2014-10-01 04:52:35 does NFS client work out of the box for anyone? 2014-10-01 04:53:32 rcp.statd keeps yelling, "failed to create RPC listeners, exiting" 2014-10-01 06:08:16 ahills: did you start rpcbind? 2014-10-01 06:08:46 i was able to mount ubuntu nfs server on an alpine nfs client 2014-10-01 06:13:53 this bash business starts to become reidiculous 2014-10-01 07:22:20 ncopa: have you tried the other way around? mount alpine-nfs on ubuntu? 2014-10-01 07:23:32 i tried it 2 weeks ago but couldnt get it mounted. probably my mistake though. 2014-10-01 08:16:03 yes, i was able to do that too 2014-10-01 08:16:10 but i dont remember if i only tried nfsv2 2014-10-01 08:16:22 i know for sure that idmap will not work 2014-10-01 08:16:44 my alpine is still 2.7 2014-10-01 08:16:54 and there was reports from upstream that fableds nfs-utils patch broke nfsv3 2014-10-01 08:17:12 the containers are 3.0. 2014-10-01 08:17:29 thats what i do on bld1 and bld2 too 2014-10-01 08:17:32 but nfs doesnt want to run in containers i remember. 2014-10-01 08:17:53 generally, i prefer moutning things on the host 2014-10-01 08:18:08 true 2014-10-01 08:18:17 so you could edit the config to mount it 2014-10-01 08:18:28 but i want to rm -rf 2.7v 2014-10-01 08:18:33 :) 2014-10-01 08:19:11 i have some openelec clients that mount nfs just fine, just my ubuntu box doesnt want to. 2014-10-01 08:19:55 i think openelec uses xbmc libnfs, so i guess its diff then regular nfs client. 2014-10-01 08:20:21 the todo list for nfs on alpine: * upstream rpcbind, libtirpc patches 2014-10-01 08:20:31 * fix fableds nfs-utils patch 2014-10-01 08:21:34 * fix the potensial mem leak in nfs-utils (addrinfo struct refactor) 2014-10-01 08:21:53 sounds like a daytime job... :| 2014-10-01 08:22:12 * add support for inotify instead of dnotify (or fallback to dnotify) 2014-10-01 08:23:09 i dont think any of those are super difficult 2014-10-01 08:23:21 but timeconsuming 2014-10-01 08:25:19 <^7heo_> life is incredibely time consuming. 2014-10-01 08:25:32 :) 2014-10-01 08:25:43 indeed 2014-10-01 08:27:47 sooner or later your time will run out. better use it carefully :) 2014-10-01 08:28:03 <^7heo_> You mean, partying and enjoying every possible moment? 2014-10-01 08:36:24 for instance 2014-10-01 08:36:32 and be kind 2014-10-01 08:36:46 im not very succesfull at that :) 2014-10-01 09:15:38 <^7heo_> you're kind to the computer scientists at least. 2014-10-01 09:17:26 rebooting brb 2014-10-01 09:22:45 <^7heo_> wb 2014-10-01 09:24:34 thanks 2014-10-01 11:12:09 Hehe. A 32M RAM Apline VM works perfectly 2014-10-01 11:12:33 openfbtd: :) 2014-10-01 11:12:43 i've got a 96MB laptop that is yet to boot alpine 2014-10-01 11:14:39 cool 2014-10-01 11:14:48 32MB ram sounds like its slow 2014-10-01 11:17:04 ncopa: filesystem cache & slab search times will be very low in 32MB :) 2014-10-01 11:17:30 damn on ia64 boxes it