinelinux.org ?
2018-12-16 20:46:44 <feuerteufel> I was searching to put more info in there ...
2018-12-16 20:46:52 <danieli> all right, fair enough
2018-12-16 20:47:14 <danieli> just checking, since there isn't much I can do about it right now
2018-12-16 20:47:39 <feuerteufel> So I'll try to get it back to run
2018-12-16 20:48:04 <feuerteufel> May be I'll find the problem
2018-12-16 20:48:23 <danieli> it sounds to me like it's fairly broken right now
2018-12-16 20:48:49 <feuerteufel> I also found some stuff like that:
2018-12-16 20:49:00 <feuerteufel> -rw-r-----    1 root     226            653 May 19  2018 fcron.conf
2018-12-16 20:49:16 <danieli> put it in the bug report :)
2018-12-16 20:50:04 <feuerteufel> I think "226" was the group "fcron" before it was killed ;)
2018-12-16 21:38:00 <Greyfn> Good morning/afternoon/evening all - I have a question about awall if anybody has any expertise?
2018-12-16 21:40:12 <danieli> some people have extensive knowledge with it - not sure if they're present on a sunday, but shoot
2018-12-16 21:41:34 <Greyfn> I'm trying to set up a bridged firewall - so no NAT involved.
2018-12-16 21:41:43 <Greyfn> I have eth0 and eth1 as part of bridge br0
2018-12-16 21:42:18 <Greyfn> on shorewall I can denote a zone as part of a bridge
2018-12-16 21:42:29 <Greyfn> so that I can have different rules in each direction
2018-12-16 21:42:57 <Greyfn> I can't figure out how to do this on awall - I can't apply policies to eth0 or eth1 as they are part of the bridge
2018-12-16 21:44:23 <Greyfn> I'm fairly convinced it's a limitation with awall at present, but I'd like somebody with some experience to confirm/deny
2018-12-16 21:44:56 <Greyfn> (specifically I'm trying to do the equivalent in awall as shorewall: http://shorewall.net/bridge-Shorewall-perl.html)
2018-12-16 21:45:43 <Greyfn> I can't find much mention of bridges in the awall documentation at all really
2018-12-16 22:05:29 <feuerteufel> danieli: Can't get it back to run ... :-((
2018-12-16 22:05:45 <feuerteufel> here is the bug:
2018-12-16 22:05:53 <feuerteufel> https://bugs.alpinelinux.org/issues/9772
2018-12-16 22:06:06 <danieli> I saw, new issues get posted in #alpine-devel
2018-12-16 22:14:20 <clandmeter> feuerteufel, you are saying that fcron always worked before?
2018-12-16 22:15:10 <feuerteufel> Yes till I upgraded to 3.8
2018-12-16 22:15:44 <feuerteufel> I didn't found that out till now
2018-12-16 22:16:37 <clandmeter> that initd looks completely broken.
2018-12-16 22:17:09 <feuerteufel> That one I made?
2018-12-16 22:17:21 <clandmeter> the one shipped
2018-12-16 22:17:28 <clandmeter> also the pkg is broken
2018-12-16 22:17:42 <clandmeter> so i cant imagine it ever worked.
2018-12-16 22:17:57 <feuerteufel> Oh ...
2018-12-16 22:18:31 <danieli> that specific initd, agreed
2018-12-16 22:19:46 <feuerteufel> I'm sorry, I don't have older installations any more
2018-12-16 22:20:01 <clandmeter> i got to get some sleep.
2018-12-16 22:20:03 <clandmeter> gnite
2018-12-16 22:20:08 <danieli> night
2018-12-16 22:20:11 <_ikke_> nite
2018-12-16 22:20:18 <_ikke_> coming week you are gone right?
2018-12-16 22:20:25 <danieli> iirc yes
2018-12-16 22:20:25 <clandmeter> wednesday
2018-12-16 22:20:28 <danieli> aha
2018-12-16 22:20:43 <_ikke_> nod
2018-12-16 22:20:48 <feuerteufel> That's what I think too
2018-12-16 22:20:48 <clandmeter> going to prague
2018-12-16 22:21:28 <danieli> looks like fcron never has worked, that initd hasn't been changed much since 2014
2018-12-16 22:21:54 <clandmeter> the permissions of fcron is also incorrect
2018-12-16 22:22:11 <danieli> the only change that has been made to the initd is /sbin/runscript -> /sbin/openrc-run in the shebang
2018-12-16 22:27:34 <clandmeter> the build scripts are a mess.
2018-12-16 22:27:39 <clandmeter> fedora also removed it
2018-12-16 22:28:03 <danieli> feuerteufel: just curious, why are you using fcron in the first place? why not something sane?
2018-12-16 22:29:59 <feuerteufel> Because I whant to use "@reboot" for example
2018-12-16 22:33:59 <clandmeter> feuerteufel, why dont you use something like local rc?
2018-12-16 22:35:38 <feuerteufel> That was yust an example ...
2018-12-16 22:37:01 <feuerteufel> But if I can't get it to run, I have to think about other way's
2018-12-16 22:37:19 <clandmeter> Add an example to the issue that makes sense for us to spend time on it.
2018-12-16 22:37:41 <danieli> if the build scripts are terrible upstream, the initd is broken, and other projects are removing it - i don't see why we should keep it around at all
2018-12-16 22:38:34 <clandmeter> Cron is not made to start thing at boot
2018-12-16 22:39:17 <clandmeter> There is an pr on GitHub about it
2018-12-16 22:41:07 <feuerteufel> But it look's like it was Changed on 2018-10-05 ...
2018-12-16 22:41:16 <feuerteufel> https://pkgs.alpinelinux.org/package/edge/community/x86/fcron
2018-12-16 22:42:55 <danieli> doesn't really matter, it's still broken - the only things that have been changed is being upgraded and cleaning the APKBUILD a little
2018-12-16 22:44:09 <feuerteufel> So if I want to use it, I have to compile it by my self?
2018-12-16 22:44:21 <danieli> it's broken in alpine right now, so yes
2018-12-16 22:50:37 <feuerteufel> clandmeter: That will by a never ending dicussion, what is made to do whatevver ...
2018-12-16 22:51:11 <danieli> cron is literally just a time-based job scheduler
2018-12-16 22:51:29 <danieli> if some obscure implementation exceeds that, it doesn't change what cron is meant to do
2018-12-16 22:51:44 <danieli> s/meant/made/
2018-12-16 22:51:46 <feuerteufel> I don't think somebody need that
2018-12-16 22:51:58 <danieli> needs what, cron?
2018-12-16 22:55:49 <feuerteufel> A discussion what will be used for what
2018-12-16 22:56:49 <danieli> I'm confused?
2018-12-16 22:57:29 <feuerteufel> <feuerteufel> clandmeter: That will by a never ending dicussion, what is made to do whatevver ...
2018-12-16 22:57:46 <danieli> cron is a time-based job scheduler
2018-12-16 22:58:03 <feuerteufel> That's what I'm talking about ...
2018-12-16 22:58:05 <danieli> there isn't really more to it
2018-12-16 23:00:17 <danieli> except this https://en.wikipedia.org/wiki/Cron#Nonstandard_predefined_scheduling_definitions
2018-12-16 23:03:50 <danieli> debian's even seems to have a fix to only run @reboot jobs on reboot and not just when cron is restarted
2018-12-16 23:04:10 <feuerteufel> But fcron is able to do things like "@yearly @ nightly" und so on
2018-12-16 23:04:39 <danieli> yeah, non-standard ones
2018-12-16 23:06:50 <feuerteufel> What ever, I'm used to use fcron, it makes many things easyer
2018-12-16 23:07:51 <danieli> fair enough, it'll probably get fixed eventually
2018-12-16 23:07:57 <feuerteufel> But it dosn't work on Alpine now so I hve to find a solution
2018-12-16 23:09:17 <feuerteufel> If that is possible, it would be great!
2018-12-16 23:09:19 <danieli> it's generally not a good idea to rely on implementation-specific (non-standard) details
2018-12-16 23:11:19 <feuerteufel> Gentoo has bcron, as fare as I know it will do the same ...
2018-12-16 23:12:48 <danieli> i don't believe bcron is packaged for alpine
2018-12-16 23:13:30 <feuerteufel> No, i don't think also
2018-12-16 23:15:32 <feuerteufel> If you able to get it fixed, will be great!
2018-12-16 23:15:47 <feuerteufel> And yust take your time!
2018-12-16 23:15:50 <danieli> I'm not looking at it, I'm busy with a bunch of other stuff, but someone else might
2018-12-16 23:16:55 <feuerteufel> it dosn't work here for almost 2 month
2018-12-16 23:17:59 <feuerteufel> What I yust found out, ist it looks like there is a newer version on edge
2018-12-16 23:18:28 <danieli> the only thing that was changed was the package version, if you're aiming at fcron
2018-12-16 23:19:56 <feuerteufel> So you think that one is also not working?
2018-12-16 23:20:19 <danieli> the initd hasn't been touched so at least one thing will still be broken
2018-12-16 23:23:58 <feuerteufel> there is also:
2018-12-16 23:24:01 <feuerteufel> fcron-openrc-3.2.1-r0 - periodical command scheduler for systems not always up (OpenRC init scripts)
2018-12-16 23:24:21 <danieli> abuild splits out the initscripts from a package into a separate subpackage
2018-12-16 23:24:47 <danieli> it isn't a different initd
2018-12-16 23:24:59 <feuerteufel> Ah, ok
2018-12-16 23:28:02 <feuerteufel> I'll try that tomorrow, I have a testsystem on edge
2018-12-16 23:28:10 <danieli> try what, exactly?
2018-12-16 23:28:49 <feuerteufel> But now I'm to tired
2018-12-16 23:28:56 <danieli> fcron isn't fixed in edge either, fwiw
2018-12-16 23:29:35 <feuerteufel> Install and run fcron
2018-12-16 23:30:01 <feuerteufel> Ah, you think it's broken there also?
2018-12-16 23:30:08 <danieli> i know it's broken there too, it hasn't been fixed anywhere in alpine
2018-12-16 23:30:13 <danieli> i looked at the repository for it earlier
2018-12-16 23:35:59 <hatramatra> Greetings everyone. I'm having troubles to mount btrfs filesystem on multiple devices.
2018-12-16 23:36:31 <hatramatra> It is mandatory to run 'btrfs device scan', but I don't think it happens when booting from iso image.
2018-12-16 23:36:45 <hatramatra> If I run it manually, then I can mount.
2018-12-16 23:37:27 <hatramatra> Bug #6903 might be fixing this for sys installations.
2018-12-16 23:42:46 <feuerteufel> N8 @ all
2018-12-16 23:43:18 <danieli> nate?
2018-12-16 23:43:21 <danieli> oh, night
2018-12-17 00:02:30 <hatramatra> Right, I just tried on the clean installation with VARFS=btrfs set. Booting from ISO and lbu saving to the usb disk.
2018-12-17 00:02:59 <hatramatra> Everything works, up until I add another device to the btrfs filesystem and run btrfs balance ...
2018-12-17 00:03:48 <hatramatra> Filesystem becomes raid1 and then after reboot /var won't get mounted. I need to run manually 'btrfs device scan'.
2018-12-17 00:04:20 <hatramatra> I wonder what is the best place to run that. Somewhere in the boot runlevel before localmount?
2018-12-17 16:58:34 <jokke> hey there o/
2018-12-17 16:58:53 <jokke> i'm having trouble linking a program in alpine
2018-12-17 16:59:02 <jokke> it uses libmagic
2018-12-17 16:59:16 <jokke> but the linker fails with /usr/lib/gcc/x86_64-alpine-linux-musl/8.2.0/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -lmagic (this usually means you need to install the development package for libmagic)
2018-12-17 16:59:37 <jokke> i have installed file-dev
2018-12-17 17:06:52 <jokke> these files exist in /usr/lib: https://p.jokke.space/BEV/
2018-12-17 18:05:53 <Shiz> hmm
2018-12-17 18:06:00 <Shiz> jokke: are you somehow trying to link statically?
2018-12-17 18:07:05 <dalias> shiz, sounds possible. lots of ppl have been told to use an alpine chroot or container to build static musl binaries easily :)
2018-12-17 18:08:31 <Shiz> looks like there's no libmagic.a (and it would require the presence of a /usr/share/misc/magic.mgc on the host system anyway)
2018-12-17 18:09:04 <Shiz> checking whether to build static libraries... no
2018-12-17 18:09:08 <Shiz> may need a fix in the file APKBUILD
2018-12-17 18:20:58 <jokke> yeah i was
2018-12-17 18:21:12 <jokke> there indeed is no libmagic.a
2018-12-17 18:30:23 <jwh> 
ACTION cries

2018-12-17 18:30:31 <jwh> got no TX on either built in or usb serial
2018-12-17 18:30:31 <dalias> ?
2018-12-17 18:32:48 <AinNero> jwh: have you tried enabling it from the software side?
2018-12-17 18:33:08 <jwh> how do you mean?
2018-12-17 18:33:38 <AinNero> i dont know.
2018-12-17 18:33:57 <AinNero> you wrote something really generic, with no context or even mentioning the device you are on
2018-12-17 18:34:15 <jwh> pretty clear tbf
2018-12-17 18:34:25 <jwh> TX is broken on mutiple serial controllers
2018-12-17 18:34:25 <jwh> :D
2018-12-17 18:34:55 <jwh> as in, it never makes it to the device
2018-12-17 18:35:39 <AinNero> ah ha, so its TX from the dongle side
2018-12-17 18:35:51 <AinNero> i thought you mean TX from the device
2018-12-17 18:35:56 <jwh> or onboard port
2018-12-17 18:35:56 <jwh> yeah
2018-12-17 18:36:00 <jwh> I get RX ok
2018-12-17 18:36:02 <jwh> its weird
2018-12-17 18:36:19 <AinNero> RX is TX as seen from the other device...
2018-12-17 18:36:33 <jwh> strictly talking from host side
2018-12-17 18:36:56 <jwh> I know the device is ok
2018-12-17 18:37:06 <AinNero> so its not related to alpine?
2018-12-17 18:37:19 <jwh> dunno, first time I've tried to use it since putting alpine on it heh
2018-12-17 18:37:42 <jwh> usb adapter is kinda flaky but normally the oboard port works ok
2018-12-17 18:38:06 <AinNero> how did you check that they are broken? using an loopback thing?
2018-12-17 18:38:38 <jwh> known working devices don't see anything on their RX
2018-12-17 18:38:42 <jwh> hm
2018-12-17 18:39:18 <jwh> just chucked the usb adapter on my laptop, fine there (windows)
2018-12-17 18:40:00 <AinNero> you could also mention how you tried to interface them
2018-12-17 18:40:22 <AinNero> like, invoking screen on it? picocom? stty and cat? getty?
2018-12-17 18:40:50 <jwh> cu, minicom, doesn't really matter, broken on both
2018-12-17 18:41:02 <AinNero> because 95% of all serial issues were misconfigured baud rates, previously
2018-12-17 18:41:19 <jwh> sure, if this was my first rodeo :D
2018-12-17 18:41:19 <AinNero> but whatever.
2018-12-17 18:42:13 <jwh> I finally threw out all my broken adapters (are there actually any decent uart/ttl adapters around anymore that aren't fakes or just crap?)
2018-12-17 18:42:35 <AinNero> i worked previously with ch340, ch341 or other FTDI-clones
2018-12-17 18:43:13 <AinNero> im surprised that you got any broken ones
2018-12-17 18:43:56 <jwh> I think most of them are prolific chips tbh (mix of fake and genuine, fake ones seem to last longer...)
2018-12-17 18:45:43 <jwh> I should really invest in some of those FTDI ones with the micro usb
2018-12-17 18:47:17 <AinNero> also, i wonder why you use the term 'host-side' for serial lines..
2018-12-17 18:47:43 <AinNero> imagine saying that to one end of an nullmodem cable?
2018-12-17 18:48:36 <jwh> host as in the pc running alpine
2018-12-17 18:48:40 <jwh> as opposed to the device
2018-12-17 18:48:51 <jwh> should probably have said "host"
2018-12-17 18:49:27 <AinNero> isn't alpine running in the device you wanted to connect to?
2018-12-17 18:49:46 <jwh> no, alpine is on the box with serial ports/adapters, device is something else
2018-12-17 18:49:56 <jwh> embedded devices
2018-12-17 18:52:34 <AinNero> hm.
2018-12-17 18:53:07 <jwh> oh another uart died, cool
2018-12-17 18:53:14 <AinNero> i learned this by accident, by swerving off topic
2018-12-17 18:53:16 <jwh> 
ACTION finds some FTDI ones

2018-12-17 18:53:48 <jedahan> Whenever the kernel gets updated with apk update I get this warning: `WARNING: Root device is not specified in /etc/update-extlinux.conf`, how do I specify the root device correctly to not get this warning?
2018-12-17 18:53:57 <AinNero> but, at least after 20 minutes of talking, i somewhat have an image of your situation, jwh
2018-12-17 18:54:08 <AinNero> except if there is something you haven't said yet
2018-12-17 18:54:23 <jedahan> should i just set root=/dev/sda4 if thats what I mount as root?
2018-12-17 18:55:19 <jwh> nothing much else to say, need to boot somethig else on it and check but my internet is just awful atm
2018-12-17 18:55:20 <AinNero> jedahan: usually you put the UUID from the disk there
2018-12-17 18:57:11 <jedahan> AinNero so root="a027741f-6526-4a31-8...."
2018-12-17 18:57:58 <AinNero> jedahan: no,
2018-12-17 18:58:13 <AinNero> its like root=UUID=123...
2018-12-17 18:58:17 <jwh> finding genuine chips is still a real nightmare
2018-12-17 18:58:21 <jedahan> Thanks AinNero !
2018-12-17 18:58:38 <AinNero> jedahan: usually setup-disk does this during installation
2018-12-17 18:58:44 <AinNero> as part of setup-alpine
2018-12-17 18:59:13 <jedahan> I think I set mine up using a chroot so maybe there was some shenanegains during that?
2018-12-17 18:59:16 <jedahan> it was a while ago
2018-12-17 18:59:28 <AinNero> if you do it using chroot, then you need to setup that manually
2018-12-17 19:00:06 <jedahan> yeah I mean this is just a box at home for friends to experiment with networking and a decent cpu
2018-12-17 19:00:19 <jedahan> surprised at how usable alpine is as the host for a bunch of things
2018-12-17 19:00:48 <AinNero> with alpine, servers with 256MB or 512MB RAM are suddenly useful again
2018-12-17 19:01:11 <danieli> ~0.00 in load avg on my laptop, with chromium and lots of terminals and stuff open
2018-12-17 19:01:28 <AinNero> danieli: i dont believe you
2018-12-17 19:01:32 <AinNero> especially with chromium
2018-12-17 19:02:26 <danieli> that's understandable, and the load is way higher on my desktop, which runs arch with discord, spotify, chromium, and lots of other stuff
2018-12-17 19:03:06 <danieli> daniel@e554 [~]$ uptime
2018-12-17 19:03:06 <danieli>  20:02:42 up 2 days, 23:12,  0 users,  load average: 0.04, 0.04, 0.03
2018-12-17 19:03:09 <jedahan> 226Mb/16042Mb :shrug:
2018-12-17 19:04:58 <danieli> I also just got "WARNING: Root device is not specified in /etc/update-extlinux.conf." but manually running the grub commands solved it
2018-12-17 19:05:16 <jedahan> we wrote our own `wall` and `users` called `say` and `nearby`
2018-12-17 19:05:21 <jedahan> no need for utmp
2018-12-17 19:05:25 <AinNero> danieli: im at loss here, do you use extlinux or grub?
2018-12-17 19:05:41 <AinNero> or is grub using the extlinux file?
2018-12-17 19:08:43 <danieli> for some reason it's doing the latter
2018-12-17 19:08:44 <danieli> i use grub
2018-12-17 19:31:40 <mps> jwh: do you trying over /dev/ttyUSBx (x = 0 - 9)
2018-12-17 19:32:13 <mps> and as 'root' user
2018-12-17 19:32:45 <mps> i have a lot of prolific usb adapter and all of them works
2018-12-17 19:33:20 <mps> about six, I think (actually not 'a lot' as I just said)
2018-12-17 19:39:27 <jwh> mps: of course
2018-12-17 19:39:30 <jwh> yeah the adapter is fine
2018-12-17 19:41:01 <g1eb> SpaceToast any chance you are online? :)
2018-12-17 19:41:14 <SpaceToast> g1eb: erm, somewhat - I am at work though, why?
2018-12-17 19:41:27 <g1eb> just wanted to thank you again for your help with docker last week!
2018-12-17 19:41:36 <SpaceToast> oh cool, no problem :)
2018-12-17 19:43:13 <g1eb> i wanted to scroll back to see other tips you had for writing proper Dockerfiles
2018-12-17 19:43:44 <g1eb> unfortunately my irc logs don't go as far back as last week :( oh well
2018-12-17 19:43:56 <mps> well, I had sometimes problem but after added ttyUSB to dialout group in /etc/mdev.conf everything is fine, screen works without problem
2018-12-17 19:44:01 <SpaceToast> https://dev.alpinelinux.org/irclogs/
2018-12-17 19:44:20 <SpaceToast> hopefully logs dating back to 2013 are sufficient :)
2018-12-17 19:44:36 <g1eb> oh sweet! I'll use that to go back and check up on them :) thanks again!
2018-12-17 19:44:36 <mps> jwh: even posted patch to aports, but don't know if it is applied
2018-12-17 19:45:11 <jwh> it opens the serial port fine, just doesn't send anything
2018-12-17 19:45:17 <jwh> receive is fine
2018-12-17 19:46:18 <mps> hmm, strange. I had similar problem when the another process run in daemon mode (pppd)
2018-12-17 19:46:30 <mps> on the same port, odc
2018-12-17 19:46:36 <mps> s/odc/ofc/
2018-12-17 19:46:38 <jwh> thats what I thought, but I've made sure theres nothing in inittab, nothing else has the port open
2018-12-17 19:47:33 <mps> did you tried to make loopback, i.e. connect rx and tx on the cable
2018-12-17 19:47:47 <jwh> not yet, will try that in a bit
2018-12-17 19:48:11 <jwh> but since I know the adapter works fine on windows, its not likely to make any difference
2018-12-17 19:48:19 <mps> that is my usual test if serial port doesn't work for no apparent reason
2018-12-17 19:48:48 <mps> you will know if the problem is on the alpine side or something else
2018-12-17 19:49:07 <mps> to narrow your investigation
2018-12-17 20:22:20 <sh-jeff> Does anyone in the channel know if it's possible to use aports to build a git snapshot, like Arch AUR?
2018-12-17 20:23:02 <sh-jeff> Or do I need to clone the repo and take a snapshot from that?
2018-12-17 20:27:14 <danieli> you can cherrypick packages from aports to build, you only need the apkbuild (and local files specified in sources, if any)
2018-12-17 20:28:13 <danieli> for what it's worth, the master branch in the aports repo is essentially what edge is
2018-12-17 20:30:43 <sh-jeff> I'm trying to modify an APK build to build a version that is in a git branch, but doesn't have a release.
2018-12-17 20:30:59 <_ikke_> right
2018-12-17 20:31:10 <sh-jeff> Matter of fact, this is actually the package I'm trying to build https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=zerotier-one-debug-git
2018-12-17 20:31:42 <danieli> do you want master or just a snapshot of master?
2018-12-17 20:32:28 <sh-jeff> Let me check the branch name
2018-12-17 20:32:55 <sh-jeff> I built it manually a week or two ago and need to check my notes.
2018-12-17 20:32:57 <danieli> https://github.com/zerotier/ZeroTierOne/archive/master.tar.gz you can use this for the latest master
2018-12-17 20:33:27 <danieli> ah, screw it, the point is that you can pull tarballs of specific commits or just latest master
2018-12-17 20:34:14 <sh-jeff> Got it
2018-12-17 20:35:10 <sh-jeff> I just need to figure out what to set for for the url?
2018-12-17 20:35:22 <danieli> uh?
2018-12-17 20:35:30 <danieli> you mean $url?
2018-12-17 20:36:03 <sh-jeff> Sorry, the url that goes in source=
2018-12-17 20:36:19 <danieli> depends on what version you want, if you want a specific commit, or latest master
2018-12-17 20:38:04 <nmeum> https://github.com/zerotier/ZeroTierOne/archive/$commit_or_branch_name.tar.gz
2018-12-17 20:38:58 <sh-jeff> thx
2018-12-17 20:39:13 <sh-jeff> that's me not knowing how to use github correctly.
2018-12-17 20:39:41 <sh-jeff> I'm looking to build the dev branch, which I think will work.
2018-12-17 20:45:02 <northox> Hello, quick question. Online I can see a package being only available if I select 'any' architecture (python-dnspython). However, simply calling `apk add <pkgname>` doesn't work. Is there's something special I should do?
2018-12-17 20:48:15 <danieli> online where, exactly?
2018-12-17 20:54:47 <northox> Here: https://www.archlinux.org/packages/?sort=&arch=any&q=dnspython&maintainer=&flagged=
2018-12-17 20:55:00 <pardis> Alpine isn't Arch
2018-12-17 20:55:14 <northox> facepalm
2018-12-17 20:57:44 <northox> Got it. That was stupid.
2018-12-17 21:00:03 <danieli> lol
2018-12-18 04:47:34 <KH405_TV> Could anyone share me their deluged init script please?
2018-12-18 05:35:28 <_ikke_> KH405_TV: Can you show what you tried and what issues you are facing?
2018-12-18 05:41:45 <KH405> I deleted what I have written, but I read what you sent me yesterday and i'm still not figuring it out
2018-12-18 05:43:31 <_ikke_> Did you try to run deluged manually first?
2018-12-18 05:43:47 <KH405> Yes, I finally got it to work
2018-12-18 05:43:51 <KH405> Deluged starts manually
2018-12-18 05:43:58 <_ikke_> ok
2018-12-18 05:44:02 <KH405> Now I want it to make it start while I boot up
2018-12-18 05:44:03 <_ikke_> That's your basis
2018-12-18 05:44:35 <_ikke_> When you start deluged manually, does it remain on the foreground, or is it forking and going in the background?
2018-12-18 05:44:55 <KH405> backgrounf
2018-12-18 05:45:00 <KH405> background *
2018-12-18 05:45:16 <_ikke_> KH405: does it create a pidfile?
2018-12-18 05:45:31 <KH405> Where are they stored?
2018-12-18 05:45:45 <_ikke_> usually somewhere in /run or /var/run
2018-12-18 05:46:49 <KH405> I see nothing mentioning deluge in those folders
2018-12-18 05:47:32 <_ikke_> So you have to find a way to either run deluged in the foreground, or for it to create a pidfile
2018-12-18 05:48:36 <KH405> OK ...
2018-12-18 05:49:02 <_ikke_> https://www.mankier.com/1/deluged
2018-12-18 05:49:17 <_ikke_> https://www.mankier.com/1/deluged#--do-not-daemonize
2018-12-18 05:50:09 <KH405> So deluge -d ?
2018-12-18 05:50:17 <_ikke_> yes
2018-12-18 05:51:12 <KH405> And then?
2018-12-18 05:51:32 <_ikke_> ok, openrc seems to prefer the process to background itself and use pidfiles
2018-12-18 05:52:01 <_ikke_> So you can use deluged -p /run/deluged.pid
2018-12-18 05:52:25 <_ikke_> That should be a capital P
2018-12-18 05:53:28 <KH405> Okay, got it!
2018-12-18 05:53:52 <KH405> Content of deluged.pid is only 453 tough lol
2018-12-18 05:53:55 <_ikke_> ok, so once you know that, you know what to provide in your service file
2018-12-18 05:54:03 <_ikke_> KH405: seems ok
2018-12-18 05:54:31 <_ikke_> command="/path/to/deluged"
2018-12-18 05:55:29 <_ikke_> command_args="-P /run/${RC_SERVICEFILE}.pid"
2018-12-18 05:56:02 <_ikke_> oh, you should also define pidfile=
2018-12-18 05:57:55 <_ikke_> KH405: try this: http://tpaste.us/BoEq
2018-12-18 05:58:58 <KH405> Aight, will give this a try, thanks alot!
2018-12-18 06:00:14 <KH405> IOError: [Errno 2] No such file or directory: '/run//etc/init.d/deluged.pid'
2018-12-18 06:00:55 <_ikke_> sorry, ${RC_SVCNAME}
2018-12-18 06:01:00 <_ikke_> not ${RC_SERVICE}
2018-12-18 06:03:37 <KH405> It "works" ?
2018-12-18 06:03:51 <KH405> I've get the output that the service is stopped but I can't connect to it
2018-12-18 06:04:03 <KH405> Umm I think it's because it's started as root?
2018-12-18 06:04:43 <KH405> Anyway to start it as a specific user ?
2018-12-18 06:05:30 <_ikke_> command_user="user:group"
2018-12-18 06:08:17 <KH405> I added it and Ps dosen't show deluge anymore
2018-12-18 06:10:56 <KH405> RC says it is started, but I can't connect to it
2018-12-18 06:13:14 <_ikke_> if ps doesn't show it, it's most likely not running
2018-12-18 06:13:47 <_ikke_> You can also try to specify a log file and config path file
2018-12-18 06:13:49 <KH405> rc-service start deluged says it's running tough?
2018-12-18 06:14:22 <KH405> Also, for future reference, where did youg et that from ? deluged -p /run/deluged.pid
2018-12-18 06:14:33 <_ikke_> rhttps://www.mankier.com/1/deluged
2018-12-18 06:15:17 <KH405> lol I completly jumped over it! Sorry!
2018-12-18 06:16:03 <KH405> Think I found the cultprit
2018-12-18 06:16:05 <KH405> IOError: [Errno 13] Permission denied: '/run/deluged.pid'
2018-12-18 06:19:07 <_ikke_> hmm
2018-12-18 06:19:33 <_ikke_> I have to go, I'm back later
2018-12-18 06:22:44 <KH405> Aight, will try to figure therest out! Thanks alot again!
2018-12-18 10:29:16 <thresh> hello.  can I use client TLS auth with apk package manager to auth against my repo of packages?
2018-12-18 10:32:02 <_ikke_> apk is using libfetch iirc
2018-12-18 11:21:08 <TBB> and that specific libfetch version doesn't support that
2018-12-18 11:21:15 <TBB> however,
2018-12-18 11:21:28 <TBB> (I'm almost an expert at this very specific issue)
2018-12-18 11:21:54 <TBB> there's a version that uses another version of libfetch that has such support built-in
2018-12-18 11:22:43 <TBB> I don't know for sure whether that version of apk-tools is now the official version though; fabled probably does
2018-12-18 11:26:04 <thresh> oh thanks TBB, it looks like it was implemented in >= 2.8.0 https://git.alpinelinux.org/cgit/apk-tools/commit/?id=52fd85a8dcfee9c93522d80693673bc95cc1caaf
2018-12-18 11:27:02 <thresh> so it should work in 3.8 and thats everything I care about
2018-12-18 16:53:40 <geosmin> getting I/O error when trying to read /dev/ttyS0 on rpi, dmesg is spamming 'ttyS ttyS0: tty_port_close_start: tty->count = 1 port count = 2'
2018-12-18 16:55:06 <mps> geosmin: is it /dev/ttyS0 or it should be /dev/ttyAMA0 on rpi, i'm not sure, just thinking aloud
2018-12-18 16:55:28 <geosmin> mps: on rpi3 it should be ttyS0 IIRC
2018-12-18 17:00:27 <mps> geosmin: i don't have rpi3 to check, so not sure, could be ttyS0
2018-12-18 18:10:59 <d4z3x> hey guys
2018-12-18 18:11:27 <d4z3x> I've done everything to install virtualbox guest additions
2018-12-18 18:11:42 <d4z3x> No go. From this wiki: https://wiki.alpinelinux.org/wiki/VirtualBox_guest_additions
2018-12-18 18:11:55 <d4z3x> I've added the repo, etc.
2018-12-18 18:15:06 <d4z3x> Anyone around to helo?
2018-12-18 18:15:14 <d4z3x> s/helo/help
2018-12-18 18:18:16 <d4z3x> yeah, great help guys
2018-12-18 18:57:48 <maldridge> does alpine recommend any particular configuration management solution?
2018-12-18 19:23:23 <sh-jeff> Can interface names be referred to by regex in awall? The "Zero-to-Awall" wiki page refers to tun+ and says it stands for "any digit," but I don't see that referenced in the user's guide.
2018-12-18 19:24:01 <sh-jeff> I'm assuming that means 0-9, but can a . be used to denote any non-whitepace character?
2018-12-18 19:24:22 <sh-jeff> Or similar
2018-12-18 19:28:33 <sh-jeff> maldrige: I'm not associated with the project, but I would assume ACF https://wiki.alpinelinux.org/wiki/Alpine_Configuration_Framework_Design
2018-12-18 19:29:05 <sh-jeff> I have no idea what it actually is or does though.
2018-12-18 19:45:13 <maldridge> ACF looks pretty slick, but it also appears to only do one system at a time
2018-12-18 19:45:22 <maldridge> I'm looking at things closer to ansible or salt
2018-12-18 19:45:31 <maldridge> or mgmt, but it has an incomprehensible config language
2018-12-18 22:40:02 <Tetsumaki> Hello. Can someone tell me why syslog is running please ? -> https://privatebin.net/?b415d18206bae473#BUfyUxFnxrtn7i4MQlMzVF5n2I7A1Z4VLAADv1Kg3BU=
2018-12-18 22:42:04 <KH405> _ikke_: Hit me up when you got a minute, giving rights to my deluged user never work, unless I did something wrong, but i'm at a loss again :(
2018-12-18 22:42:49 <KH405> Tetsumaki: Don't you want syslog to run ?
2018-12-18 22:42:56 <danieli> Tetsumaki: you should be able to just stop it, but I don't see why you don't want logs
2018-12-18 22:43:06 <danieli> KH405: what exactly is the issue?
2018-12-18 22:43:53 <Tetsumaki> I want him to start but i want to understand why he then starts
2018-12-18 22:44:14 <Tetsumaki> rc-update -v show
2018-12-18 22:44:27 <Tetsumaki> syslog is not set to default
2018-12-18 22:44:59 <KH405> danieli: Creating a init script in general, I receive this error when I try to run it as something else than root IOError: [Errno 13] Permission denied: '/run/deluged.pid'
2018-12-18 22:57:45 <danieli> Tetsumaki: i believe it's set to boot by default
2018-12-18 22:57:58 <danieli> is it in your case?
2018-12-18 22:58:15 <Tetsumaki> no danieli
2018-12-18 22:58:50 <Tetsumaki> # rc-update | grep syslog --> no output
2018-12-18 22:58:57 <danieli> KH405: you could probbaly rip off this https://gitweb.gentoo.org/repo/gentoo.git/tree/net-p2p/deluge/files/deluged.init-2
2018-12-18 23:01:23 <danieli> Tetsumaki: it might just start if it's a dependency of other services, i don't remember how it works but that's what makes sense
2018-12-18 23:01:27 <danieli> i don't see why you'd want to turn it off
2018-12-18 23:01:45 <danieli> $ grep -r "use logger" /etc/init.d/ | wc -l
2018-12-18 23:01:45 <danieli> 6
2018-12-18 23:02:16 <danieli> s/probbaly/probably/
2018-12-18 23:03:23 <Tetsumaki> hum, ok danieli
2018-12-18 23:03:26 <Tetsumaki> 3
2018-12-18 23:04:10 <KH405> danieli: Yeah I checked this script out like last week and it assumes plenty of thing that I don't really understand, as I never really done anything with openRC, but do you think it would work ?
2018-12-18 23:04:40 <KH405> Where is this defined 	if [ "${DELUGED_USER}" = "" ] ; then ?
2018-12-18 23:05:02 <Tetsumaki> danieli: i dont want stop syslog but only understand
2018-12-18 23:05:18 <danieli> KH405: https://gitweb.gentoo.org/repo/gentoo.git/tree/net-p2p/deluge/files/deluged.conf-2
2018-12-18 23:05:51 <maldridge> I believe cfengine in 3.8 is broken, where should I report this to?
2018-12-18 23:05:53 <Tetsumaki> unlike windows or systemd, we can understand
2018-12-18 23:06:01 <Tetsumaki> and i like that
2018-12-18 23:06:07 <danieli> maldridge: bugs.alpinelinux.org
2018-12-18 23:06:14 <maldridge> danieli: thnx
2018-12-18 23:06:17 <danieli> yw
2018-12-18 23:06:35 <danieli> please do provide some tracebacks/logs/the reason you think it's broken
2018-12-18 23:06:49 <Tetsumaki> and thanks danieli ;)
2018-12-18 23:07:40 <maldridge> sure, I'm confirming now that it actually is, but it looks like its missing a number of files that are required to make it work
2018-12-18 23:07:50 <maldridge> I'll try building from source before raising an issue
2018-12-18 23:08:00 <danieli> if you explore it a bit, that'd be lovely
2018-12-18 23:08:17 <maldridge> can you point me to where I can find the "template" for cfengine?
2018-12-18 23:08:22 <maldridge> dunno what aports calls those
2018-12-18 23:08:35 <maldridge> I want to build mine with the same options
2018-12-18 23:08:45 <danieli> aha, the APKBUILD
2018-12-18 23:08:47 <danieli> just use abuild for it
2018-12-18 23:08:52 <danieli> https://git.alpinelinux.org/cgit/aports/tree/community/cfengine/APKBUILD
2018-12-18 23:08:54 <KH405> And where is this file suppose to be exactly ?
2018-12-18 23:12:27 <danieli> KH405: the confd file? look at line 1 :)
2018-12-18 23:15:13 <KH405> Alright! will try that out after eating, thanks alot!
2018-12-18 23:16:43 <maldridge> danieli: I think its an interaction with the subpkg that's broken, but I'm still playing with it
2018-12-18 23:16:59 <danieli> maldridge: what's the expected behavior, and how does it deviate from it?
2018-12-18 23:17:27 <maldridge> the expected behavior is that cfengine works, the observed behavior is that the masterfiles aren't in the correct location
2018-12-18 23:18:07 <danieli> $builddirmasterfiles, what a version
2018-12-18 23:18:24 <maldridge> I'm on Alpine 3.8
2018-12-18 23:18:25 <danieli> so cfengine-masterfiles practically just puts them in the wrong location
2018-12-18 23:18:37 <danieli> where exactly are they installed, and where are they supposed to be?
2018-12-18 23:19:25 <danieli> right, it might just have been fixed in edge, who knows
2018-12-18 23:19:46 <maldridge> with the understanding that I'm still learning this, the error expects them to be at /var/lib/cfengine/inputs/promises.cf for the entrypoint
2018-12-18 23:19:54 <maldridge> I am happy to try from edge if you'd like
2018-12-18 23:20:00 <maldridge> though I am hesitant to run edge in prod
2018-12-18 23:20:46 <danieli> the only thing that has been changed recently is the package version being bumped
2018-12-18 23:20:54 <danieli> so it's probably still broken in edge
2018-12-18 23:21:24 <maldridge> this also expects some binaries in different places, so I'm inclined to think its just broken
2018-12-18 23:22:01 <danieli> it doesn't look like we're doing any funky business with the binaries
2018-12-18 23:22:20 <maldridge> yeah, I'm not sure why this expects binaries to live in /var/lib
2018-12-18 23:25:51 <maldridge> still in the process of building to see if this can be reproduced in an external build
2018-12-18 23:28:25 <danieli> masterfiles are installed to /var/lib/cfengine/materfiles
2018-12-18 23:29:00 <maldridge> yes, but that's not where the tools look
2018-12-18 23:29:46 <danieli> well... that's dumb
2018-12-18 23:30:10 <danieli> we either have to patch it to look there, or install to the proper location
2018-12-18 23:30:47 <maldridge> my personal opinion is to install the the correct location
2018-12-18 23:30:52 <maldridge> this is several tools to patch to keep them moved
2018-12-18 23:31:01 <danieli> indeed, might as well do it properly
2018-12-18 23:31:28 <maldridge> I question if the subpkg is necessary here, but that is my opinion as a void-dev
2018-12-18 23:32:16 <danieli> we generally split stuff up a bit more, i don't have an opinion on this exact package, i haven't used cfengine a lot
2018-12-18 23:32:35 <maldridge> neither have I, but its in tree, so I'm giving it a try before I look for something else
2018-12-18 23:32:42 <maldridge> I also like how small it is
2018-12-18 23:56:19 <maldridge> danieli: I'll keep working on this, something doesn't seem right with the build, but also with the cfengine tarball itself
2018-12-19 02:16:44 <KH405> danieli: Also, if i want to keep the config at the default location can I just delete the last line in the second file you sent ?
2018-12-19 02:17:15 <danieli> KH405: uh? which one?
2018-12-19 02:20:05 <KH405> This: DELUGED_HOME="/var/lib/deluge"
2018-12-19 02:20:25 <danieli> I'd rather set it to the correct path, if it's incorrect
2018-12-19 02:20:56 <KH405> Is there a correct path ?
2018-12-19 02:23:17 <danieli> I don't use deluge so I'd have to check, but I'm a little busy right now
2018-12-19 02:23:55 <KH405> It's all good, my point is you can store the config everywhere and the default path is /home/user/.config/deluge
2018-12-19 02:24:09 <KH405> I just think they changed it to /var/lib/deluge for some reason
2018-12-19 02:24:10 <danieli> it *probably* has a local and a global config
2018-12-19 02:25:20 <KH405> I ran deluge a few times and I have nothing in /var/lib
2018-12-19 02:25:44 <danieli> if you're running deluge as a service, it doesn't make sense to read a config file in $HOME
2018-12-19 02:26:36 <KH405> Why not :O ?
2018-12-19 02:26:46 <KH405> We can talk about this later if you're busy tough ^^
2018-12-19 02:26:54 <KH405> I'm just intrigued by what you said
2018-12-19 02:26:57 <danieli> just a little preoccupied, still on/off available in chat
2018-12-19 02:27:24 <KH405> Ohh, just ping me when you'll be done ^^
2018-12-19 02:27:50 <danieli> you can just check the contents of the tarball / the config / initd
2018-12-19 02:27:55 <danieli> or make an issue at bugs.alpinelinux.org
2018-12-19 02:29:05 <danieli> I don't use or know deluge too well, so in order to know, I'd have to research it for myself
2018-12-19 02:40:19 <maldridge> danieli: cfengine masterfiles are in the canonically correct place, but --enable-fhs may have a bug
2018-12-19 02:40:36 <maldridge> so this is at this time unlikely to be an alpine bug
2018-12-19 02:41:02 <danieli> maldridge: if it works on other platforms, it *might* be a side-effect of something that's unique to alpine
2018-12-19 02:41:13 <danieli> be it musl, some alpine configuration, or something else
2018-12-19 02:41:19 <danieli> or it's just plain broken upstream too
2018-12-19 02:41:42 <maldridge> yeah, I'm looking to see if void's is broken in the same way
2018-12-19 02:41:56 <maldridge> the problem is its missing where it put the bins
2018-12-19 02:43:16 <danieli> "its missing where it puts the bins"?
2018-12-19 02:44:50 <maldridge> sorry that could have done with more words
2018-12-19 02:45:09 <maldridge> the executables wind up in /usr/bin as expected, but cfengine expects them to be in /var/lib/cfengine/bin
2018-12-19 02:45:22 <danieli> that's a bit strange
2018-12-19 02:45:34 <danieli> did you build it with the binaries installed in the path cfengine expects them?
2018-12-19 02:45:47 <danieli> s/did you/did you try/
2018-12-19 02:45:57 <danieli> ah, screw it, you know what I meant
2018-12-19 02:47:00 <maldridge> yes, and the alpine version works if you symlink them
2018-12-19 02:47:16 <maldridge> so the bins themselves are fine, there's just something wrong with the pathing
2018-12-19 03:26:57 <KH405> danieli: I might have explained myself wrong, this is what interrested me if you're running deluge as a service, it doesn't make sense to read a config file in $HOME
2018-12-19 03:27:03 <KH405> Why ?
2018-12-19 03:27:19 <danieli> because homes are meant to be private to users, if you had ran deluge as a regular user, it would make sense
2018-12-19 03:27:23 <danieli> but it's a system service
2018-12-19 03:27:44 <danieli> system configuration, as you know, resides in /etc
2018-12-19 03:32:08 <KH405> Ohhh! I also seen alot of people putting it in /var/lib
2018-12-19 03:35:56 <danieli> `man hier`
2018-12-19 03:38:13 <KH405> Damn, definetly bookmarking that page :P
2018-12-19 03:53:43 <danieli> page?
2018-12-19 03:53:53 <danieli> oh right, you're probably looking at it online
2018-12-19 03:54:05 <danieli> it varies with OSes, so you should just use the actual `man` program
2018-12-19 04:18:30 <KH405> Okay ^^ I'm still wondering why the official gentoo wiki says to place it in /var/lib tough ^^
2018-12-19 04:21:51 <jwh> hmm
2018-12-19 04:22:12 <jwh> 
ACTION builds a (U)EFI image

2018-12-19 04:23:52 <danieli> you may have to adapt it
2018-12-19 04:23:56 <danieli> @ KH405
2018-12-19 04:24:05 <KH405> Adapt what ?
2018-12-19 04:24:11 <danieli> the files from gentoo, if you decide to use them
2018-12-19 04:38:44 <KH405> Yeah but I mean, /var/lib in gentoo is suppose to be the same in alpine ^^
2018-12-19 04:38:56 <KH405> Why are they putting the config there :P
2018-12-19 04:39:50 <_ikke_> KH405: I think it's just a consequence of putting the home there
2018-12-19 04:40:37 <KH405> Ohh I see, and how will I go to put my home there if my user is already created?
2018-12-19 04:40:48 <danieli> for deluge?
2018-12-19 04:49:05 <danieli> @ KH405
2018-12-19 04:50:58 <KH405> well the user deluge
2018-12-19 04:51:06 <KH405> I want to change his ''home'' folder location
2018-12-19 04:51:19 <KH405> Sorry, I'm also doing alot of things at the same time ^^'
2018-12-19 04:52:10 <_ikke_> easiest is to just edit /etc/passwd
2018-12-19 04:53:22 <_ikke_> alternatively, you can install the shadow package, and use usermod -m /path/to/new/home <username>
2018-12-19 04:53:59 <KH405> Okay, and for a user that only is gonna run service, is bin:ash okay ?
2018-12-19 04:54:24 <_ikke_> either that, or not even a valid shell at all
2018-12-19 04:54:52 <_ikke_> /bin/false or /sbin/nologin if you have util-linux installed
2018-12-19 04:56:51 <KH405> Good, thanks!
2018-12-19 04:57:29 <KH405>  /bin/false needs utli-linux?
2018-12-19 05:04:29 <danieli> only coreutils provides /bin/false by the looks of it _ikke_
2018-12-19 05:10:24 <KH405> Good, thanks to you two!
2018-12-19 05:10:43 <KH405> Also one more for the road, deluge:x:1001:1001:Linux User,,,:/home/deluge:/bin/ash what does ''Linux User,,,'' means in this line
2018-12-19 05:21:10 <danieli> iirc default full name
2018-12-19 05:21:56 <danieli> field 5 (colon delimited) is the info field
2018-12-19 07:16:04 <KH405> So I can put anything there ?
2018-12-19 07:18:22 <danieli> nearly, check the man pages
2018-12-19 07:19:03 <danieli> don't see why you would though, except full name on accounts for real people
2018-12-19 07:43:00 <KH405> Cause deluge:deluge:deluge in case I forget
2018-12-19 07:43:00 <KH405> lol
2018-12-19 07:47:50 <danieli> seems utterly pointless, and field 5 is comma separated
2018-12-19 15:14:12 <tdtrask> What is the recommended desktop for alpine these days? I'm installing a new box from scratch
2018-12-19 15:14:24 <tdtrask> Looks like the Mate doc has gotten some love recently
2018-12-19 15:14:49 <tdtrask> And, I don't see Plasma support imported from Adelie yet
2018-12-19 15:15:13 <kahiru> waaait, adelie has plasma?
2018-12-19 15:18:28 <tdtrask> kahiru: don't know the current state, but I know they were working on it
2018-12-19 15:18:54 <kahiru> ah, that would be definitely nice to have
2018-12-19 15:20:02 <tdtrask> XFCE is the only one I've tried before on alpine
2018-12-19 15:20:16 <tdtrask> maybe I'll stick with that
2018-12-19 15:53:06 <fcolista> I'm using mate with satisfaction
2018-12-19 15:56:16 <tdtrask> thanks fcolista
2018-12-19 16:05:31 <dalias> xfce is the least-awful "DE" i've tried anywhere
2018-12-19 16:06:28 <dalias> i tried lxde (not on alpine) at one point but despite being promoted as "light" it seemed to use more resources than lxde, had comparable levels of klunkiness, and seems unmaintained now
2018-12-19 16:33:44 <tdtrask> dalias: thanks
2018-12-19 16:34:07 <tdtrask> finished following the Mate doc and wound up with a black screen with cursor in upper left :(
2018-12-19 17:50:14 <dalias> something is wrong with cgit on alpine server:
2018-12-19 17:50:27 <dalias> see https://git.alpinelinux.org/cgit/aports/plain/main/sdl/SDL-1.2.15-const_XData32.patch for example
2018-12-19 17:50:50 <dalias> rather than a raw patch i get html header/footer junk with a patch in the middle
2018-12-19 17:52:47 <danieli> dalias: an invalid patch was pushed to git, by the looks of it
2018-12-19 17:52:58 <danieli> https://git.alpinelinux.org/cgit/aports/tree/main/sdl/SDL-1.2.15-const_XData32.patch
2018-12-19 17:53:01 <danieli> cc ncopa
2018-12-19 17:53:26 <dalias> oh i see it's in the content
2018-12-19 17:53:27 <dalias> never mind
2018-12-19 17:54:05 <dalias> well never mind on the issue i reported; there's a different problem that's actually an issue :)
2018-12-19 17:54:42 <dalias> lol it's been there since 2013 :)
2018-12-19 17:54:57 <dalias> and ncopa committed it
2018-12-19 17:55:13 <danieli> shit happens
2018-12-19 17:55:16 <dalias> yeah
2018-12-19 17:55:24 <dalias> i'm just surprised nothing broke
2018-12-19 17:56:08 <dalias> i was poking around with SDL1 source because i've been having trouble with it not setting fullscreen modes that should be available
2018-12-19 17:56:24 <dalias> looks like for some reasons xrandr is hardcoded off by default in it
2018-12-19 17:58:35 <danieli> seems odd that the patch applies and sdl builds
2018-12-19 18:02:10 <dalias> patch is usually forgiving about malformed crap around the patch
2018-12-19 18:02:14 <dalias> since it's made to handle emails
2018-12-19 18:02:23 <danieli> indeed
2018-12-19 18:02:39 <dalias> but i bet some patch versions would break on it
2018-12-19 18:02:51 <dalias> i only noticed it trying to manually apply
2018-12-19 18:23:23 <geosmin> looking for a telnet client, doesn't seem to be much in the repos
2018-12-19 18:24:27 <_ikke_> Anyone knows where I can find the documentation for alpines / busyboxs network/interfaces? Specifically I'm looking for the documentation on bridge-ports
2018-12-19 18:29:10 <Tetsumaki> geosmin: curl ?
2018-12-19 18:29:40 <danieli> geosmin: netcat?
2018-12-19 18:29:43 <_ikke_> geosmin: busybox-extras
2018-12-19 18:29:50 <danieli> and busy--
2018-12-19 18:29:56 <danieli> yes, that :) an actual telnet client
2018-12-19 18:30:56 <Tetsumaki> a lot of choices
2018-12-19 18:32:33 <dalias> geosmin, seen the news lately? most telnet clients have major vulns
2018-12-19 18:32:58 <dalias> (rce, not just the usual telnet being exposed)
2018-12-19 18:33:05 <geosmin> dalias: i appreciate the warnings but this is just to connect to a piece of industrial equipment
2018-12-19 18:33:13 <geosmin> rce. lovely.
2018-12-19 18:34:43 <danieli> ics. lovely.
2018-12-19 18:38:21 <dalias> :)
2018-12-19 18:38:36 <dalias> i would just use the one in busybox
2018-12-19 18:38:54 <dalias> alpine omits it these days for size
2018-12-19 18:39:04 <dalias> but you can install the busybox-extras package to get it
2018-12-19 18:39:38 <danieli> _ikke_: i'd think it relates to ifupdown, not really sure in alpine
2018-12-19 18:39:59 <danieli> for some reason i'm not seeing any busybox*-doc packages
2018-12-19 18:58:14 <_ikke_> danieli: yes, it's busybox ifup / ifdown, but it's barely documented
2018-12-19 18:58:40 <danieli> _ikke_: yeah, it's better to just look at the code tbh
2018-12-19 19:27:57 <_ikke_> busybox source is not very enlightening
2018-12-19 19:33:16 <maldridge> you say this as though you are surprised
2018-12-19 19:34:34 <_ikke_> no, not really
2018-12-19 19:36:04 <maldridge> libcrypto seems to own a config file that is also owned by libressl
2018-12-19 19:36:18 <maldridge> I understand that things are in flux, but is there a way to suppress this error in the meantime
2018-12-19 20:43:24 <greyeax> hey there, is the virtualbox-guest-modules-grsec package deprecated?
2018-12-19 20:43:27 <greyeax> i cant seem to find it
2018-12-19 20:44:37 <_ikke_> greyeax: correct, alpinelinux no longer uses grsec
2018-12-19 20:45:06 <_ikke_> greyeax: you probably want this: https://pkgs.alpinelinux.org/package/edge/community/x86_64/virtualbox-guest-modules-vanilla
2018-12-19 20:45:32 <_ikke_> if you have alpine 3.7, it's this one
2018-12-19 20:46:53 <_ikke_> https://pkgs.alpinelinux.org/package/v3.7/community/x86_64/virtualbox-guest-modules-hardened
2018-12-19 20:49:27 <greyeax> _ikke_: I'm actually using 3.8.1 - would it be the same as 3.7?
2018-12-19 20:50:44 <_ikke_> no
2018-12-19 20:50:58 <_ikke_> than you need vanilla as well
2018-12-19 20:51:06 <_ikke_> https://pkgs.alpinelinux.org/packages?name=virtualbox-gues*&branch=v3.8&arch=x86_64
2018-12-19 20:56:12 <ncopa> dalias: re SDL patch: *facepalm*
2018-12-19 20:56:33 <greyeax> thanks!
2018-12-19 21:05:54 <maldridge> quick bump on if there's any way to suppress libressl/openssl related errors
2018-12-19 21:07:46 <_ikke_> maldridge: what packages do you have that still depend on libressl?
2018-12-19 21:08:24 <maldridge> I think its cf-engine, but I'm trying again in a clean docker container, for reasons unknown edge became enabled in the vm I was testing in
2018-12-19 21:09:40 <maldridge> right, doing this in a clean env doesn't raise errors
2018-12-19 21:09:47 <maldridge> so this is a VM error in the vagrant box I was using
2018-12-19 23:01:46 <dalias> ncopa, :)
2018-12-19 23:46:01 <jwh> hmm, no dhcpv6 stuff on the iso?
2018-12-19 23:46:02 <jwh> :(
2018-12-19 23:46:43 <jwh> nm
2018-12-19 23:49:53 <jwh> doesn't matter, can't do v6 only setup anyway :(
2018-12-19 23:53:49 <jwh> aha! found a mirror that has v6
2018-12-20 00:07:35 <maldridge> ncopa: the cfengine package which I believe you maintain has some serious problems w.r.t. lmdb, namely its broken
2018-12-20 00:08:02 <maldridge> I think you're already aware of this, but wanted to reach out and ask if you had any workarounds to fix the package
2018-12-20 00:10:35 <maldridge> I'm looking right now at trying to rebuild it to see if that will fix the problem, but I'm kind of stuck trying to figure out aports
2018-12-20 00:48:30 <georgios> so.. what happened to grsec & alpine, grsec public betas etc etc?
2018-12-20 00:50:50 <dalias> grsec still exists? :)
2018-12-20 00:50:53 <dalias> </troll>
2018-12-20 00:51:07 <georgios> hi dalias
2018-12-20 00:51:15 <georgios> musl still exists?
2018-12-20 00:51:23 <georgios> </xtroll>
2018-12-20 00:55:23 <dalias> last i checked we're shipping genuine FOSS under a permissive license, not GPL-violating derivative works of the kernel :)
2018-12-20 01:04:50 <georgios> true
2018-12-20 01:05:44 <georgios> but then again grsecurity is just a patch...
2018-12-20 01:06:36 <georgios> which apparently nobody leaks so that it seems it is in parallel universe
2018-12-20 01:06:43 <georgios> a*
2018-12-20 01:07:06 <georgios> strange story in the world of FOSS
2018-12-20 01:08:28 <georgios> it seems the only way to uncover it either pay for a subscription or buy an appliance using it and demanding a copy of the source!
2018-12-20 02:27:47 <Greyfn> Anybody know what /etc/udhcpd.conf relates to? udhcpd doesn't appear to be part of busybox, and there is no apk available for udhcpd either
2018-12-20 02:28:21 <Greyfn> /etc/udhcpd.conf appears to be part of the base install too
2018-12-20 04:44:33 <danieli> Greyfn: it is part of busybox-extras
2018-12-20 04:45:14 <danieli> udhcpd (server) in busybox-extras and udhcpc (client) in busybox
2018-12-20 05:04:20 <jwh> hm
2018-12-20 05:23:50 <jwh> wish they'd hurry up and release grub 2.03
2018-12-20 06:01:01 <jwh> I'd *really* like bootable f2fs
2018-12-20 07:12:53 <ncopa> maldridge: what is the problem with cfengine?
2018-12-20 07:13:08 <maldridge> there's an as yet unidentified problem with lmdb
2018-12-20 07:13:12 <maldridge> but mainly it doesn't work
2018-12-20 07:13:31 <maldridge> the problem manifests as a failure of lmdb to obtain a transaction lock
2018-12-20 07:13:41 <ncopa> ok, i may have heard it before, but i cannot find any bug report on it
2018-12-20 07:13:55 <maldridge> I believe you got an email on this from a cfengine dev some time ago
2018-12-20 07:14:01 <maldridge> they stumbled onto it before me
2018-12-20 07:14:01 <ncopa> ah, ok
2018-12-20 07:16:14 <ncopa> https://tracker.mender.io/browse/CFE-2936
2018-12-20 07:16:27 <maldridge> yep, that would be the bug and error
2018-12-20 07:16:31 <ncopa> seems to be 2 problems, one seems like a packaging problem
2018-12-20 07:16:49 <maldridge> right, the fhs one though is in my opinion a cfengine problem
2018-12-20 07:16:55 <maldridge> since it shouldn't be expecting binaries in /var
2018-12-20 07:17:22 <ncopa> agree
2018-12-20 07:18:17 <maldridge> I've thus far been fixing it with symlinks, I'm not sure if this is an acceptable solution for packaging
2018-12-20 07:24:15 <ncopa> seems like arch linux adds symlinks while suse has a patch
2018-12-20 07:24:30 <ncopa> i suppose symlinks is the easiest
2018-12-20 07:25:47 <maldridge> the cfengine people seemed open to either earlier today when I spoke with them
2018-12-20 07:36:48 <ncopa> maldridge: what alpine version did you use?
2018-12-20 07:36:57 <ncopa> ncopa-desktop:~$ cf-agent -Kf ./hello.cf
2018-12-20 07:36:57 <ncopa> R: hello from alpine
2018-12-20 07:37:31 <maldridge> I'm on alpine 3.8
2018-12-20 07:39:33 <ncopa> what architecture?
2018-12-20 07:39:36 <ncopa> i cannot reproduce
2018-12-20 07:39:47 <ncopa> # cf-agent -Kf ./hello.cf
2018-12-20 07:39:47 <ncopa> R: hello from alpine
2018-12-20 07:39:58 <ncopa> thats in alpine:3.8 docker
2018-12-20 07:40:48 <maldridge> x86_64
2018-12-20 07:40:56 <maldridge> full vm, not a container
2018-12-20 07:49:04 <ncopa> can you run the testcase with strace?
2018-12-20 07:49:47 <maldridge> I can in the morning, its quite late here
2018-12-20 07:49:57 <maldridge> any particular things you want to see in a test?
2018-12-20 07:53:50 <ncopa> just want to try figure why it ends up with error in lmdb, if a syscall is failing
2018-12-20 07:56:19 <maldridge> ok
2018-12-20 07:56:27 <maldridge> I'll see what I can come up with
2018-12-20 07:56:43 <maldridge> do you want me to update first or are you holding off on the symlink one until you have the strace?
2018-12-20 08:02:09 <georgios> ncopa: ping
2018-12-20 08:05:37 <ncopa> i can deal with the first, which is relatively easy to fix
2018-12-20 08:05:49 <ncopa> i tried the patch from suse, but then it looked in my $HOME
2018-12-20 08:05:57 <ncopa> i think i'll just create symlinks
2018-12-20 08:06:05 <ncopa> georgios: pong
2018-12-20 08:06:25 <danieli> iirc georgios was wondering about why we dropped grsec (it wasn't really grsec, it was linux-hardened, an unofficial port)
2018-12-20 08:06:29 <danieli> he left before i could reply
2018-12-20 08:06:33 <danieli> s/he/they/
2018-12-20 08:18:18 <maldridge> ncopa: it will look in your home if it doesn't run as root
2018-12-20 08:18:33 <maldridge> I'll look into the strace tomorrow morning
2018-12-20 08:23:28 <ncopa> georgios: we dropped grsec because upstream devs removed it from public
2018-12-20 08:24:00 <ncopa> we ran an unofficial port of it for a while, be we always knew we relied on upstream
2018-12-20 08:24:33 <ncopa> upstream used to release "testing" patches for new kernels, which was the base of our unofficial fork
2018-12-20 08:24:38 <ncopa> they stopped with that too
2018-12-20 08:26:11 <ncopa> grsecurity is not for general consumption anymore
2018-12-20 08:26:46 <ncopa> and even if you buy the patch, you can (in practice) only use it internally for yourself and not redistribute it
2018-12-20 08:27:05 <ncopa> because if you do, you need to redist the patch (due to GPL)
2018-12-20 08:27:18 <ncopa> so they dont sell the patch if you redist
2018-12-20 08:29:56 <georgios> so basically i become a client using grsec in a service
2018-12-20 08:30:15 <georgios> i ve seen on arch a linux-hardened package
2018-12-20 08:30:29 <georgios> and i was wondering what it constituted off
2018-12-20 08:31:10 <danieli> this https://github.com/anthraxx/linux-hardened
2018-12-20 08:33:23 <georgios> i dont see any clue of grsecurity there
2018-12-20 08:33:28 <georgios> but really
2018-12-20 08:33:47 <georgios> is it so impossible for the community to fork the damn thing?
2018-12-20 08:33:58 <danieli> they closed it
2018-12-20 08:34:01 <danieli> it's not publicly accessible anymore
2018-12-20 08:34:29 <danieli> https://grsecurity.net/passing_the_baton.php
2018-12-20 08:34:44 <georgios> there was a time that when i wanted a good patch+kernel i would go to alpine's git
2018-12-20 08:35:13 <georgios> so alpine still has a good copy of it for a given kernel
2018-12-20 08:36:12 <georgios> why doesnt vanilla incorporate it...
2018-12-20 08:36:27 <danieli> incoporate what?
2018-12-20 08:36:34 <georgios> the patch
2018-12-20 08:36:44 <danieli> which one are you aiming at?
2018-12-20 08:37:11 <georgios> the  latest stable linux-grsec of alpine
2018-12-20 08:37:22 <georgios> the last even
2018-12-20 08:37:39 <danieli> it isn't grsec - it's an unofficial port
2018-12-20 08:37:45 <danieli> and as ncopa said, we dropped it for various reasons
2018-12-20 08:38:38 <danieli> in addition to that, vanilla is supposed to be vanilla (ordinary/standard linux), linux-hardened was the hardened counterpart
2018-12-20 08:39:08 <georgios> and i wonder... why dont the kernel fellas adopt it?
2018-12-20 08:39:13 <danieli> ask them
2018-12-20 08:39:23 <georgios> isnt it wort it?
2018-12-20 08:39:29 <_ikke_> georgios: fyi, they tried to port the patch to newer versions, but failed, mostly due to spectre/meltdown fixes
2018-12-20 08:39:43 <georgios> would it boost standard linux' security
2018-12-20 08:40:21 <danieli> grsec hasn't been public for a while now
2018-12-20 08:40:27 <danieli> also, politics
2018-12-20 08:44:03 <ncopa> georgios: there are various reasons why upstream does not accept grsecurity as it is
2018-12-20 08:44:13 <ncopa> grsec patch is huge, 5MB IIRC
2018-12-20 08:44:30 <ncopa> you cannot review a 5MB patch
2018-12-20 08:44:48 <ncopa> you cannot merge it as is and expect it to not break anything for anyone
2018-12-20 08:44:58 <ncopa> so linus wanted it submitted in peaces
2018-12-20 08:45:26 <ncopa> grsec devs misunderstood a bit linus meant and got offended
2018-12-20 08:45:32 <_ikke_> And I guess that grsec looses it revenue model once they upstreamed the patch
2018-12-20 08:45:40 <ncopa> linus called it "garbage", "useless"
2018-12-20 08:46:09 <ncopa> grsec devs also think its unfair that they do the huge amount of work to split it in pieces unpaid
2018-12-20 08:46:28 <ncopa> and nobody (afaik) offerend to pay them to do it
2018-12-20 08:46:30 <_ikke_> ncopa: But they find it fair they get a linux kernel for free
2018-12-20 08:47:08 <ncopa> int their opinion, they dont "get a linux kernel", they "fix a linux kernel"
2018-12-20 08:47:39 <_ikke_> still they benefit from it
2018-12-20 08:47:40 <ncopa> i think they were envious that big companies pays kernel devs, but nobody pays grsec
2018-12-20 08:47:51 <danieli> indeed
2018-12-20 08:48:15 <ncopa> then there was some google sponsored project, KSPP, that started to do that work, split up the grsec kernel
2018-12-20 08:48:20 <danieli> i find it odd that it's derivative work (of the GPL'd Linux kernel) yet is allowed to live in private
2018-12-20 08:48:26 <ncopa> and submit it in pieces to upstream
2018-12-20 08:48:35 <ncopa> that offended grsecurity even more
2018-12-20 08:48:54 <ncopa> they claimed that someone else get the credit for their inventions
2018-12-20 08:49:10 <ncopa> someone else get paid for their work (eg the work to upstream it)
2018-12-20 08:49:27 <ncopa> and they were unhappy with how it was done too
2018-12-20 08:49:54 <ncopa> so they got offended that the thing merged into mainline was called grsecurity (eg trademark issue)
2018-12-20 08:50:15 <ncopa> because in their opinion it was a broken/inferior version of their work
2018-12-20 08:50:49 <ncopa> and even if they didnt say it out loud, i also think that they got offended because it threatened their business model
2018-12-20 08:51:32 <ncopa> their way to get funding is to sell the patch expensive
2018-12-20 08:52:47 <ncopa> in any case, what grsecurity does is extremely complicated, and its very difficult to get it right
2018-12-20 08:53:07 <ncopa> i doubt any person except the grsecurity devs are able to make it properly
2018-12-20 08:54:55 <_ikke_> and as a fork, grsec can do whatever they want
2018-12-20 08:55:05 <danieli> it is indeed very difficult, and they've been quite the divas in public
2018-12-20 08:55:08 <_ikke_> if you want to upstream things, concessions have to be made
2018-12-20 08:57:53 <danieli> indeed, especially with such a huge patchset
2018-12-20 08:58:08 <ncopa> i think that you should always try upstream your patches
2018-12-20 08:58:17 <ncopa> however, grsecurity is the exception
2018-12-20 08:58:18 <_ikke_> unless you want to sell them
2018-12-20 08:58:35 <ncopa> the grsecurity is an exception, and i think it makes sens to keep it separate
2018-12-20 08:59:24 <_ikke_> I think a major part in this is that you can opt-in into the patch, but if for some reason it breaks something, you can choose to use vanilla
2018-12-20 08:59:40 <_ikke_> So you can make choices that upstream can never make
2018-12-20 08:59:54 <_ikke_> at least, not easily
2018-12-20 09:00:54 <danieli> it does make sense to keep them, and it does make sense to keep it proprietary if you want to make money off it, but it's derivative work so I'm not sure why it's "okay"
2018-12-20 09:01:57 <_ikke_> danieli: from what I understood from gpl is that you only have to provide the source to the users of your work when they ask for it
2018-12-20 09:02:07 <_ikke_> you don't have to make it publically available
2018-12-20 09:03:15 <ncopa> brb
2018-12-20 09:03:18 <_ikke_> But one point someone made in a similar discussion is that you cannot apply any extra restrictions to the users of your derivative works
2018-12-20 09:03:46 <_ikke_> And their point was not being alowed to distribute the patch (by refusing service if you do) is such a restriction
2018-12-20 09:04:36 <georgios> @_ikke_ | you don't have to make it publically available
2018-12-20 09:04:41 <georgios> exactly
2018-12-20 09:04:51 <georgios>  @_ikke_ | And their point was not being alowed to distribute the patch (by refusing service if you do) is such a restriction
2018-12-20 09:04:57 <georgios> illegal
2018-12-20 09:05:39 <georgios> we could raise funds to get our hands on grsec and leak it :p
2018-12-20 09:06:17 <_ikke_> georgios: but you would not get any newer versions after that
2018-12-20 09:06:20 <georgios> but then... who could fork it??
2018-12-20 09:06:39 <georgios> i beleive we are saying the same thing
2018-12-20 09:07:23 <danieli> we're geeks, not lawyers. and honestly, I'm fine with just using Linux vanilla instead
2018-12-20 09:22:04 <georgios> i am a bit of a lawyer when it comes to fair use of s/w and respect to the original author(s) will
2018-12-20 09:23:00 <georgios> ncopa, could you hand me your last grsec stable patch. just for archival
2018-12-20 09:24:11 <_ikke_> should be still in git
2018-12-20 09:24:33 <danieli> we haven't removed it, we just quit shipping it as default
2018-12-20 09:24:40 <georgios> buried in some old commit
2018-12-20 09:24:46 <georgios> 
ACTION is not good at git

2018-12-20 09:25:22 <_ikke_> danieli: well, it's not in the current tree anymore, you have to retreive it from an older commit
2018-12-20 09:25:28 <danieli> fair enough
2018-12-20 09:26:21 <_ikke_> georgios: https://git.alpinelinux.org/cgit/aports/tree/main/linux-hardened?id=8ed4b3d76ab839286b82474e75a6f61dc3a6bfbb
2018-12-20 09:27:33 <georgios> thank you _ikke_
2018-12-20 09:27:38 <_ikke_> https://dev.alpinelinux.org/~ncopa/grsec/hardened-3.1-4.9.73-201704252333-alpine2.patch
2018-12-20 09:28:22 <georgios> so remind me please, what is the difference of the regular grsec and the hardened kernel?
2018-12-20 09:28:49 <georgios> arch still offers a hardened kernel but i cant see the difference
2018-12-20 09:28:58 <danieli> linux-hardened is an unofficial port
2018-12-20 09:29:32 <georgios> and was abandoned due to overcomplication?
2018-12-20 09:29:58 <danieli> 09:23:28 <@ncopa> georgios: we dropped grsec because upstream devs removed it from public
2018-12-20 09:29:59 <danieli> 09:24:00 <@ncopa> we ran an unofficial port of it for a while, be we always knew we relied on upstream
2018-12-20 09:29:59 <danieli> 09:24:33 <@ncopa> upstream used to release "testing" patches for new kernels, which was the base of our unofficial fork
2018-12-20 09:31:02 <Hamilton> Do I need to also install guest additions by mounting the VBox iso?
2018-12-20 09:31:20 <Hamilton> I can't `mount -o loop ~/VBoxGuestAdditions_5.2.0_RC1.iso /media/GuestAdditionsISO` in the first place. invalid argument
2018-12-20 09:31:58 <danieli> Hamilton: context..?
2018-12-20 09:32:15 <Hamilton> danieli, Trying to run Xorg on Alpine inside Vbox
2018-12-20 09:32:34 <danieli> there are virtualbox guest additions in the repos
2018-12-20 09:32:34 <Hamilton> Installed virtualbox guest addition packages from apk
2018-12-20 09:33:10 <danieli> i don't believe anything in the vbox-provided ISO is built against musl
2018-12-20 09:33:43 <danieli> i also had to fiddle a bit with that, but i got it to work eventually - don't remember what i did, but i can check real quick
2018-12-20 09:34:19 <Hamilton> danieli, So no need to manualy mount their ISO and run their bash script?
2018-12-20 09:34:29 <danieli> uh?
2018-12-20 09:34:40 <Hamilton> VBox guest additions .iso
2018-12-20 09:34:56 <georgios> i understand... they are the only ones able to fork their own product :p
2018-12-20 09:36:56 <danieli> Hamilton: I'm still confused
2018-12-20 09:38:05 <Hamilton> danieli, I mean, for making Alpine work inside vbox, is it sufficient to apk add vbox-guest additions from repositories?
2018-12-20 09:38:17 <danieli> Hamilton: should be, also make sure the kernel modules get loade
2018-12-20 09:38:20 <danieli> loaded*
2018-12-20 09:38:47 <Hamilton> danieli, How should I check if kernel modules are loaded?
2018-12-20 09:38:53 <_ikke_> lsmod
2018-12-20 09:38:57 <danieli> as usual, lsmod
2018-12-20 09:39:01 <danieli> there's no magic to it
2018-12-20 09:40:04 <Hamilton> danieli, Sorry I'm n00b. It is ok? http://termbin.com/vven
2018-12-20 09:42:31 <Hamilton> My main problem for these stuff is that I want to install X but startx issues error. If I install xf86-video-vesa, the error would be gone but desktop would be really slow and laggy
2018-12-20 09:46:26 <georgios> good bye for now. thanks to everybody for the valuable information. dalias , ncopa , _ikke_ , danieli , cheers!
2018-12-20 09:47:19 <danieli> 01:50:50 <dalias> grsec still exists? :)
2018-12-20 09:47:19 <danieli> 01:50:53 <dalias> </troll>
2018-12-20 09:47:19 <danieli> 01:51:07 <georgios> hi dalias
2018-12-20 09:47:19 <danieli> 01:51:15 <georgios> musl still exists?
2018-12-20 09:47:22 <danieli> this gave me a good giggle
2018-12-20 14:15:01 <Greyfn> @daneli: thanks for answering my udhcpd question. I'm still struggling with udhcpd though, it appears to not be writing it's pidfile. I'm investigating further.
2018-12-20 15:30:08 <greyeax> hello - does the virtualbox-guest-additions package not support screen resizing?
2018-12-20 15:34:27 <ncopa> i havent tried it
2018-12-20 15:34:35 <danieli> greyeax: you didn't highlight me properly, so I didn't see the message until now - I could have a quick look at it in a bit
2018-12-20 15:57:05 <mtnt> hi there!
2018-12-20 16:00:49 <TuXAPuK> re: mtnt >> Hello!
2018-12-20 16:02:52 <mtnt> Hey, TuXAPuK! Got question: my soundcard is PCI based. It is E-MU 1212M PCI. Alsamixer won't see it. Is it even supported by Alpine& I'm noob, still learning.
2018-12-20 16:04:22 <mtnt> And that [KMS] drm modesetting not supported... Ouch, my head going to blow
2018-12-20 16:05:32 <mtnt> Video is AMD R9 390. Xorg possibly won't work for me?
2018-12-20 16:17:03 <danieli> if linux supports it, alpine *should*, unless it requires super wonky drivers
2018-12-20 16:17:08 <danieli> in which case they might just be in a separate package
2018-12-20 16:22:57 <mtnt> Hey, danieli, thanks for the heads up. But i'm to exhausted to continue research. Way too hard for me. Everything looks okay, bit "something" won't let me go any further. Maybe i got to run the alsamixer and xorg not from root... Got to read more. I like Alpine, it is clean and all. But that blobbed hardware killing my interest. Possibly got to find some old laptop and try to install it there, not on mu desktop.
2018-12-20 16:23:13 <mtnt> but*
2018-12-20 16:23:35 <danieli> as far as I see, alsa *should* support it, with firmware loaded - they used to have pretty poor linux support
2018-12-20 16:23:54 <danieli> looks like the firmware is in alsa-firmware, which I don't believe we have packaged, judging by a quick search
2018-12-20 16:24:30 <mtnt> Yep, no firmware (
2018-12-20 16:26:54 <mtnt> Thanks, got to smoke more manuals )))
2018-12-20 16:29:39 <greyeax> danieli: My bad, that would be dope, thanks
2018-12-20 17:43:38 <hatramatra> Any opinions here on full disk encryption, that is LVM on top of LUKS where even /boot is an LV and the whole thing is booted via GRUB?
2018-12-20 18:13:10 <hiro> hatramatra: useless IMO
2018-12-20 18:13:38 <hiro> hatramatra: if somebody can replace your GRUB you're fucked either way.
2018-12-20 18:14:46 <hiro> hatramatra: my advice i give people is to ignore full-root stuff, and instead make more independent mounts that each have different escalation levels, and different keys.
2018-12-20 18:15:28 <hiro> hatramatra: this way there is a higher chance that somebody who sits in front of the computer at least doesn't have ALL encrypted partitions mounted at the same time, cause you at least have the option of mounting only some of them at a time
2018-12-20 18:16:01 <hiro> and anyway, it's quite common to not need the most secret stuff on a daily basis
2018-12-20 18:16:05 <hatramatra> hiro: I feel you are correct. For my use case (the computer is stolen and I can still sleep well not to worry about my files running wild), /boot can stay unencrypted.
2018-12-20 18:16:31 <hiro> yeah, if you want to obsess more, you can do this fine-grained partitioning
2018-12-20 18:16:50 <hiro> to cover the scenario of person having access to the RUNNING computer
2018-12-20 18:17:22 <hiro> otoh it's probably good to encrypt swap in case you use suspend to disk
2018-12-20 18:17:37 <hiro> i don't even have root encrypted.
2018-12-20 18:17:41 <hatramatra> Encrypted /boot makes sense for someone who goes all the way and uses something like cryptboot
2018-12-20 18:17:49 <hiro> and i don't use suspend really, so that doesn't matter for me.
2018-12-20 18:18:12 <hiro> hatramatra: never heard of that, but i guess it's some hardware stage that secures the bootloader?
2018-12-20 18:18:30 <hatramatra> https://github.com/xmikos/cryptboot
2018-12-20 18:19:08 <hiro> oh secureboot
2018-12-20 18:19:11 <hiro> yeah, uefi stuff
2018-12-20 18:19:50 <hiro> hatramatra: sounds good. if it works...
2018-12-20 18:23:21 <hatramatra> hiro: I've never tried it, though. I really don't have use for it at the moment.
2018-12-20 18:23:28 <hatramatra> BRB
2018-12-20 18:50:19 <maldridge> ncopa: as promised, http://ix.io/1wsi is an strace of cf-serverd starting up exhibiting the error
2018-12-20 18:54:13 <greyeax> does anyone here use alpine as a daily driver?
2018-12-20 18:55:02 <dalias> i assume by that you mean a system you sit down at (desktop/laptop) rather than some sort of server/remote?
2018-12-20 18:55:29 <greyeax> yea
2018-12-20 18:55:41 <greyeax> like your everyday desktop os usage
2018-12-20 18:56:22 <maldridge> I don't see a reason you couldn't use it that way
2018-12-20 18:56:28 <maldridge> but like all things it depends on your workload
2018-12-20 18:57:49 <greyeax> makes sense
2018-12-20 18:58:06 <dalias> then yes. but it depends on what you want to do
2018-12-20 18:58:15 <greyeax> my only real concern would be lack of productivity packages and issues building with musl
2018-12-20 18:58:20 <greyeax> because ive heard thats a thing
2018-12-20 18:58:31 <greyeax> i was just curious thought it might be fun to do
2018-12-20 18:58:49 <dalias> the major productivity packages are there: firefox and chromium, openoffice, inkscape, gimp, ...
2018-12-20 18:59:58 <dalias> more obscure stuff like solvespace needs to be built from source still
2018-12-20 19:00:09 <dalias> i was going to say blender but apparently there's a blender package now!
2018-12-20 19:00:21 <dalias> i should try it out
2018-12-20 19:01:07 <greyeax> nice
2018-12-20 19:03:02 <dalias> if you want games it's definitely more work
2018-12-20 19:03:40 <dalias> most emulators build with minimal patching (typical stuff like missing #include directives)
2018-12-20 19:04:12 <dalias> and all the prereqs generally have -dev packages so you don't have to build those yourself too
2018-12-20 20:22:43 <ncopa> greyeax: i use alpine as my daily driver on my work desktop
2018-12-20 20:35:14 <mps> I use alpine on three machines as main working desktops, my wife mostly for web and similar things, and my daughter for her daily and at the university where she study IT
2018-12-20 20:53:58 <maldridge> ncopa: I'm certainly happy to believe this is an issue in my system, but I've not been able to isolate anything that would imply that so far
2018-12-20 22:12:14 <ncopa> maldridge: i dont think its an issue in your system. i think its a bug in cfengine
2018-12-21 01:57:39 <Greyfn> Right, so I'm absolutely stumped. When I build busybox from source directly it seems to write pidfiles correctly. udhcpd from Busybox-extras from the repos seems not to do so. Still trying to work out why.
2018-12-21 02:14:37 <maldridge> ncopa: do you have any feedback on where you think that issue might be that I can take back to the cfengine folks?
2018-12-21 02:16:19 <danieli> ncopa: it's efficient *when it's stable*, alpine still has a long way to go to be a desirable daily driver for me. I installed it on my primary laptop, but I might have to reinstall some other OS when I get back to work
2018-12-21 02:18:31 <jwh> heh
2018-12-21 05:07:54 <chr0mag> hi folks, I'm trying (and failing) to boot an Alpine ISO using Grub's loopback feature. I've copied the APPEND options from the ISO's syslinux.cfg file and added 'alpine_dev=usbdisk:ext4 (and added ext4 to modules=...)
2018-12-21 05:09:34 <chr0mag> ...but it can't find the loop disk. I know other distros have a option to specify the path to the ISO file (eg. Debian has findiso=$iso_path). Does Alpine have an equivalent?
2018-12-21 06:15:13 <chr0mag> OK, it can find the loop disk now but the boot fails and drops me into an emergency shell, presumably because it can't find the ISO.
2018-12-21 06:15:29 <chr0mag> It doesn't seem to matter what value I set for 'alpine_dev'.
2018-12-21 08:15:04 <ncopa> can someone with rpi test if linux-rpi kernel works in alpien 3.8.2? https://bugs.alpinelinux.org/issues/9732
2018-12-21 08:15:14 <ncopa> i tested the aarch64, and it works
2018-12-21 08:15:34 <ncopa> i need to go out for a while...
2018-12-21 11:07:41 <danieli> maldridge: did you get any wiser with cfengine?
2018-12-21 13:17:08 <hatramatra> fcolista: lxd 3.8 version depends on something called libdqlite.so.0. Should there be a package for that, please?
2018-12-21 13:25:59 <danieli> it's apparently distributed sqlite (?)
2018-12-21 13:26:13 <danieli> https://github.com/CanonicalLtd/dqlite and go-dqlite (go bindings)
2018-12-21 13:32:10 <fcolista> umh
2018-12-21 13:32:21 <fcolista> hatramatra, you are right
2018-12-21 13:33:17 <hatramatra> >>> lxd*: Tracing dependencies...
2018-12-21 13:33:24 <hatramatra> >>> ERROR: lxd*: libdqlite.so.0: path not found
2018-12-21 13:33:35 <hatramatra> That is from the build log
2018-12-21 13:33:41 <danieli> doesn't look like we have dqlite anything in our repos
2018-12-21 13:34:25 <Greyfn> I've figured out why udhcpd in busybox-extras isn't working properly (not writing pidfiles; can't start/stop the process using rc-service). It's a problem with the busybox config used. How do I go about proposing a fix/patch?
2018-12-21 13:34:38 <fcolista> lxcd is shipped with his own version of sqlite /0\
2018-12-21 13:34:44 <fcolista> WHY!?!!?
2018-12-21 13:34:50 <fcolista> *lxd
2018-12-21 13:46:43 <hatramatra> fcolista: I checked how it's done in ubuntu. In their bionic release,  /usr/lib/lxd/libdqlite.so.0 is part of lxd package directly.
2018-12-21 13:47:31 <fcolista> seems to be patched version of lxc which allows to work in cluster mode
2018-12-21 13:47:46 <hatramatra> That is 3.0.3 version. In their Cosmic release, they install snap package instead and not .deb anymore.
2018-12-21 13:48:21 <hatramatra> sorry, 3.0.2
2018-12-21 13:49:00 <hatramatra> lxc command in itself works. Just lxd doesn't run due to this missing dependency.
2018-12-21 13:49:09 <fcolista> yes, gotcha
2018-12-21 13:52:31 <danieli> most likely the same thing as the last link I posted
2018-12-21 13:52:48 <danieli> C lib and its Go bindings
2018-12-21 13:53:23 <fcolista> danieli, I think that would be better to package dqlite and add it as dependency of lxd
2018-12-21 13:53:24 <hatramatra> There is some information here https://discuss.linuxcontainers.org/t/lxd-natively-on-debian/1318/3
2018-12-21 13:53:26 <fcolista> what do you think?
2018-12-21 13:53:45 <danieli> fcolista: +1
2018-12-21 13:53:50 <danieli> I'll give it a try, one sec
2018-12-21 13:54:18 <danieli> oh BLEH
2018-12-21 13:54:38 <danieli> To build libdqlite from source you'll need:
2018-12-21 13:54:38 <danieli> - A reasonably recent version of libuv (v1.8.0 or beyond)
2018-12-21 13:54:38 <danieli> - A patched version of SQLite with support for WAL-based replication.
2018-12-21 13:54:56 <fcolista> yes
2018-12-21 13:55:13 <fcolista> i need to package sqlite patched  and dqlite
2018-12-21 13:55:17 <fcolista> as different packages
2018-12-21 13:55:23 <fcolista> then add to lxd
2018-12-21 13:55:27 <hatramatra> :-|
2018-12-21 13:55:32 <fcolista> in sqlite i need to call it differently
2018-12-21 13:55:37 <danieli> god, what a pain
2018-12-21 13:55:44 <fcolista> and set "conflicts" with sqlite
2018-12-21 13:55:53 <fcolista> a royal pain
2018-12-21 13:56:05 <fcolista> why are they not using standard tool?
2018-12-21 13:56:11 <fcolista> /0\
2018-12-21 13:56:19 <danieli> #JustGolangThings
2018-12-21 13:56:21 <hatramatra> Did I have to bring this before Christmas? /me feel bad.
2018-12-21 13:56:29 <fcolista> +1 danieli
2018-12-21 13:56:48 <danieli> hatramatra: this fades in comparison with some of the other difficult TODO tasks
2018-12-21 13:56:54 <danieli> don't worry
2018-12-21 13:57:19 <danieli> having to jump through three hoops to get what we want is a little annoying though
2018-12-21 13:58:35 <danieli> fcolista: are you currently looking at it?
2018-12-21 13:58:40 <fcolista> yes
2018-12-21 13:58:45 <danieli> good to know
2018-12-21 14:05:30 <fcolista> what arch does is adding the patch to sqlite to enable replica
2018-12-21 14:05:43 <fcolista> while CanonicalLtd ships sqlite already patched
2018-12-21 14:05:52 <fcolista> I woudld go with the second option
2018-12-21 14:05:59 <fcolista> calling it sqlite-lxd
2018-12-21 14:06:08 <fcolista> and set conflicts="sqlite"
2018-12-21 14:07:21 <danieli> that sounds good to me
2018-12-21 14:07:27 <danieli> or just call it dqlite
2018-12-21 14:07:45 <danieli> oh nevermind, a patched sqlite is required to build dqlite at all, my bad
2018-12-21 14:07:51 <danieli> +1 sqlite-lxd
2018-12-21 14:09:51 <hatramatra> does it mean that patched sqlite will be drop in replacement for unpatched sqlite for other packages?
2018-12-21 14:10:54 <danieli> 'for other packages'?
2018-12-21 14:10:58 <danieli> the patched sqlite is intended only for dqlite
2018-12-21 14:14:15 <hatramatra> If it has conflicts="sqlite", how one can have reqular sqlite packages installed?
2018-12-21 14:18:32 <danieli> sqlite-lxd is intended only for building dqlite
2018-12-21 14:19:19 <danieli> fcolista: we could statically link it, unless it's huge?
2018-12-21 14:19:35 <fcolista> it's what arch does for lxc
2018-12-21 14:19:37 <fcolista> *lxd
2018-12-21 14:28:07 <fcolista> wondering how to set the pkgver in order to be valid//
2018-12-21 14:28:58 <danieli> fcolista: what exactly is the issue with it?
2018-12-21 14:29:18 <fcolista> pkgver should be formatted in a certain way to be valid
2018-12-21 14:29:19 <fcolista> 3.26.0_repl3 is not a valid version
2018-12-21 14:29:31 <fcolista> I was thinking how to set it to differenciate with sqlite pkgver
2018-12-21 14:29:45 <fcolista> 3.26.0_replication3 is not a valid version
2018-12-21 14:29:58 <fcolista> i can set r2
2018-12-21 14:30:01 <fcolista> *r3
2018-12-21 14:30:10 <fcolista> but it can be confusing with pkgrel
2018-12-21 14:32:00 <mtnt> Hiya! Just came back to say thanks. Alpine is just the best distro on the scene now. Keep up! Happy holydays!
2018-12-21 14:34:32 <danieli> 12 seconds, didn't even get to say happy holidays :(
2018-12-21 14:34:38 <danieli> s/didn't/I didn't/
2018-12-21 15:07:27 <fcolista> it's a can of worm this lxd with sqlite patched
2018-12-21 15:10:44 <danieli> fcolista: sounds like a real pandora's box
2018-12-21 15:10:46 <danieli> what's up with it?
2018-12-21 15:14:05 <fcolista> the sqlite patched version (replication3) does not find the sqlite3.h header
2018-12-21 15:14:23 <fcolista> the header is shipped with sqlite-dev
2018-12-21 15:14:38 <fcolista> but sqlite conflicts with it /0\
2018-12-21 15:15:05 <danieli> wait, the patched sqlite depends on sqlite to build?
2018-12-21 15:15:07 <danieli> that sounds very strange
2018-12-21 15:15:14 <danieli> or am I being an idiot?
2018-12-21 15:16:45 <fcolista> https://github.com/CanonicalLtd/sqlite
2018-12-21 15:23:33 <danieli> fcolista: make: *** No rule to make target '/some/path/sqlite/manifest', needed by 'sqlite3.h'.  Stop.
2018-12-21 15:23:44 <fcolista> correct
2018-12-21 15:25:12 <danieli> manifest.uuid?
2018-12-21 15:25:31 <danieli> bah, i have no clue, i'll dig into it a little
2018-12-21 15:25:54 <fcolista> danieli, I think I'll do something different
2018-12-21 15:29:55 <danieli> aha
2018-12-21 15:30:04 <danieli> fcolista: https://github.com/CanonicalLtd/dqlite#build
2018-12-21 15:30:07 <danieli> it has pretty specific instructions
2018-12-21 15:30:33 <fcolista> danieli, yes
2018-12-21 15:30:50 <fcolista> I don't have issue with dqlite
2018-12-21 15:30:58 <fcolista> but with the sqlite patched version
2018-12-21 15:31:09 <danieli> look closer
2018-12-21 15:31:10 <fcolista> https://github.com/CanonicalLtd/sqlite/releases/tag/version-3.26.0+replication3
2018-12-21 15:31:24 <danieli> aah right, you're using the release, hmm
2018-12-21 15:31:42 <danieli> it doesn't include the manifest or manifest.uuid files the dqlite readme uses git for creating
2018-12-21 15:31:58 <danieli> following the instructions under build does work, but seems to require git
2018-12-21 15:32:10 <fcolista> I've not tried with git
2018-12-21 15:32:15 <danieli> try it, it just worked for me
2018-12-21 15:32:32 <fcolista> https://dpaste.de/j4Dv/raw
2018-12-21 15:32:36 <fcolista> I'm working on this
2018-12-21 15:33:07 <danieli> i see, you're doing it all in lxd
2018-12-21 15:33:07 <fcolista> I've almost done
2018-12-21 15:33:12 <fcolista> >>> lxd*: Stripping binaries
2018-12-21 15:33:12 <fcolista> strip: ./usr/lib/lxd/stOPNLbB: not enough room for program headers, try linking with -N
2018-12-21 15:33:12 <fcolista> strip:./usr/lib/lxd/stOPNLbB[.gnu.hash]: bad value
2018-12-21 15:33:12 <fcolista> >>> ERROR: lxd*: prepare_package failed
2018-12-21 15:33:32 <danieli> that... sounds painful
2018-12-21 15:33:40 <fcolista> last package steps fail with this error
2018-12-21 15:34:33 <danieli> sounds like the sqlite-lxd idea would be easier, given that we can create manifest{,.uuid} files without using git
2018-12-21 15:35:45 <fcolista> yes, that's the other option
2018-12-21 15:36:05 <fcolista> this would mean maintain 2 other packages
2018-12-21 15:36:15 <fcolista> dqlite + sqlite patched version
2018-12-21 15:36:24 <danieli> that is unfortunately true
2018-12-21 15:36:29 <fcolista> with the headache of upgrading them with a criteria
2018-12-21 15:36:46 <fcolista> from maintenance perspective this is more painful
2018-12-21 16:43:18 <CcxWrk> When installing x86 alpine edge chroot using x86_64 host I get post-install script errors (127) and chroot does not work (No such file or directory).
2018-12-21 16:43:39 <CcxWrk> I presume broken shared library somewhere, but musl package installed successfully.
2018-12-21 16:44:33 <CcxWrk> The /lib/ld-musl-i386.so.1 is present and ELF formatted at least.
2018-12-21 16:48:58 <CcxWrk> latest-stable seems to have the same issue
2018-12-21 16:49:41 <TBB> this is basically a guessing game
2018-12-21 16:51:06 <TBB> 32/64 bit incompatibility (due to host not being multilib or such?), flawed method of installation ...
2018-12-21 16:53:38 <CcxWrk> I'm using apk.static, just updated it just to be sure. This used to work, didn't change anything in my scripts.
2018-12-21 16:54:45 <CcxWrk> Multilibness of host should not matter for chroot, just if kernel can execute x86 binaries.
2018-12-21 16:55:34 <TBB> 127 usually refers to a missing binary somewhere, so you could be right about a missing library
2018-12-21 16:56:35 <TBB> wonder how quickly I could alter my own install method to do 32-bit chroot, I usually only do 64 bit on 64 bit
2018-12-21 16:57:48 <CcxWrk> I need this for i586 router. I do have the signing keys installed so the packages should be verified.
2018-12-21 17:00:37 <TBB> just a moment, I'll see if I can reproduce this
2018-12-21 17:02:51 <CcxWrk> https://paste.pound-python.org/show/kTmextioQcbALXmjPKrE/
2018-12-21 17:03:49 <danieli> hmm, busybox trigger exited with 127
2018-12-21 17:04:00 <danieli> several scripts did, actually
2018-12-21 17:04:26 <CcxWrk> Presumably /bin/sh doesn't work in there.
2018-12-21 17:04:28 <danieli> your mirror URL is invalid by the way
2018-12-21 17:04:38 <danieli> latest-stable is not edge, only edge has the testing repo
2018-12-21 17:05:23 <CcxWrk> Yeah, I tried edge and then changed it to latest-stable. Should not matter really, can run it against edge if you want.
2018-12-21 17:06:04 <danieli> it *shouldn't* matter, it looks like the system failed to bootstrap due to the apk errors
2018-12-21 17:07:02 <CcxWrk> Same thing if I remove the invalid repo. Or use edge.
2018-12-21 17:08:00 <danieli> would almost think it's incapable of running 32-bit code for whatever reason, the only thing the busybox post-install script runs is busybox --install -s
2018-12-21 17:11:39 <TBB> can't reproduce using a 64 bit Alpine host
2018-12-21 17:12:24 <CcxWrk> Yeah, seems that x86 binary support is somehow broken on this kernel. Just tried with 32bit apk-static.
2018-12-21 17:13:23 <CcxWrk> Does anyone know the .config flag name for it out of top of the head?
2018-12-21 17:15:23 <CcxWrk> CONFIG_IA32_EMULATION=y  this is weird
2018-12-21 17:15:47 <danieli> i don't remember which config flag is for 32 bit support on 64 bit kernel off the top of my head
2018-12-21 17:16:03 <danieli> might be COMPAT_32 or IA32_EMULATION or something along those lines
2018-12-21 17:16:24 <danieli> i have both fo those plus COMPAT set on mine
2018-12-21 17:16:26 <danieli> s/fo/of
2018-12-21 17:18:12 <CcxWrk> Yeah, I do have all of those on but I'm getting "Invalid ELF image for this architecture" for "./apk/sbin/apk.static: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, stripped"
2018-12-21 17:18:30 <CcxWrk> But I guess I can just go to another box and make the chroot there. Thanks.
2018-12-21 17:19:20 <danieli> what's the qemu version?
2018-12-21 17:19:24 <hatramatra> fcolista: Is there anything I could do to help with that latest lxd version? What is the conclusion about the way to fix it?
2018-12-21 17:19:25 <danieli> I remember it having a bug with stripped ELFs
2018-12-21 17:19:44 <danieli> @ CcxWrk
2018-12-21 17:20:36 <CcxWrk> qemu?
2018-12-21 17:20:40 <danieli> yes
2018-12-21 17:21:31 <danieli> if it's remotely up-to-date, that's not it
2018-12-21 17:22:01 <danieli> oh right, my bad, this is probably chroot on a real box
2018-12-21 17:22:16 <danieli> nevermind my ramblings :)
2018-12-21 17:22:53 <CcxWrk> Actually you were right. Apparently binfmt_misc was breaking this.
2018-12-21 17:23:11 <CcxWrk> It's a real box, but old version of qemu was lying around and interfering.
2018-12-21 17:23:36 <danieli> hah, and I feel senile with my ramblings and selective memory :)
2018-12-21 17:26:20 <CcxWrk> Is there a list of available kernels somewhere?
2018-12-21 17:26:40 <danieli> elaborate?
2018-12-21 17:27:48 <danieli> thinking of flavors or versions?
2018-12-21 17:28:32 <CcxWrk> Flavors. There used to be a bunch, but I presume both grsec and vserver now to be long gone.
2018-12-21 17:29:02 <danieli> nah, linux-virt is still around
2018-12-21 17:29:13 <danieli> linux-hardened (unofficial grsec port) is obsolete, yes
2018-12-21 17:33:24 <CcxWrk> -virt doesn't seem to use the VServer patchset, unless I'm reading the APKBUILD wrong
2018-12-21 17:33:33 <CcxWrk> Not that I need it, was mostly just curious.
2018-12-21 17:36:16 <CcxWrk> So the default to use is linux-vanilla?
2018-12-21 18:09:42 <maldridge> danieli: marginally
2018-12-21 18:09:54 <maldridge> alpine's cfengine is broken in some way that I have yet to nail down
2018-12-21 18:16:13 <maldridge> due to timing pressures I'm bootstrapping with ansible, and then I'll likely spend the time to do it right later
2018-12-21 18:46:02 <maldridge> right now I'm dealing with openrc seeming to have a mind of its own
2018-12-21 18:53:59 <maldridge> ah, SIGPIPE, one of the few inits I've seen that actually does close the pipes
2018-12-21 20:42:54 <Latrina> 4da9efe9-4904-cacd-d4ac-c3a34b119472
2018-12-21 20:42:59 <Latrina> wrong paste
2018-12-21 21:02:56 <kgmoney> i'm new to alpine and am trying to setup a diskless system as a docker host.  install appears to work fine and then i 'lbu commit', however when i reboot, docker is gone.  anybody know what is happening?
2018-12-21 22:00:46 <ddevault> I have a package which won't upgrade
2018-12-21 22:00:51 <ddevault> it's not pinned according to /etc/apk/world
2018-12-21 22:00:58 <ddevault> but the version shown in apk search and in apk info disagree
2018-12-21 22:02:32 <ddevault> apk fix sorted me out
2018-12-21 22:02:34 <ddevault> ¯\_(ツ)_/¯
2018-12-21 22:13:34 <kgmoney> so...I've figured out my issue as well...docker disappearing after reboot
2018-12-21 22:14:22 <kgmoney> i created my install from the extended iso and didn't create a package cache
2018-12-21 22:14:52 <kgmoney> so some things would persist fine (because i had the packages local already
2018-12-21 22:15:21 <kgmoney> but things like docker weren't local, so they couldn't get reloaded on reboot
2018-12-22 01:38:30 <CcxWrk> Can I set multiple static IP addresses in /etc/networks/interfaces? Or is there some better network configuration system available?
2018-12-22 01:38:47 <CcxWrk> I'd prefer netifrc but that doesn't seem available.
2018-12-22 02:20:09 <maldridge> CcxWrk: yes, ifupdown understand how to do that, just use the correct format
2018-12-22 02:40:18 <Arahael> CcxWrk: That's a multihomed interface, btw.
2018-12-22 03:06:48 <chr0mag> Can anyone point me to documentation for Alpine specific boot options? I'm trying to boot the Alpine ISO using Grub2 loopback and am looking for a way to pass the ISO path as a parameter.
2018-12-22 11:58:34 <mtnt> hi! Just mentioned that acpid (on fresh sys install) won't stop on reboot and, also marked as crashed at rc-status. What may be the cause?
2018-12-22 14:17:27 <alpha_Aquilae> 13 18 quel merdier
2018-12-22 14:18:03 <danieli> uh?
2018-12-22 16:22:58 <Hund_> How come rTorrent was removed from the repos?
2018-12-22 16:27:35 <danieli> Hund: it's in community?
2018-12-22 16:27:39 <danieli> uh
2018-12-22 16:27:43 <Hund> Oh
2018-12-22 16:27:48 <danieli> oh, nick change
2018-12-22 16:28:00 <Hund> It's moved. I didn't know that.
2018-12-22 16:31:44 <danieli> it belongs in community and not in main
2018-12-22 16:31:57 <danieli> anyway, looks like it was moved june 15th
2018-12-22 16:32:09 <danieli> for future reference: https://pkgs.alpinelinux.org/packages?name=rtorrent&branch=edge&arch=x86_64
2018-12-22 16:32:27 <Hund> Thanks!
2018-12-22 19:49:28 <Dreamer3> isntaller really not meant to work in 64mb, huh?
2018-12-22 19:59:29 <desnudopenguino> so, just learned about alpine linux, went to the downloads page, and there are a ton of different versions. my use case is this: i run openbsd, and alpine seems to be a good one to run virtualized on that host. i'm looking to use it as a jump-off point for getting into containers. would the virtual or mini root filesystem be a good choice in that case (or something else?)
2018-12-22 20:00:09 <_ikke_> For containers, I think minirootfs should be enough
2018-12-22 20:04:03 <desnudopenguino> cool. i have a lot to learn, and i'm excited
2018-12-22 20:07:40 <maldridge> desnudopenguino: iirc openbsd doesn't have native containers, so you're looking at running alpine in vmm and then containers in alpine?
2018-12-22 20:07:51 <desnudopenguino> maldridge: correct
2018-12-22 20:07:53 <maldridge> at least my 6.1 box certainly doesn't have native containers
2018-12-22 20:08:08 <desnudopenguino> yeah, there won't be native containers anytime soon (if ever)
2018-12-22 20:08:23 <maldridge> you probably want the virt flavor then, but honestly vmm doesn't expose most of paravirt, so even vanilla should be fine
2018-12-22 20:09:04 <desnudopenguino> ok.
2018-12-22 20:11:22 <maldridge> desnudopenguino: how stable is vmm now?  I was driven off from using it due to numerous problems with UFS stability on unclean shutdown
2018-12-22 20:12:24 <desnudopenguino> i'm not really sure, but we are on 6.4 now, and it seems to be getting better. i haven't followed it too closely because i've been running virtualization on a separate host
2018-12-22 20:14:44 <desnudopenguino> the talk is that alpine is a good guest OS on vmm/vmd, and having the ability to play with containers makes me want to mess with them for some of my web dev work. that's currently where i am with it all. lol
2018-12-22 20:15:16 <_ikke_> alpine is a good host OS as well :-)
2018-12-22 20:15:41 <maldridge> I find alpine to be very nice, I am tempted to use it as the base OS for my personal VM cluster since I could switch to xen
2018-12-22 20:15:50 <desnudopenguino> cool
2018-12-22 20:16:01 <maldridge> though I am quite satisfied with my void+kvm base
2018-12-22 20:17:04 <desnudopenguino> i have a host running proxmox
2018-12-22 20:19:15 <maldridge> ah, the million line perl wonder of the modern world
2018-12-22 20:19:20 <desnudopenguino> lol
2018-12-22 20:19:30 <desnudopenguino> it just works(tm)
2018-12-22 20:19:31 <desnudopenguino> lol
2018-12-22 20:20:03 <maldridge> I spent a few months trying to build packages for it that could be installed outside of its custom debian base, and eventually gave up after discovering a tripply nested build loop
2018-12-22 20:20:27 <maldridge> its tempting to build a vm control frontend for alpine and just be done with it
2018-12-22 20:20:36 <maldridge> like a replacement to how oVirt works
2018-12-22 20:22:05 <desnudopenguino> hmm
2018-12-22 20:24:11 <Dreamer3> it's too bad IP space is still limited
2018-12-22 20:29:49 <Dreamer3> was just looking at linode long long time ago when they had 64mb vpses
2018-12-22 20:30:02 <Dreamer3> now you could fit 16 inside a single $5 VPS
2018-12-22 20:30:23 <Dreamer3> but without IPs hard to do treat them as mini VPSes
2018-12-22 20:35:40 <TomJepp> unique IPv6 addresses and ipv4 proxying of services?
2018-12-22 20:52:56 <Dreamer3> but if everyone wants a website on port 80... :)
2018-12-22 20:53:16 <Dreamer3> but yeah if ipv6 was enough it'd be much easier me thinks
2018-12-22 20:53:50 <desnudopenguino> on ipv4, throw a proxy in front like nginx or haproxy?
2018-12-22 20:56:55 <Dreamer3> but now they aren't really independent boxes then
2018-12-22 21:49:56 <Dreamer3> what is armhf vs arm64?
2018-12-22 21:50:13 <maldridge> 32bit vs 64bit arm
2018-12-22 21:50:24 <danieli> practically, yes
2018-12-22 21:50:57 <danieli> Dreamer3: you'll get more through and exhaustive answers to that question if you google it
2018-12-22 22:00:34 <Dreamer3> https://stackoverflow.com/questions/37790029/what-is-difference-between-arm64-and-armhf
2018-12-22 22:01:07 <maldridge> you can also wind up down the wormhole to answer that question
2018-12-22 22:01:16 <maldridge> its worth learning imho, but know that this is a deep pit to go into
2018-12-22 22:05:00 <Dreamer3> llol
2018-12-23 15:28:00 <georgios> hello
2018-12-23 15:28:07 <georgios> https://github.com/jameshilliard/linux-grsec
2018-12-23 15:28:09 <georgios> :)
2018-12-23 15:32:16 <mps> georgios: it is for 4.4.x
2018-12-23 15:33:09 <mps> maybe I have somewhere on archive old patches
2018-12-23 16:13:12 <georgios> 
ACTION wishes he was skilled enough to be able to develop it :(

2018-12-23 16:23:40 <georgios> i just found a "b2sum" utility. at last blake2 csums!
2018-12-23 16:23:51 <_ikke_> nice
2018-12-23 16:31:05 <georgios> /usr/bin/b2sum is owned by coreutils 8.30-1
2018-12-23 16:31:06 <georgios> :D
2018-12-23 16:32:55 <nmeum> clandmeter: just a heads up: community/ocrmypdf is currently broken since our py-ruffus version is too new. A fix would either be downgrading ruffus or upgrading ocrmypdf. The latter will require new dependencies which aren't packaged currently
2018-12-23 16:33:05 <nmeum> don't have enough time on my hands to fix it myself atm
2018-12-23 18:37:59 <_ikke_> Hmm, somehow, interface lo is not enabled for me on boot on a specific machine
2018-12-23 19:39:05 <Dreamer3> "New users are not allowed to add ip addresses and phone numbers within 5 hrs of account creation""
2018-12-23 19:39:06 <Dreamer3> what?
2018-12-23 19:39:18 <Dreamer3> i'm not trying to add an address ot phone number just fix an error
2018-12-23 19:56:15 <_ikke_> Dreamer3: heh, we had to deal with a lot of spam
2018-12-23 19:56:40 <Dreamer3> so is the error just weirdly worded or what?
2018-12-23 19:56:45 <Dreamer3> does it mean i can't fix anything for 5 hours?
2018-12-23 19:56:59 <_ikke_> Let me check if I can find what it might be detecting
2018-12-23 19:57:30 <clandmeter> It's it an edit?
2018-12-23 19:58:09 <clandmeter> Could be already included url
2018-12-23 19:58:51 <_ikke_> It probably does not look at the diff
2018-12-23 20:10:41 <Dreamer3> yes i'm trying to fix  typo
2018-12-23 20:41:38 <_ikke_> Anyone got an idea why my lo interface does not come up at boot?
2018-12-23 20:41:46 <_ikke_> Some services wouldn's tart because of it
2018-12-23 20:45:10 <mps> _ikke_: hmmm, stupid question, but did you checked syntax of /etc/network/interfaces
2018-12-23 20:47:56 <_ikke_> yes
2018-12-23 20:48:48 <_ikke_> printf '#!/bin/sh\n\nsed
2018-12-23 20:48:50 <_ikke_> sed "s#:PREFIX:#/usr#g" $1 >$2
2018-12-23 20:48:52 <_ikke_> arg
2018-12-23 20:49:24 <_ikke_> http://tpaste.us/lZyg
2018-12-23 20:49:26 <_ikke_> http://tpaste.us/lZyg
2018-12-23 20:50:12 <mps> there is no lo interface
2018-12-23 20:50:54 <_ikke_> I tried adding it
2018-12-23 20:51:03 <_ikke_> auto lo, but it would not accept that
2018-12-23 20:51:17 <mps> auto lo eth0
2018-12-23 20:51:34 <mps> and 'iface lo inet loopback'
2018-12-23 20:51:48 <_ikke_> Strange that it suddenly stopped working
2018-12-23 20:51:53 <_ikke_> I haven't touched that file
2018-12-23 20:52:13 <mps> what says 'ifup lo'
2018-12-23 20:52:33 <_ikke_> RTNETLINK answers: File exists
2018-12-23 20:52:56 <mps> and 'ifdown lo'
2018-12-23 20:53:02 <_ikke_> or ifup: ignoring unknown interface lo
2018-12-23 20:53:09 <_ikke_> interface lo not configured
2018-12-23 20:53:54 <mps> try to add it to top in /etc/network/interfaces and ifup/ifdown
2018-12-23 20:54:12 <_ikke_> http://tpaste.us/dkPo
2018-12-23 20:54:22 <_ikke_> still complains about it existing, but it does bring it up
2018-12-23 20:55:16 <mps> that's ok, probably remembered by init script
2018-12-23 20:57:06 <mps> iirc, kernel 'installs' interface lo by default but don't bring it up, this is left to OS
2018-12-23 20:57:31 <_ikke_> right
2018-12-23 20:59:51 <mps> with added 'auto lo' and 'iface lo inet loopback' it could be controlled manually by ifup and ifdown
2018-12-23 21:00:35 <_ikke_> yes, makes sense, I just find it strange that this was not necessary before
2018-12-23 21:01:44 <mps> if you find what removed it from /etc/network/interfaces report it, maybe some package did that
2018-12-24 00:38:31 <midasi> Hello, I'm looking for a custom 4.14 kernel with grsec enabled. Anybody still using grsec?
2018-12-24 00:58:34 <pardis> Yes, you can easily use grsec by paying the grsec developers money
2018-12-24 00:59:14 <midasi> yes I know...I'm fine with paying them... but I'm not sure if it's straightforward to include the grsec patch in the alpine linux kernel
2018-12-24 01:00:58 <midasi> I suspect issues with musl c
2018-12-24 01:22:24 <jn__> midasi: Alpine used to have a kernel with grsec, so i *guess* it'll work fine
2018-12-24 01:38:30 <Dreamer3> yeah wasn't that long ago grsec was default
2018-12-24 06:44:54 <danieli> yes, but that wasn't grsec though
2018-12-24 06:44:59 <danieli> it was linux-hardened
2018-12-24 06:45:08 <danieli> there is a difference @ jn__ Dreamer3
2018-12-24 08:20:05 <_ikke_> before it was hardened, it was called grsec
2018-12-24 10:03:42 <spacelander> Hi! I'm wirting a shell script (default busybox-ash) and I need to use sort -d. However, my list is in German and German diacritics (e.g. ü) are simply ignored. I'v read I can't set LC_AL like in glibc systems. How can I make sort put e.g. Müa* after Mu*, not before Me*?
2018-12-24 10:06:42 <AinNero> i'd be surprised if busybox sort has support for unicode sorting
2018-12-24 10:07:19 <danieli> you're probably better off using coreutils sort for that
2018-12-24 10:11:00 <spacelander> OK, so it's a busybox thing not a musl one. Thanks!
2018-12-24 10:11:50 <danieli> spacelander: we just said it's unlikely busybox has support, i haven't actually looked into your issue
2018-12-24 11:10:02 <Arahael> spacelander: If you just need "sorting", and don't need it to be lexically sorted, you could perhaps pretend it's latin1?
2018-12-24 11:29:12 <spacelander> <Arahael> lexical sorting would have been nice. I'm not sure what you mean by pretend it's latin1, however.
2018-12-24 15:07:01 <Han> I've set up unbound on this host and notice about a 2 second delay for each request to localhost, but if I make the same request to the external interface it goes fast.
2018-12-24 15:07:53 <Han> I vaguely remember some old bug about slow localhost traffic with linux...
2018-12-24 15:08:10 <Han> Does it ring a bell with anyone here?
2018-12-24 15:35:18 <brebs> Perhaps it is trying ipv6 first, and timing out? Try disabling ipv6, in the unbound config.
2018-12-24 15:39:16 <Han> brebs, first check, then change. ;-)  host foo localhost is slow and host foo 127.0.0.1 is fast. We have a winner.
2018-12-24 15:41:23 <Han> Added ::0 to the listening interfaces. Now it works fine. Thanks brebs.
2018-12-24 15:41:47 <brebs> The file /etc/hosts should contain the localhost to 127.0.0.1 lookup
2018-12-24 15:45:00 <Han> true!
2018-12-24 23:19:56 <Arahael> SpaceToast: latin1 is just an 8-bit encoding.
2018-12-24 23:20:07 <SpaceToast> ???????
2018-12-24 23:20:37 <Arahael> SpaceToast: Erm, you ain't spacelander - nevermind!
2018-12-24 23:20:43 <SpaceToast> ah, I see, carry on then
2018-12-25 01:02:15 <JankyGaming> Just set up alpine on a rasp pi. followed instructions, set a password for root, commited changes, rebooted. When I try to ssh in as root with the confirmed working password, it says its the wrong password. Am I missing something? Never had any issues with other distros so im a bit confused
2018-12-25 03:50:01 <agris> hello
2018-12-25 03:50:31 <agris> I've been looking all over, but am unable to find out how Alpine releases work
2018-12-25 03:50:48 <agris> what is edge, testing, vx.x, community, etc
2018-12-25 04:49:03 <qman__> edge is the rolling release, always current; edge-testing is packages that are in the testing phase before being put into edge, and the versions such as 3.8 are separate releases at various points in time, with the intent of being more stable than edge
2018-12-25 04:49:29 <qman__> community is not a release, it's a repo with community-supported packages as opposed to main, which is the main, official set of packages
2018-12-25 04:50:48 <qman__> similar to 'contrib' in debian or 'universe' in ubuntu
2018-12-25 10:27:13 <danieli> qman__: great explanation :)
2018-12-25 11:43:49 <hatramatra> JankyGaming: PermitRootLogin is by default set to 'without-password', that can be confirmed with "sshd -T". Thus password authentication is disabled for root.
2018-12-25 13:16:16 <silverwillow> hey all - been looking around for a recent version (>5.14) of Mono. From what I understand there was an issue with musl and the Mono threadstack? in versions >5.10.. Does anyone know if this has been fixed?
2018-12-25 14:04:12 <danieli> silverwillow: doesn't look like it's been changed much for a good while
2018-12-25 14:05:30 <silverwillow> ok - that's what I thought. I'll be installing the laatest alpine over the next few hours. moving back to alpine from Devuan ASCII
2018-12-25 14:06:17 <silverwillow> then I might bang my head against Mono for a bit, see if I can get it to compile.
2018-12-25 14:08:50 <silverwillow> there's this: https://github.com/mono/mono/issues/7167 and the guy at the bottom: 'frebib' seems to have something that works for 5.12...
2018-12-25 14:09:10 <silverwillow> wish me luck.
2018-12-25 14:10:17 <danieli> looks like upstream has some work to do until we can make a clean and up-to-date alpine package of mono
2018-12-25 14:10:23 <danieli> without patching the living shit out of it
2018-12-25 14:14:57 <silverwillow> nods. but 5.18 was just released and there is mention of a musl related fix... so i;m hoping/assuming the rest of compilation on musl has been fixed too.
2018-12-25 14:15:16 <silverwillow> i know. naive in the extreme...
2018-12-25 16:07:55 <hatramatra> I'm seeking a wisdom on how to solve bug #9539. The nutshell, one has to run 'btrfs device scan' command before btrfs filesystem spanning multiple disks is mountable.
2018-12-25 16:09:13 <hatramatra> As a learning exercise I created apk package which if installed does exactly that. But I don't know if that is the right approach.
2018-12-25 16:10:11 <hatramatra> Maybe it's so trivial, that it shouldn't exists as a separate package. Maybe it should be a subpackage of btrfs-progs. Or something completely different.
2018-12-25 16:10:45 <hatramatra> Shall I try to make pull request for it?
2018-12-25 16:18:42 <_ikke_> hatramatra: a package just for running a command?
2018-12-25 16:19:29 <hatramatra> _ikke_: Well, one has to start somewhere. Even if it's pointless. :-)
2018-12-25 16:20:50 <hatramatra> Of course, the package puts the script running that command into /etc/init.d/ and adds it boot runlevel to be executed before localmount.
2018-12-25 16:21:40 <hatramatra> So although it works and "solves" that bug. Perhaps it's not the right way of solving it.
2018-12-25 16:22:54 <mps> hatramatra: put init script and set localmount to depend on that script, maybe
2018-12-25 16:25:43 <hatramatra> mps: yes, that what it does in essence, see https://dpaste.de/CFAv
2018-12-25 16:26:20 <hatramatra> Also drops file into /etc/conf.d and /etc/module-load.d to be sure btrfs module is loaded.
2018-12-25 16:30:10 <hatramatra> Maybe this should be a subpackage of btrfs-progs (btrfs-progs-openrc?).
2018-12-25 16:30:15 <mps> then the script could be added to btrfs-tools, if it become useful
2018-12-25 16:30:34 <danieli> it's more of a workaround than a fix, I'd say
2018-12-25 16:30:45 <mps> ok, btrfs-progs
2018-12-25 16:32:31 <mps> danieli: everything with btrfs is workaround :)
2018-12-25 16:32:54 <hatramatra> there is no other way to get multi disk btrfs filesystems online, the question is if everybody should hack their way or if there can be distribution standardized way
2018-12-25 16:33:40 <mps> hatramatra: did you looked at some other distro how that problem is solved
2018-12-25 16:33:57 <silverwillow> i had the same issue just now... i simply stuck trivial btrfs.{start|stop) scripts in /etc/local.d and added the 'local' service to the default runlevel
2018-12-25 16:34:21 <silverwillow> works a charm. But then my btrfs is data only, not a root filesystem
2018-12-25 16:34:54 <mps> silverwillow: local init script starts to late for elegant solution
2018-12-25 16:35:28 <silverwillow> I accept that, but as this is on a server that reboots once in forwever, it work fine for my use case.
2018-12-25 16:36:17 <mps> ofc, because that i say 'not elegant' but did't say it is bad
2018-12-25 16:36:22 <silverwillow> :)
2018-12-25 16:37:27 <mps> and, i have some of the 'not elegant' solutions, too
2018-12-25 16:37:43 <silverwillow> the harder part was remembering/finding that you need to run 'btrfs device scan' to get rid of the mount error you get.
2018-12-25 16:38:04 <mps> in the end, it is important for the system to work
2018-12-25 16:38:44 <silverwillow> yeah - which is why i was eager to get a recent mono to work, else I can't run opensimulator. sniff :(
2018-12-25 16:41:07 <silverwillow> well, I'll get back to rebuilding the box. l8trs!
2018-12-25 16:42:50 <hatramatra> silverwillow: the initramfs is already dealing with rootfs being multi-device btrfs, so that is not a concern.
2018-12-25 16:43:39 <hatramatra> however, the device scan runs only if rootfs is btrfs during the initramfs
2018-12-25 16:44:35 <silverwillow> yes i saw that here when I was looking for the solution for my particular non-root btrfs mount issue: https://bugs.alpinelinux.org/issues/9539 (also mentioned above)
2018-12-25 16:46:09 <silverwillow> then I went looking for a place to hook-in the 'btrfs device scan' and thenm mount - and found local.d - seemed a reasonable place to put it.
2018-12-25 16:46:43 <hatramatra> mps: from btrfs wiki: "In most distributions udev runs btrfs device scan automatically when a block device is discovered."
2018-12-25 16:47:00 <hatramatra> but I can't find how exactly it is done
2018-12-25 16:48:37 <mps> hmm, could that bee done with mdev?
2018-12-25 16:52:38 <mps> not quite to your problem but could give you some hints https://wiki.gentoo.org/wiki/Btrfs/System_Root_Guide
2018-12-25 16:53:10 <mps> if you didn't read it already, ofc
2018-12-25 16:53:29 <hatramatra> how is it in Alpine, is mdev(d) installed/used by default?
2018-12-25 16:53:57 <danieli> hatramatra: eudev
2018-12-25 16:57:26 <mps> eudev could be right solution but i had a hope that the mdev could help
2018-12-25 16:57:51 <danieli> no, what i mean is that alpine uses eudev
2018-12-25 17:06:37 <mps> yes, but not by default
2018-12-25 17:12:23 <hatramatra> OK, systemd-udev handles this all internally. They don't run 'btrfs device scan' at all.
2018-12-25 17:13:28 <hatramatra> thus all distributions using systemd rely on that
2018-12-25 17:15:47 <mps> how solved that devuan or void linux, if they solved
2018-12-25 17:16:07 <mps> I don't have btrfs so cannot check
2018-12-25 17:17:50 <hatramatra> but default, neither eudev nor mdevd is installed, correct?
2018-12-25 17:20:13 <mps> on Alpine mdev is installed by default
2018-12-25 17:20:54 <mps> busybox mdev, I mean
2018-12-25 17:21:53 <hatramatra> yes, that is what I guessed
2018-12-25 17:26:36 <hatramatra> it seems that both Devuan (confirmed) and Void (don't have an installation around) are using eudev
2018-12-25 17:27:15 <hatramatra> I can't mess with that exising Devuan installation, but maybe I can quickly install new one to see how it behaves there.
2018-12-25 17:28:50 <hatramatra> but if eudev magically handled multi-device btrfs filesystems, would the solution be just go ahead and install eudev? What are consequences of that?
2018-12-25 17:30:03 <mps> no problem, I install eudev nearly on all my boxes
2018-12-25 17:30:23 <mps> running X without it is pain
2018-12-25 17:46:15 <hatramatra> thank you all for the tips. I will look closer at eudev and update the bug report with my findings.
2018-12-25 17:47:31 <hatramatra> apparently, one can list multiple device= parameters in fstab, but then device names needs to be stable
2018-12-25 19:28:19 <sydney112> Hey, I have a question about mirroring the repo - are you guys open to new mirrors?
2018-12-25 19:28:25 <sydney112> Is this the right channel to ask?
2018-12-25 19:31:51 <_ikke_> sydney112: good enough
2018-12-25 19:33:17 <_ikke_> sydney112: We are open to new mirrors, though, adding mirrors in locations that already have enough mirrors wouldn't add a lot
2018-12-25 19:33:36 <sydney112> That's fair, thanks
2018-12-25 19:33:51 <_ikke_> https://mirrors.alpinelinux.org for a list of current mirrors
2018-12-25 19:34:40 <sydney112> I'm consulting that list now
2018-12-25 19:35:02 <mps> sydney112: maybe you can set source packages archive for Alpine if you have enough disk space
2018-12-25 19:38:49 <sydney112> Depends on how much disk space, I suppose. I assume I have to be able to host every single Alpine version in the main repos?
2018-12-25 19:40:46 <sydney112> I'm speccing new servers, so I have the advantage of being able to be flexible on that depending on how much it would take
2018-12-25 19:40:50 <mps> i suspect that anyone knows how much space is needed for all source files, but that could be calculated with some scripts which could scan source packages
2018-12-25 19:42:42 <mps> that could be discussed on #alpine-devel or #alpine-infra
2018-12-25 19:48:27 <TBB> I have a rough idea, even a rather exact number but it's at work, I'm not and won't be until the 27th
2018-12-25 19:48:47 <TBB> but it's, let's say, a manageable size
2018-12-25 19:51:11 <_ikke_> v3.8 seems to be 20G if I look at the correct place
2018-12-25 19:51:13 <TBB> I run a full source fetch a couple times per year, and I think the build team also does it
2018-12-25 19:51:20 <TBB> _ikke_: sounds about right
2018-12-25 19:52:04 <_ikke_> How come the repository itself is 64G?
2018-12-25 19:52:29 <TBB> builds for all platforms perhaps?
2018-12-25 19:53:02 <_ikke_> That's just one platform
2018-12-25 21:13:57 <silverwillow> Finally getting round to building Mono. Have found this APKBUILD file: https://raw.githubusercontent.com/spritsail/mono-apk/684da8065349abe94f44fb9cf84cd55f0e6d4a30/APKBUILD
2018-12-25 21:14:25 <silverwillow> but getting this error:  not founduild: /mnt/store/home/avi/tmp/src/mono/APKBUILD: line 42:
2018-12-25 21:14:34 <silverwillow> what am I doing wrong?
2018-12-25 21:15:28 <silverwillow> am calling: 'abuild checksum && abuild -r' as per the wiki.
2018-12-25 21:18:41 <Lord> hi, i have a little problem : i have an init script that i put in the default runlevel but it's not started at boot
2018-12-25 21:18:57 <Lord> if i /etc/init.d/isso start, it starts well
2018-12-25 21:19:08 <Lord> but nothing at boot time
2018-12-25 21:19:24 <Lord> what can i do to debug this ?
2018-12-25 21:20:01 <_ikke_> Logging?
2018-12-25 21:20:31 <mps> rc-service isso status
2018-12-25 21:22:04 <Lord> rc-service isso status doesn't return anything
2018-12-25 21:22:14 <Lord> nothing in the logs
2018-12-25 21:23:33 <clandmeter> rc-update --update
2018-12-25 21:24:17 <Lord> i tried putting it in the boot runlevel but still nothing
2018-12-25 21:24:29 <Lord> i wonder if the service file is good
2018-12-25 21:26:04 <Lord> clandmeter : it doesn't change anything
2018-12-25 21:27:27 <clandmeter> does it list in rc-status --all?
2018-12-25 21:28:22 <Lord> no
2018-12-25 21:29:39 <clandmeter> find /etc/runlevels |grep isso
2018-12-25 21:29:54 <Lord> it is in default
2018-12-25 21:31:04 <clandmeter> what is isso?
2018-12-25 21:31:18 <Lord> it's a commenting services for websites written in python
2018-12-25 21:31:34 <Lord> https://posativ.org/isso/
2018-12-25 21:31:38 <clandmeter> can you pastbin the initd?
2018-12-25 21:31:39 <Lord> like disqus but selfhosted
2018-12-25 21:32:49 <Lord> http://vpaste.net/Ki6w2 here it is
2018-12-25 21:33:27 <clandmeter> thats not a openrc service
2018-12-25 21:34:22 <Lord> ho ?
2018-12-25 21:34:42 <Lord> it's a copy/paste from motd init file
2018-12-25 21:34:46 <_ikke_> https://github.com/OpenRC/openrc/blob/master/STYLE-GUIDE.md
2018-12-25 21:34:48 <Lord> just tweaked a bit
2018-12-25 21:34:51 <clandmeter> https://github.com/OpenRC/openrc/blob/master/service-script-guide.md
2018-12-25 21:35:07 <_ikke_> sysv initscript != openrc init script
2018-12-25 21:46:41 <Lord> thank you, i converted the script (very easily) and it works perfectly ;-)
2018-12-25 21:46:56 <clandmeter> yw
2018-12-26 00:36:44 <Greyfn> can anybody confirm to me if the 'apkovl=' kernel command line parameter actually accepts http as the documentation suggests?
2018-12-26 03:34:45 <chzbacon> hey guys, if i wanted to get a package in the repo updated how would i go about doing that? is there some place/person i could contact? or maybe someone could point me in the right direction to do it myself?
2018-12-26 05:25:05 <danieli> Greyfn: can't confirm it, but judging by docs, it should indeed work if you've also set `ip`
2018-12-26 15:07:44 <Processus42> Hello guys. Do someone have an automated way to build a toolchain for Alpine ? (I'm using Gentoo, and I don't want to build my packages on Alpine)
2018-12-26 15:32:23 <_ikke_> Processus42: Why don't you want to build on alpine?
2018-12-26 15:37:54 <PureTryOut[m]> You can just use an Alpine chroot on your Gentoo system, works quite well
2018-12-26 15:38:01 <_ikke_> nod
2018-12-26 15:38:12 <_ikke_> I tend to use docker for it, but you can use whatever you like
2018-12-26 15:38:57 <danieli> i wrote a script to make small disposable alpine chroot environments
2018-12-26 15:38:59 <Processus42> The only solution I have right now is to setup an Alpine container (I'm using unprivileged containers) or a VM (That's my current setup). I want to get free of theses, plus it will soon be integrated in a GitLab runner, and I want to avoid these solutions (Yet, I might considere the container as a good solution here).
2018-12-26 15:39:08 <danieli> but yeah, i tend to use docker and qemu for other architectures
2018-12-26 15:39:26 <Processus42> I never tried choorting, I think I'll give it a try
2018-12-26 15:39:46 <_ikke_> Processus42: gitlab runners can use docker containers to run the builds in
2018-12-26 15:40:11 <Processus42> Yep, but I don't want to use Docker as well :p
2018-12-26 15:40:21 <Processus42> (My own containers are LXC ones)
2018-12-26 15:40:59 <Processus42> I realise that containers are a good solutions, but I'd want to try alternatives
2018-12-26 15:41:06 <Processus42> I'm interested in your script, danieli
2018-12-26 15:42:45 <_ikke_> advantage of docker is that you already have existing images that you can easily download
2018-12-26 15:43:03 <_ikke_> But if you want to arrange that yourself, a chroot should work as well
2018-12-26 15:45:40 <danieli> it's not released, sorry
2018-12-26 15:45:45 <danieli> i'd have to clean it up a bit
2018-12-26 15:45:49 <danieli> it's pretty trivial stuff anyway
2018-12-26 15:46:23 <Processus42> _ikke_: I prefer to have full control of my container (I'm not experienced enough with Docker, and I got a bit repelled the last time I checked internals)
2018-12-26 15:46:51 <mtnt> Hi there!
2018-12-26 15:46:54 <Processus42> Containers offer more security than chrooted environments, though. But dangerous code shouldn't be merged on master branch ;)
2018-12-26 15:47:09 <mtnt> Acpid marked as crashed at rc-status also acpid failed to stop on reboot. Am i missing something?
2018-12-26 15:47:28 <_ikke_> I use docker more for convenience than security
2018-12-26 15:47:43 <danieli> same, neither are a very good option if you're running untrusted code
2018-12-26 15:58:24 <mtnt> do i need to install the udev for acpid to stop those errors? sorry to be pia here
2018-12-26 16:15:12 <Processus42> mtnt:  "those" errors ?
2018-12-26 16:16:05 <danieli> which errors, specifically? I didn't see you posting any
2018-12-26 16:16:38 <Processus42> I just realised I didn't got an answer for Alpine toolchain building. When did self-building got implemented ? I'll try to get a look at the old way things where compiled.
2018-12-26 16:17:01 <_ikke_> probably already a long time ago
2018-12-26 16:18:37 <Processus42> Okay, I'll start "from scratch" then :)
2018-12-26 16:21:31 <mtnt> Processus42, you just missed my two lines
2018-12-26 16:21:33 <mtnt> Acpid marked as crashed at rc-status also acpid failed to stop on reboot. Am i missing something?
2018-12-26 16:21:37 <mtnt> do i need to install the udev for acpid to stop those errors? sorry to be pia here
2018-12-26 16:22:18 <mtnt> and thanks for reply
2018-12-26 16:23:20 <danieli> but that's not an error message, that's just the symptom
2018-12-26 16:23:31 <danieli> logs?
2018-12-26 16:24:01 <mtnt> sympthom of a noob? i can't give a log, since im writing from the different machine
2018-12-26 16:24:37 <mtnt> one thing i mentioned before is that those "symptoms" lost after setup-xorg-base
2018-12-26 16:24:52 <mtnt> but i don't want to install the xorg for now
2018-12-26 16:26:21 <mtnt> it looks like im missing some package to install
2018-12-26 16:27:04 <mtnt> and danieli, thanks you so kind to reply also
2018-12-26 16:34:12 <mtnt> may be the cause is lack of linux-firmware package? Is it installed by default on fresh system?
2018-12-26 16:36:43 <mtnt> okay. i see that best way is to give logs. which logs exactly and where better to put (is pastebin link is allowed here) them?
2018-12-26 16:46:21 <mtnt> i got it, better go now. thanks
2018-12-26 17:28:27 <Processus42> Does abuild have a man ? Skimming through its source-code makes me think it can cross-build natively.
2018-12-26 17:58:53 <_ikke_> Processus42: Someone is working on man pages, but not atm
2018-12-26 17:59:04 <_ikke_> Most information is on the wiki atm
2018-12-26 19:38:59 <Processus42> Okay
2018-12-26 19:42:42 <bleb> anyone know why firefox-esr is stuck on version 52?
2018-12-26 19:52:39 <Processus42> I looked too far, what I was looking for is in the usage of abuild: "To activate cross compilation specify in environment:" "CHOST  Arch or hostspec of machine to generate packages for" "CTARGET  Arch or hostspec of machine to generate compiler for"
2018-12-26 19:53:25 <AinNero> bleb: afaik its rust support
2018-12-26 19:57:11 <bleb> AinNero: yet firefox 62 works fine?
2018-12-26 19:57:28 <AinNero> yeah, not esr
2018-12-26 19:57:42 <bleb> 62 doesn't require rust?
2018-12-26 20:00:32 <AinNero> dont know
2018-12-26 20:02:16 <AinNero> but after the crap mozilla people pull, i'm glad if any version builds at all
2018-12-26 20:06:22 <danieli> the holy browser war is probably worse than the holy init system war
2018-12-26 20:06:38 <danieli> but not more than the holy OS (or distro) wars
2018-12-26 20:07:39 <bleb> the war is not between browsers, but between those who don't want to use browsers and those who want to force you to use browsers
2018-12-26 20:08:01 <danieli> it was just in a manner of speaking, take it wiht a pinch of salt
2018-12-26 20:08:03 <AinNero> i'd be more happy if browser developers would care a bit more about maintainability
2018-12-26 20:08:36 <AinNero> currently we are at the state that upstream is close to just build it itself and ship a docker container or flatpak instead
2018-12-26 20:09:07 <danieli> that... would honestly not surprise me
2018-12-26 20:09:16 <AinNero> which is cancer for free software, that depends on the fact that people are able to compile stuff themselves
2018-12-26 20:10:21 <bleb> browser complexity has been cancer for free software for a long time
2018-12-26 23:09:19 <danieli> Processus42: i forgot about this https://github.com/alpinelinux/alpine-chroot-install
2018-12-26 23:44:11 <desnudopenguino> so how's the vm stack on alpine other than running containers on it? with such a small footprint from the actual OS, it might be my next favorite thing
2018-12-27 09:31:48 <danieli> desnudopenguino: 'the vm stack' is a bit ambiguous, mind being a little more specific/
2018-12-27 09:44:31 <desnudopenguino> danieli: from what i've seen, a lot of people like it for containers, how's it handle as a kvm host or xen dom0?
2018-12-27 09:51:11 <desnudopenguino> i've been playing around with docker in alpine, and i've been slowly piecing together a vm server, now i'm wondering if alpine linux would be a good choice as a host, and which way i should set it up if so.
2018-12-27 09:52:18 <desnudopenguino> i've been runnnig proxmox for a little over a year, but i wouldn't mind something leaner as the base.
2018-12-27 10:20:43 <danieli> desnudopenguino: handles quite well
2018-12-27 10:20:56 <danieli> i've used it as a guest in xen, qemu, containers, and vmware
2018-12-27 10:21:03 <danieli> and as a xen, qemu, container host
2018-12-27 10:21:19 <danieli> don't touch proxmox much though, so i have no personal experience with alpine there
2018-12-27 10:23:00 <desnudopenguino> proxmox is debian with a few baked in useful tools and a nice web interface for VM management. i just like things minimalist, lean and clean.
2018-12-27 10:24:56 <desnudopenguino> is there a preference one way or the other between xen and qemu/kvm? i plan on running windows, linux, and BSD guests if that makes any difference
2018-12-27 10:29:45 <danieli> i knwo what it is
2018-12-27 10:29:50 <danieli> s/wo/ow
2018-12-27 10:38:46 <desnudopenguino> danieli: do you have any suggestion/opinion between qemu and xen with alpine as the host?
2018-12-27 10:39:10 <danieli> desnudopenguino: nothing that's alpine-specific
2018-12-27 15:22:24 <mtnt> Hiya. Me again, sorry for my rudeness yesterday, also for my poor language. I returned with logs, can someone take a look? https://pastebin.com/raw/CqnDBBRG Thanks!
2018-12-27 15:25:45 <danieli> the issue was acpid crashing and not stopping on reboot, right?
2018-12-27 15:27:32 <mtnt> yes, danieli
2018-12-27 15:27:39 <mtnt> thanks that you remember
2018-12-27 15:30:18 <danieli> what's the bugs.alpinelinux.org issue number for it?
2018-12-27 15:31:07 <mtnt> did not searched it yet
2018-12-27 15:31:12 <_ikke_> Not everything should immediately be reported as a bug. You first find out if it's actually a bug or just something with the setup
2018-12-27 15:31:27 <danieli> true, i've seen something like this before though, on other distros
2018-12-27 15:31:29 <mtnt> it is acting so out of the box
2018-12-27 15:31:52 <danieli> acpid has to start before hald
2018-12-27 15:31:57 <mtnt> and as i mentioned it's gone after setup-xorg-base
2018-12-27 15:31:59 <_ikke_> danieli: the way you worded it sounded like: "and why didn't you file a redmine bug for it yet"
2018-12-27 15:32:12 <danieli> _ikke_: that was unintentional
2018-12-27 15:32:14 <_ikke_> ok
2018-12-27 15:33:36 <mtnt> did not find something similar @ bugs.alpinelinux.org
2018-12-27 15:35:19 <_ikke_> Do you have some special hardware?
2018-12-27 15:35:53 <mtnt> like my E-MU 1212m PCI card? Not any more special here
2018-12-27 15:36:27 <mtnt> PSU? It is seasonic 850w prime
2018-12-27 15:36:46 <mtnt> mobo itself kinda special
2018-12-27 15:40:00 <mtnt> not overclocked. HPET disabled
2018-12-27 15:40:36 <mtnt> hyperthreading and other Intel's features is disabled also
2018-12-27 15:41:10 <mtnt> i can try to reset bios to defaults an d try to reinstall
2018-12-27 15:42:02 <danieli> sounds unlikely it's related to that, it sounds like something is causing acpid to lose access to event0
2018-12-27 15:42:13 <danieli> i've only seen that happen in one case, but i haven't seen that on alpine
2018-12-27 15:42:33 <mtnt> why then it came to normal after setup-xorg-base?..
2018-12-27 15:44:15 <danieli> it might be because either because it installs eudev, or because it installs xorg-input-libinput
2018-12-27 15:44:56 <mtnt> thats it. i'll try to install those and come back with results. do i need to stop and disable mdev?
2018-12-27 15:45:19 <danieli> this is all setup-xorg-base is https://git.alpinelinux.org/cgit/alpine-conf/tree/setup-xorg-base.in
2018-12-27 15:46:14 <mtnt> pciutils maybe?
2018-12-27 15:46:45 <danieli> hm?
2018-12-27 15:47:10 <mtnt> i'll check it out, brb
2018-12-27 15:47:12 <mtnt> thanks
2018-12-27 16:09:54 <mtnt> hiya, me again. Probably i find the cause, got to check on fresh Alpine install though
2018-12-27 16:10:16 <mtnt> Added "kms" to mkinitfs.conf
2018-12-27 16:10:43 <mtnt> and it's working like a charm now. now errors
2018-12-27 16:10:47 <mtnt> no
2018-12-27 16:13:05 <mtnt> I'll go and do reinstall now. Only thing i'm going to install after this will be nano. I'll return with results after that.
2018-12-27 16:22:41 <mtnt> ta-daaa! proceed to the next level!
2018-12-27 16:23:09 <mtnt> only three letters make it works "kms" in mkinitfs.conf
2018-12-27 16:24:36 <mtnt> is logs going to be useful at this state?
2018-12-27 17:27:20 <mtnt> logs https://pastebin.com/raw/edYFEnNd
2018-12-27 17:29:35 <mtnt> btw, what do you think, is it possible to have alsa working on my rig? and the xorg desktop (dwm would be enough for me)?..
2018-12-27 18:50:37 <itguys> Weird, random question: Is Alpine primarily for server use (that's what I've mainly seen) or do a lot of people also use it for desktop use?
2018-12-27 18:52:09 <danieli> there are people who use it for workstations/desktops
2018-12-27 18:52:28 <danieli> personally, i'd say the experience is quite a bit more pleasant on servers at the moment
2018-12-27 18:52:57 <itguys> That's what I figured. I'm redoing my entire network and I was looking at using different distributions. Thank you.
2018-12-27 18:55:32 <itguys> I'm more interested in using small specialized tools over large multifunction ones, so Alpine interested me. The server I'm going to set up will only have a couple functions so I don't need a bunch of features or packages.
2018-12-27 18:56:30 <danieli> the base system is small, makes it pretty resource efficient
2018-12-27 19:43:08 <Marbug> Hi, I installed perl-cache-cache due to the error I have in munin: "Missing dependency Cache::Cache". Still after I had installed the package, the error persists, is there something I'm missing, or is it another package I might need to use? it's on alpine 3.7
2018-12-28 02:53:57 <KH405_TV> I know I ask before, but I'm not sure of the answer, Where should config file for a program installed in /usr/bin be ?
2018-12-28 02:54:20 <KH405_TV> Where should I set the home directory of the user running the program at ?
2018-12-28 02:56:18 <KH405_TV>  /etc/appname ?
2018-12-28 03:22:18 <KH405_TV>  /etc/conf.d/appname ?
2018-12-28 03:48:00 <AinNero> KH405_TV: /etc/conf.d/appname is for openrc services
2018-12-28 03:49:01 <AinNero>  /etc/appname is appropiate for config files
2018-12-28 03:49:07 <AinNero> homedir should be in /var, probably
2018-12-28 03:55:50 <KH405_TV> I'm not sure I understand the difference between /etc/conf.d and /etc
2018-12-28 03:56:22 <KH405_TV> You mean things that are run by openRC should be in /etc/conf.d ?
2018-12-28 04:00:04 <AinNero> no, /etc/conf.d is configuration for openrc services
2018-12-28 04:00:18 <AinNero> for each service in /etc/init.d
2018-12-28 04:00:54 <AinNero> if you want to define a variable for /etc/init.d/dickbutt, you'd write FOO=BAR into /etc/conf.d/dickbutt
2018-12-28 04:01:24 <AinNero> and the application will read its own configuration from /etc/dickbutt/dickbutt.conf or so
2018-12-28 04:01:54 <KH405_TV> Aight
2018-12-28 04:02:16 <KH405_TV> So /etc/appname it is then?
2018-12-28 04:02:28 <AinNero> for configuration files, yeah
2018-12-28 04:03:00 <AinNero> look at /etc/mkinitfs for reference
2018-12-28 04:03:12 <AinNero> thats generic unix-like convention and not specific to alpine
2018-12-28 04:25:54 <KH405_TV> I don't even have that lol
2018-12-28 04:29:58 <AinNero> then you are probably using containers wrong :P
2018-12-28 07:57:14 <mtnt> Hi! Thinking how does correlate kms in mkinitfs.conf and nomodeset kernel param...
2018-12-28 14:19:38 <daggs1> Greetings, is there a way to install alpine Linux using serial console?
2018-12-28 14:29:27 <_ikke_> yes
2018-12-28 14:29:42 <_ikke_> https://wiki.alpinelinux.org/wiki/Enable_Serial_Console_on_Boot
2018-12-28 14:43:33 <daggs1> _ikke_: the iso is rw?
2018-12-28 14:43:42 <_ikke_> no
2018-12-28 14:44:27 <daggs1> mmmm so how I can change syslinux,cfg?
2018-12-28 14:44:40 <_ikke_> can you make a bootable usb disk?
2018-12-28 14:44:51 <_ikke_> drive*
2018-12-28 14:45:12 <daggs1> I intend to install alpine as kvm guest so I guess I can
2018-12-28 14:45:20 <daggs1> is there any tutorial on this?
2018-12-28 14:45:51 <AinNero> there exist unofficial images with serial console enabled per default
2018-12-28 14:46:01 <AinNero> this is a use case that pops up quite often..
2018-12-28 14:46:37 <daggs1> unofficial how?
2018-12-28 14:47:04 <AinNero> i build such an image and uploaded it to a site i own, with console=ttyS0 enabled per default
2018-12-28 14:47:32 <daggs1> ok, can you point me to it?
2018-12-28 14:47:38 <AinNero> https://w1r3.net/paste/alpine-standard-v3.8-x86_64.iso
2018-12-28 14:47:48 <AinNero> its some months old, because i built it for another person
2018-12-28 14:48:55 <daggs1> so the only diff from the original is the console support?
2018-12-28 14:49:42 <AinNero> its signed with unofficial keys and has 'console=ttyS0,115200' added to the kernel command line
2018-12-28 14:50:48 <AinNero> what device are you trying to install on?
2018-12-28 14:50:55 <AinNero> pcengine board?
2018-12-28 14:51:36 <daggs1> AinNero: qemu x64
2018-12-28 14:51:44 <AinNero> uhhmm
2018-12-28 14:52:04 <AinNero> if you boot qemu, you can just use the netboot images
2018-12-28 14:52:41 <daggs1> I went with the qemu entry in the wiki
2018-12-28 14:52:49 <AinNero> link?
2018-12-28 14:53:09 <AinNero> http://boot.alpinelinux.org/, scroll down to 'testing netboot'
2018-12-28 15:01:12 <daggs1> thanks, will look into it
2018-12-28 15:03:30 <AinNero> i'll throw this into the room without context, maybe its of use to you https://github.com/nero/apkovl/blob/master/utils/run-qemu.sh
2018-12-28 19:03:30 <daggs1> AinNero: tried the netboot images, I get the installer to boot but for some reason it fails to mount the initramfs, any ideas?
2018-12-28 19:03:50 <daggs1> I get "initramfs emergency recovery shell launched."
2018-12-28 19:04:02 <daggs1> and it drops to shell
2018-12-28 19:04:13 <AinNero> daggs1: thats the symptom of another error
2018-12-28 19:04:28 <AinNero> actual error should be above in the logs
2018-12-28 19:06:10 <AinNero> i guess its that its unable to find the repositories
2018-12-28 19:06:15 <AinNero> because it has no network
2018-12-28 19:06:22 <daggs1> AinNero: I see it mounts vda as boot but vda is the file I want to install it on
2018-12-28 19:06:43 <daggs1> sec, I can handle that
2018-12-28 19:06:46 <AinNero> uhm
2018-12-28 19:07:01 <AinNero> why should it mount vda....
2018-12-28 19:07:09 <AinNero> you did not specify root=, right?
2018-12-28 19:07:18 <daggs1> that is correct
2018-12-28 19:07:32 <daggs1> what should be the root value?
2018-12-28 19:07:37 <AinNero> not set
2018-12-28 19:08:02 <AinNero> not set = start diskless system (aka installer)
2018-12-28 19:08:24 <KH405> So user ran as a service home directory should be in /var and the config for those user should be in /etc/appname right ?
2018-12-28 19:08:37 <daggs1> I don't I have console=ttyS0, ip=dhcp modloop=<path to local modloop file>
2018-12-28 19:08:59 <AinNero> daggs1: where is your alpine_repo ?
2018-12-28 19:09:16 <AinNero> uhm, <path to local modloop file> does not work, it needsa to be http url
2018-12-28 19:09:36 <AinNero> the quest cannot access any files on the host except for the kernel and initrd
2018-12-28 19:09:52 <daggs1> where can I find the url to them?
2018-12-28 19:10:41 <AinNero> either in the netboot config http://boot.alpinelinux.org/boot.ipxe or in the script i pasted earlier
2018-12-28 19:11:02 <daggs1> script? can you repaste it?
2018-12-28 19:11:37 <AinNero> https://github.com/nero/apkovl/blob/master/utils/run-qemu.sh
2018-12-28 19:11:56 <daggs1> sec
2018-12-28 19:16:23 <daggs1> yay!
2018-12-28 19:16:27 <daggs1> installation underway
2018-12-28 19:16:54 <daggs1> that's it?
2018-12-28 19:17:36 <AinNero> yeah
2018-12-28 19:17:50 <daggs1> ahh, that the installer, now the actual installation
2018-12-28 19:22:11 <daggs1> whats the diff between data and sys layouts?
2018-12-28 19:22:43 <daggs1> nice help!
2018-12-28 19:22:44 <AinNero> data boots in ram, like the installer, but loads the configuration and user data from a tarball
2018-12-28 19:22:56 <AinNero> sys is regular installation on disk
2018-12-28 19:23:17 <AinNero> hm, i think data is only having /var on disk or so... weird thing i never used
2018-12-28 19:23:51 <daggs1> install failed: mount: mounting /dev/vda3 on /mnt failed: No such device
2018-12-28 19:25:49 <AinNero> as part of setup-alpine?
2018-12-28 19:26:43 <daggs1> yes, started it again, dded vda, maybe it was the issue
2018-12-28 19:28:48 <daggs1> failed again, same error
2018-12-28 19:30:18 <AinNero> can you do fdisk /dev/vda and do 'p' ?
2018-12-28 19:30:28 <AinNero> curious if the partitioning failed
2018-12-28 19:31:15 <daggs1> nope, all ok
2018-12-28 19:31:45 <AinNero> what okay? does it have vda3?
2018-12-28 19:32:06 <daggs1> yes, all the partitions are ok
2018-12-28 19:32:41 <AinNero> you tried to install with sys mode?
2018-12-28 19:33:28 <AinNero> its difficult for me to understand what went wrong here
2018-12-28 19:33:30 <daggs1> yes, selected sys mode
2018-12-28 19:34:37 <AinNero> maybe something got messed up by the repeated tries
2018-12-28 19:34:50 <AinNero> otherwise, bootling like that is how i installed all my machines
2018-12-28 19:36:02 <daggs1> ok, I'll dd vda again, restart the installer and paste it to a file, if it fails I'll upload it do you can inspect it
2018-12-28 19:38:40 <AinNero> quick hack: if you modify /sbin/setup-disk for the first line to be #!/bin/sh -x, you'll get much verbose output from the script
2018-12-28 19:40:59 <daggs1> added that, will see what is in the logs
2018-12-28 19:42:45 <daggs1> same shit
2018-12-28 19:42:48 <daggs1> checking the log
2018-12-28 19:45:54 <AinNero> ohhh
2018-12-28 19:46:13 <AinNero> daggs1: 'No such device' is ENODEV, "filesystemtype not configured in the kernel."
2018-12-28 19:46:15 <daggs1> AinNero: http://paste.debian.net/1057811/
2018-12-28 19:46:23 <AinNero> the kernel module for ext4 or ext3 is probably not loaded
2018-12-28 19:46:44 <daggs1> check the log, it doesn't find that modile
2018-12-28 19:46:47 <daggs1> module
2018-12-28 19:46:56 <AinNero> > modprobe: can't change directory to '4.14.89-0-virt': No such file or directory
2018-12-28 19:47:03 <AinNero> right in your bootlog
2018-12-28 19:47:17 <daggs1> ok
2018-12-28 19:47:18 <AinNero> your modloop is vanilla, your kerne is virt
2018-12-28 19:47:29 <daggs1> good catch
2018-12-28 19:48:16 <daggs1> starting again
2018-12-28 19:52:58 <daggs1> passed that point :)
2018-12-28 20:01:20 <daggs1> install is done
2018-12-28 20:04:39 <daggs1> I'm unable to install w_scan, getting unsatisfiable constraints errors
2018-12-28 20:07:42 <AinNero> w_scan is in testing, which likely depends on edge
2018-12-28 20:08:17 <daggs1> ok, how can I upgrade to edge/testings?
2018-12-28 20:08:41 <AinNero> change the repository urls in /etc/apk/repositories
2018-12-28 20:09:31 <AinNero> here is mine for reference: http://0x0.st/snS0.txt
2018-12-28 20:10:31 <daggs1> done
2018-12-28 20:11:40 <AinNero> apk update, apk upgrade, should update a hell of stuff
2018-12-28 20:11:54 <daggs1> 29 pkgs
2018-12-28 20:12:34 <daggs1> thing is, reboot didn't went well, I've killed the vm and now it doesn't boots
2018-12-28 20:14:29 <AinNero> 
ACTION sighs

2018-12-28 20:14:58 <AinNero> how does it error out?
2018-12-28 20:15:18 <AinNero> i liked the terminal transscript you sent me earlier
2018-12-28 20:15:19 <daggs1> cannot see, no console
2018-12-28 20:17:17 <AinNero> crap
2018-12-28 20:29:31 <daggs1> how can I edit the cmdline?
2018-12-28 20:31:07 <_ikke_> daggs1: What bootloader do you use
2018-12-28 20:31:41 <daggs1> no idea, how can I check thar?
2018-12-28 20:31:47 <daggs1> *that
2018-12-28 20:32:02 <_ikke_> What do you see when it boots?
2018-12-28 20:32:26 <_ikke_> At the beginning
2018-12-28 20:32:43 <daggs1> _ikke_: I'm running it in a headless vm, so I cannot see anything at the start
2018-12-28 20:32:44 <_ikke_> If you see "Boot:"
2018-12-28 20:32:49 <_ikke_> ah ok
2018-12-28 20:33:12 <_ikke_> check the /boot directory
2018-12-28 20:33:45 <daggs1> edit extlinux.conf and that's it?
2018-12-28 20:34:03 <_ikke_> yes
2018-12-28 20:34:42 <daggs1> cool, it works
2018-12-28 20:35:16 <AinNero> /boot/extlinux.conf is overwritten on update by update-extlinux
2018-12-28 20:35:29 <AinNero> permanent changes can be done in /etc/update-extlinux.conf
2018-12-28 20:35:48 <AinNero> (assuming a system installed in sys mode)
2018-12-28 20:37:26 <daggs1> I've reinstalled the system, ran upgrade to edge, anything else I need to run before the reboot?
2018-12-28 20:42:04 <daggs1> AinNero: ?
2018-12-28 20:42:30 <AinNero> when a new kernel is installed, update-extlinux will be executed
2018-12-28 20:42:43 <AinNero> which regenerates /boot/extlinux.conf unless thats prohibited in /etc/update-extlinux.conf
2018-12-28 20:43:14 <daggs1> so just rebbot then
2018-12-28 20:45:07 <daggs1> I see what is happening now, it is stuck after  * Starting busybox crond ... [ ok ]
2018-12-28 20:45:21 <danieli> i had an issue with chrony when i had no network, it kept waiting for chrony to connect to ntp servers
2018-12-28 20:45:24 <danieli> try hitting ctrl+c
2018-12-28 20:45:29 <danieli> see if that works
2018-12-28 20:46:07 <_ikke_> AinNero: thanks, I always forget that
2018-12-28 20:46:23 <daggs1> ok after some timeout it booted to login
2018-12-28 20:46:51 <daggs1> is there a boot log to see what caused the delay?
2018-12-28 20:47:08 <danieli> daggs1: we don't use systemctl, so you should be able to just look at /var/log/messages or something along those lines
2018-12-28 20:47:13 <danieli> s/systemctl/systemd/
2018-12-28 20:47:53 <daggs1> chronyd too more than a minute to boot
2018-12-28 20:51:07 <daggs1> is that logica?
2018-12-28 20:51:12 <daggs1> logical?
2018-12-28 20:53:52 <daggs1> danieli: looks like you are right, it is chrony waiting to connect but it has network
2018-12-28 20:56:11 <danieli> honestly, it sounds like a bug
2018-12-28 20:56:15 <danieli> it happened with chrony as well
2018-12-28 20:57:39 <danieli> the system had a really long timeout for ntp when there was no network connection
2018-12-28 20:58:24 <daggs1> I get ip before that, how it is possible it doesn't have network?
2018-12-28 20:58:43 <danieli> is ntp configured right?
2018-12-28 21:00:04 <daggs1> no idea, I didn't configured it
2018-12-28 21:01:13 <daggs1> how can I check it?
2018-12-28 21:08:39 <_ikke_> /etc/chrony*
2018-12-28 21:10:59 <daggs1> no idea what can be bad in that config file
2018-12-28 21:16:00 <daggs1> found a bug in tvheadend, it looks for /etc/tvheadend-git when /etc/tvheadend exists
2018-12-28 21:44:12 <sy> Hi, where can I find the setup-disk script? Is there a package for it?
2018-12-28 21:44:53 <_ikke_> https://git.alpinelinux.org/cgit/alpine-conf/tree/
2018-12-28 21:45:14 <_ikke_> https://pkgs.alpinelinux.org/package/edge/main/x86_64/alpine-conf
2018-12-28 21:46:06 <sy> sweet, thanks
2018-12-28 21:51:24 <sy> hi, i'm getting " does not seem to be a mount point" when trying to "setup-disk -m sys /" from within a chroot on arch
2018-12-28 21:51:50 <sy> should i rbind the root filesystem inside... itself to trick the script?
2018-12-28 22:04:05 <sy> so i did that, then this happens: http://ix.io/1x0W
2018-12-28 22:04:59 <sy> tldr ERROR: unsatisfiable constraints: \\ linux-amd64 (missing): \\ required by: world[linux-amd64] after running setup-disk -m sys /mnt/ with / mounted on /mnt (yes really)
2018-12-28 22:05:19 <danieli> mounting it like that seems pretty strange, when you chroot, / is supposed to be the root
2018-12-28 22:06:22 <_ikke_> linux-amd64 is not correct afaik
2018-12-28 22:06:27 <danieli> indeed
2018-12-28 22:06:42 <sy> danieli: setup-disk doesn't like operating on /
2018-12-28 22:07:42 <danieli> when i install it from a chroot, i do it by hand, some of the scripts tend to screw up in wonky chroots
2018-12-28 22:08:23 <sy> oh
2018-12-28 22:08:42 <danieli> anyway, the package name is linux-{vanilla,virt} and apk automatically pulls the package built for the correct architecture
2018-12-28 22:25:58 <sy> so i followed https://wiki.alpinelinux.org/wiki/LVM_on_LUKS and i'm getting this: https://0x0.st/sn1T.png
2018-12-28 22:27:34 <danieli> grub-mkconfig didn't work properly for me due to some udev issues, i had to rbind mount it under /mnt
2018-12-28 22:27:46 <danieli> even then, it complained at lvm memory leaks, but it did owrk
2018-12-28 22:27:54 <danieli> s/ow/wo/
2018-12-28 22:28:32 <TBB> LVM file descriptor leaks?
2018-12-28 22:30:26 <sy> danieli: i'm using syslinux here
2018-12-28 22:30:39 <sy> could it not like the ext4 boot directory?
2018-12-28 22:30:39 <danieli> ah, purely using syslinux
2018-12-28 22:30:49 <danieli> did you follow that guide to detail?
2018-12-28 22:30:51 <sy> to be honest seabios should be able to read it, right?
2018-12-28 22:30:54 <danieli> for all i know, it might be outdated
2018-12-28 22:31:07 <sy> no, i followed this as well https://www.linode.com/docs/tools-reference/custom-kernels-distros/install-alpine-linux-on-your-linode/
2018-12-28 22:31:09 <TBB> there's that annoying little requirement for ext4 boot with syslinux
2018-12-28 22:31:10 <sy> (this isn't linode)
2018-12-28 22:31:16 <sy> but i did this from a system rescue cd
2018-12-28 22:31:27 <TBB> something about 64-bit something
2018-12-28 22:33:48 <TBB> -O ^64bit at mkfs.ext4
2018-12-28 22:34:03 <sy> here, some detail about the setup http://ix.io/1x10
2018-12-28 22:34:15 <sy> i ran mkinitfs from within the filesystem
2018-12-28 22:35:53 <sy> is this message important? http://ix.io/1x11
2018-12-28 22:36:06 <sy> it says to ignore the errors in the wiki
2018-12-28 23:20:47 <sy> okay i've done the things again
2018-12-28 23:21:02 <sy> nop same thing
2018-12-28 23:24:05 <sy> TBB: mkfs.ext4 ^64bit /dev/vda1 ?
2018-12-28 23:26:24 <TBB> https://www.syslinux.org/wiki/index.php?title=Filesystem#ext
2018-12-28 23:26:31 <TBB> almost correct, yes
2018-12-28 23:31:29 <sy> alright, thanks
2018-12-28 23:31:31 <sy> here we go
2018-12-28 23:31:43 <sy> i always use ext2 for the boot
2018-12-28 23:31:53 <danieli> it's an old habit of mine too
2018-12-28 23:31:55 <sy> that also works, right?
2018-12-28 23:37:30 <TBB> I'm strictly a VFAT guy
2018-12-28 23:38:20 <TBB> reason? UEFI
2018-12-28 23:38:57 <TBB> fortunately you can even avoid that if you're willing to do some extra work...
2018-12-28 23:39:39 <sy> https://0x0.st/sn1T.png again...
2018-12-28 23:39:43 <sy> let's try ext2
2018-12-28 23:40:26 <danieli> TBB: that's only necessary on the ESD partition
2018-12-28 23:40:30 <danieli> mkfs.vfat -F32
2018-12-28 23:40:55 <danieli> /boot however can be whatever afaik
2018-12-28 23:41:05 <sy> ESP
2018-12-28 23:41:13 <sy> yeah, i have /boot on the root filesystem
2018-12-28 23:41:20 <sy> the efi stubs have the right bits to load it
2018-12-28 23:41:29 <TBB> depending on the bootloader and your other requirements, yes it can
2018-12-28 23:41:32 <sy> on my working pc anyway
2018-12-28 23:45:10 <danieli> oh yes, esp partition, I'm a bit tired
2018-12-28 23:45:22 <sy> http://ix.io/1x1m is this right?
2018-12-28 23:55:35 <sy> trying again with ext2...
2018-12-28 23:55:43 <sy> this isn't going to work...
2018-12-28 23:57:06 <sy> oh well
2018-12-28 23:59:38 <sy> TBB: any ideas?
2018-12-29 00:24:57 <sy> alright, i;m giving grub a go to see if it's any better
2018-12-29 00:25:58 <sy> i have managed to load a grub shell
2018-12-29 00:27:01 <sy> damn! missed out mkconfig!
2018-12-29 00:27:13 <sy> it's far too late for this
2018-12-29 00:27:55 <sy> i notice https://wiki.alpinelinux.org/wiki/Bootloaders mentions extlinux --install /boot however https://wiki.alpinelinux.org/wiki/LVM_on_LUKS does not
2018-12-29 00:28:11 <sy> or is this part of what setup-disk -m sys /mnt/ does?
2018-12-29 02:01:39 <sy> so in /boot/extlinux.conf,   APPEND root=/dev/mapper/void-root modules=sd-mod,usb-storage,ext3 cryptroot=/dev/vda2 cryptdm=lvm -- do I need to change modules?
2018-12-29 02:23:43 <sy> So I've added ext4 to the modules, along with virtio_blk, and now I get this https://0x0.st/sneQ.png
2018-12-29 02:23:49 <sy> definitive progress I'd say!
2018-12-29 02:24:04 <sy> obviously i'm still here so it's still bork :(
2018-12-29 03:34:49 <sy> derp! void-root is my void lvm name... simple typos
2018-12-29 03:35:49 <sy> i got it to boot but didn't remember to set the root password
2018-12-29 06:10:41 <daggs1> Greetings, is there a pkg for vanilla-virt kernel that contains dvb modules or I need to recompile it?
2018-12-29 09:17:55 <Dekedro> Hey, is Alpine moving from libressl -> openssl? https://git.alpinelinux.org/cgit/aports/commit/?id=50bb5f8772dffd19e400f299c195322ba21ca5d3 https://git.alpinelinux.org/cgit/aports/commit/?id=708b95a9dd94a00a59f6493687dab2a1b044b522
2018-12-29 09:23:46 <_ikke_> Dekedro: yes
2018-12-29 09:26:13 <Dekedro> I am not subscribed to mailing list, so might've missed something but what's the reasoning behind this change?
2018-12-29 09:27:55 <_ikke_> https://lists.alpinelinux.org/alpine-devel/6308.html
2018-12-29 09:28:31 <Dekedro> Thank you!
2018-12-29 09:34:58 <beli3ver> could somebody help me to get the network manager scan for networks, wifi => device not ready.
2018-12-29 09:35:17 <beli3ver> Linux pinetree 4.19.12-0-vanilla #1-Alpine SMP Thu Dec 27 15:30:13 UTC 2018 x86_64 Linux
2018-12-29 09:35:29 <beli3ver> 3: wlan0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN qlen 1000
2018-12-29 09:43:31 <beli3ver> product: QCA9565 / AR9565 Wireless Network Adapter
2018-12-29 09:44:52 <beli3ver> with iwlist scan I find all my networks
2018-12-29 09:46:45 <beli3ver> ok forget it I do ith without nm
2018-12-29 12:04:10 <mps> is it possible somehow to build same package with different build options in one pass, i.e. how to do that with APKBUILD build function? any hint or guide?
2018-12-29 12:06:27 <mps> have a libzip in context, it's build (cmake) allows to build shared or static but not both at the same build (make) time
2018-12-29 12:09:07 <_ikke_> mps: You probably want to have two seperate build directories
2018-12-29 12:11:26 <mps> _ikke_: sounds reasonable
2018-12-29 12:13:55 <mps> or maybe is better to create independent package, for examle libzip-static and put it in aports
2018-12-29 12:14:07 <daggs1> dump question, why I need more than 2.5 gb on my hdd to build a custom version of the kernel?
2018-12-29 12:28:10 <Shiz> mps: the precedent is to do it a single apkbuild and with separate build dirs yea
2018-12-29 12:28:22 <Shiz> daggs1: because the kernel and all its modules are big :p
2018-12-29 12:29:35 <danieli> yup, linux is a big boy
2018-12-29 12:30:51 <mps> Shiz: then need to have two 'cd $srcdir ;configure ; make' in build function? right?
2018-12-29 12:31:03 <Shiz> $builddir but yes
2018-12-29 12:31:03 <daggs1> Shiz: no that big, at least not more than 2.5 gb per version
2018-12-29 12:31:57 <Shiz> embargo:/usr/src/linux-4.9.9$ du -hcs .
2018-12-29 12:31:59 <Shiz> 751.7M.
2018-12-29 12:32:02 <mps> yes, different build dirs, and also in package function to install files in pkgdir
2018-12-29 12:32:04 <Shiz> this is a minimal build with almost no drivers/modules
2018-12-29 12:32:14 <Shiz> i can trivially see how it can reach up that high
2018-12-29 12:33:30 <daggs1> anyway, I guess I'll have to wait until my server will be ready... :(
2018-12-29 13:55:58 <mps> j8r (nick) from #crystal-lang asked me to help build libzip-dev with added static version of library, libzip.a
2018-12-29 13:57:28 <mps> I made some changes to APKBUILD and it worked, although i have feeling it is crude hack, so I'm asking if someone have time to review it
2018-12-29 13:57:38 <mps> it is here: http://tpaste.us/kZ6Z
2018-12-29 13:59:00 <mps> pkgrel should be bumped, ofc. comments and advices, please :)
2018-12-29 14:18:28 <_ikke_> mps: I think the *.a files usually go in the -dev subpackage
2018-12-29 14:19:06 <_ikke_> 0https://pkgs.alpinelinux.org/contents?file=*.a&path=&name=&branch=edge
2018-12-29 14:22:32 <mps> _ikke_: thanks for look at it. yes, changed it to put libzip.a to libzip-dev
2018-12-29 14:25:11 <mps> but libzip CMakeFiles.txt have flip switch to allow build only shared or static in one pass, not both. because that I add two pass build with copying file
2018-12-29 18:01:06 <mps> _ikke_: here is patch for libzip https://patchwork.alpinelinux.org/patch/4348/ if you have time to review
2018-12-29 19:33:47 <hatramatra> I'm wondering how configuration files from a package are handled at installation time and when package is being upgraded.
2018-12-29 19:35:18 <hatramatra> when a package is first time installed, some files go to /etc/
2018-12-29 19:35:48 <hatramatra> what happens to those files when that package is upgraded? what if they were modified in the meantime?
2018-12-29 19:36:54 <hatramatra> if anyone can give me some rtfm pointer, I would be grateful.
2018-12-29 19:37:28 <hatramatra> I'm trying to package https://github.com/arkanis/minidyndns
2018-12-29 19:49:50 <sy> if a file was modified and you try to uninstall the package it belongs to, apk won't touch it
2018-12-29 19:50:16 <sy> if you try to install a package but a file that it provides already exists, it will create the file with a certain extension
2018-12-29 19:50:22 <sy> apknew i think, or something to that effect
2018-12-29 19:51:02 <hatramatra> sy: great! thank you
2018-12-29 19:57:33 <danieli> iirc it'll install $file.apk-new
2018-12-30 00:39:35 <sy> hi
2018-12-30 00:39:54 <sy> is there an iso for aarch64?
2018-12-30 00:40:04 <sy> i'm going to try running it on rock64
2018-12-30 00:49:53 <milliardo> can't you unpack the rootfs in a flash card?
2018-12-30 05:23:58 <porjo> I have a docker image based off alpine container. I noticed the version of package 'youtube-dl' is not the latest, so trying to update (run apk update + apk upgrade youtube-dl) but doesn't work. What else should I try?
2018-12-30 05:28:50 <porjo> ah, I see the latest version is in 'edge' branch and not '3.8' yet
2018-12-30 10:09:44 <beli3ver> Hi
2018-12-30 10:10:42 <beli3ver> Has anyone here using jdownloader and get it to run under alpine?
2018-12-30 10:10:57 <beli3ver> Alpine is my desktop system
2018-12-30 10:15:10 <beli3ver> And teamviewer?
2018-12-30 13:50:53 <danieli> beli3ver: no teamviewer, it's proprietary and you're at mercy of teamviewer gmbh to build it for musl libc
2018-12-30 13:51:29 <danieli> jdownloader i've never touched, and it *looks* like parts of it are closed source, but there's no alpine package for it either way
2018-12-30 13:54:34 <beli3ver> Thanks
2018-12-30 13:59:51 <dalias> gcompat might work
2018-12-30 14:00:02 <beli3ver> What is gcompat
2018-12-30 14:00:39 <danieli> iirc the official "teamviewer for linux" is just an older teamviewer bundled with wine
2018-12-30 14:35:35 <beli3ver> danieli did you try anydesk
2018-12-30 14:35:52 <danieli> I don't need it, so no
2018-12-30 14:37:22 <_ikke_> I haven't managed to get teamviewer to work lately
2018-12-30 16:20:16 <ddevault> this may be relevant to your interests https://lists.sr.ht/~sircmpwn/sr.ht-dev/%3C20181230161951.GA4488%40homura.localdomain%3E
2018-12-31 02:23:13 <ceska> hi.
2018-12-31 02:25:28 <ceska> just quick question: I'm using Alpine-linux appliance on GNS3 trying to restart networking as deskribed in Alpine-L wiki site /etc/init.d/networking restart - and this doesnt work, I have mesage that comand not found. DO I mising something?
2018-12-31 03:58:48 <daurnimator> Is there a tool in the alpine repositories that allows me to run something with ambient capabilities?
2018-12-31 03:59:11 <daurnimator> capsh is too old to have --addamb ; even in edge
2018-12-31 07:33:36 <daggs1> Greetings, I've used the twiki to prepare a custom kernel. I've added the path to the repos file, how can I install the new kernel?
2018-12-31 10:26:04 <ceska> hi there. Happy new year to everybody. I've just quick question: I'm using Alpine-linux appliance on GNS3 trying to restart networking as deskribed in Alpine-L wiki site /etc/init.d/networking restart - and this doesnt work, I have mesage that comand not found. DO I mising something?
2018-12-31 10:29:59 <mps> ceska: which command is not found?
2018-12-31 10:32:46 <ceska>  this one -> /etc/init.d/networking restart
2018-12-31 10:33:19 <ceska> not counting my arrow pointer :))
2018-12-31 10:35:36 <ceska> https://wiki.alpinelinux.org/wiki/Configure_Networking#Activating_Changes_and_Testing_Connectivity
2018-12-31 10:36:05 <mps> you want to say that the file '/etc/init.d/networking' does not exists?
2018-12-31 10:41:57 <ceska> https://drive.google.com/open?id=1RjJSiuOfzqDNrABxgMHbMzotGDHKP0Qq
2018-12-31 10:43:03 <ceska> Fors exemple when you changing IP addresses on interface you wouls like to restart networking stack? Corect?
2018-12-31 10:43:32 <ceska> acording the Alpine Wiki this is the comand to restart network stack.
2018-12-31 10:43:58 <mps> yes, it is. but your question in not clear
2018-12-31 10:44:19 <ceska> but this doesnt work on GNS3 appliance , not sure if its same in normal Alpine L instalation or just on GNS3?
2018-12-31 10:44:32 <mps> what says: 'ls -l /etc/init.d/networking'
2018-12-31 10:46:37 <ceska> ls: /etc/init.d/networking: No such file or directory
2018-12-31 10:47:14 <ceska> ls -l /etc/init.d/
2018-12-31 10:47:19 <ceska> total 0
2018-12-31 10:47:23 <mps> aha, something is bad with installation.
2018-12-31 10:47:52 <ceska> because its just 1.99MB
2018-12-31 10:48:05 <ceska> this can be an issue.
2018-12-31 10:48:20 <mps> what is this GNS3?
2018-12-31 10:48:37 <mps> container image or what?
2018-12-31 10:48:43 <ceska> a....
2018-12-31 10:49:07 <ceska> in emulator
2018-12-31 10:49:14 <ceska> an emulator
2018-12-31 10:49:28 <ceska> build mainly for networkinging profesionals.
2018-12-31 10:50:07 <ceska> that can emulate random networking gear, al together and lot more. Mainly just for test labs.
2018-12-31 10:50:34 <mps> not sure if they omits networking init :)
2018-12-31 10:51:06 <ceska> that's what I;m thinking. but it works just doesnt have this function..
2018-12-31 10:51:55 <mps> question was: what is it, container or fully virtual system
2018-12-31 10:52:53 <ceska> sorry. yes. it runs everything in come kinde of containers.
2018-12-31 10:53:19 <ceska> some kind of containers.
2018-12-31 10:53:29 <ceska> dont realy trying to understand this..
2018-12-31 10:53:42 <mps> docker or lxc, or something else
2018-12-31 10:54:17 <ceska> mainly all runs in separate ubuntu based system, that is already as VM
2018-12-31 10:55:01 <ceska> it supports lots of platforms.
2018-12-31 10:55:23 <mps> and they have ready made Alpine install image which you installed following their guides?
2018-12-31 10:55:31 <ceska> Dynamips, QUEMU, Virtual Box, VMware Docker
2018-12-31 10:55:57 <ceska> mps: yes sure, you are right
2018-12-31 10:56:36 <ceska> https://gns3.com/marketplace/appliances
2018-12-31 10:56:59 <ceska> https://gns3.com/marketplace/appliance/alpine-linux-2
2018-12-31 10:57:27 <mps> then their image have errors if it doesn't have anything in '/etc/init.d/' or the install is broken
2018-12-31 10:58:22 <mps> huh, crappy site with a lot of javascript
2018-12-31 10:58:24 <ceska> not sure.. maybe.. I just want to ask someone if this comands runs well on normals installation. Which looks like it is, but not on GNS3
2018-12-31 10:59:07 <ceska> mps: sorry mate. :)) i;m standart GUI guy :)
2018-12-31 10:59:16 <mps> it works, if the admin don't screw someting. never had any problem with it
2018-12-31 11:00:52 <mps> and, i installed Alpine in containers, VMs, different real hardware. never had problem with /etc/init.d/networking
2018-12-31 11:03:41 <ceska> that;s what I need to know. Thanks for your time and patience. Happy new year :)
2018-12-31 11:06:41 <mps> no problem. you should ask them about your issue. thanks for your wishes about new year, but for me every day is new year :)
2018-12-31 11:07:36 <mps> btw, OT question: ceska means chech?
2018-12-31 11:11:32 <ceska> this is very modified version of normal name. but it can be understandable only for my language speaking individuals :)
2018-12-31 11:12:35 <ceska> this is from name Česlovas , then modification for slang use Čieska and then my modification as Ceska
2018-12-31 11:13:59 <mps> again we are OT: in my language ceska means exactly chech. ok, thanks for explanation
2018-12-31 11:14:58 <mps> there is #alpine-offtopic channel for such discussions, or some sort of
2018-12-31 11:29:40 <daggs1> how can I force apk to install a pkg from my local repository?
2018-12-31 11:35:44 <mps> doppo: apk add --allow-untrusted /path/to/apk
2018-12-31 11:37:00 <mps> sorry, that should be daggs1 not doppo
2018-12-31 11:37:38 <daggs1> mps: no need to add the file tree to the repos file?
2018-12-31 11:38:25 <mps> daggs1: you could add local repo to /etc/apk/repositories
2018-12-31 11:39:48 <mps> for example: file:///home/mps/packages/testing
2018-12-31 11:40:32 <daggs1> mps: I tried that but I was unable to install the custom pkg
2018-12-31 11:40:42 <mps> and you have to add your public key to /etc/apk/keys
2018-12-31 11:41:15 <daggs1> that might be right
2018-12-31 11:41:24 <daggs1> I'll try that
2018-12-31 11:42:29 <daggs1> one more question, incase the same pkg exists in my local repo and in the tree, how does apk know which to pick?
2018-12-31 11:43:46 <mps> you can 'pin' repos, for example '@local file:///home/mps/packages/testing' and then you 'apk add pkgname@local'
2018-12-31 11:48:32 <daggs1> add @local file:///home/mps/packages/testing to /etc/apk/repositories?
2018-12-31 11:49:15 <TBB> yup
2018-12-31 11:49:48 <daggs1> nice, thanks!
2018-12-31 11:49:48 <mps> daggs1: no, you should use path where your packages are stored
2018-12-31 11:50:07 <daggs1> mps: of course
2018-12-31 11:51:19 <mps> daggs1: sorry, but sometimes there are people who do some thing literally without thinkig :)
2018-12-31 12:06:54 <daggs1> mps: :)
2018-12-31 12:07:19 <daggs1> mps: on my main system I run gentoo, no room for not thinking there
2018-12-31 12:09:26 <mps> ofc, understood from your question that you are not novice, but in any case added 'disclaimer' :)
2018-12-31 12:10:51 <daggs1> (Y) thanks for the help
2018-12-31 12:11:20 <mps> np, you are welcome
2018-12-31 12:24:53 <TBB> no room for not thinking or, indeed, updating every week
2018-12-31 12:33:24 <mps> TBB: with gentoo you waste a lot of cpu clocks save some cpu clocks
2018-12-31 12:34:02 <mps> s/save/to save/
2018-12-31 12:35:38 <TBB> that part I don't really have a problem with, it's just that you're basically forced to run bleeding edge (forgot already what it was called in gentoo terminology) and updating it in that state is always hit and miss
2018-12-31 12:49:21 <danieli> yup, no file:// scheme for local files in /etc/apk/repositories
2018-12-31 13:00:02 <daggs1> danieli: not sure I follow
2018-12-31 13:00:27 <danieli> i don't think there should be a file:// scheme anyway
2018-12-31 13:00:42 <danieli> but none of it is f-ing documented so it's hard to know
2018-12-31 13:01:51 <mps> danieli: why not, it is standard scheme
2018-12-31 13:02:52 <danieli> i know it works without "file://"
2018-12-31 13:04:37 <mps> can't remember exactly, but when tried first time it didn't worked in my case, but file:// worked
2018-12-31 13:04:57 <daggs1> danieli: in that case, how I can force the local version over the remote version?
2018-12-31 13:05:10 <danieli> ... uh?
2018-12-31 13:05:23 <danieli> what exactly is the problem you're trying to solve?
2018-12-31 13:05:40 <mps> daggs1: pin
2018-12-31 13:06:07 <daggs1> mps: apk add pkgname@local?
2018-12-31 13:06:29 <mps> yes
2018-12-31 13:06:46 <danieli> oh well, i'll be pretty much absent in this channel for a bit, happy new year
2018-12-31 13:06:57 <daggs1> danieli: I'm running alpine over kvm, but I need some modules that aren't enabled in the config so I've built a kernel with them
2018-12-31 13:07:04 <daggs1> (Y)
2018-12-31 13:08:25 <mps> daggs1: there is article on wiki about pinnig on Alpine
2018-12-31 13:13:25 <mps> daggs1: https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning
2018-12-31 13:19:58 <daggs1> mps: thanks!
2018-12-31 21:32:22 <sy> crap. i installed 3.9
2018-12-31 21:32:28 <sy> any way to fix this?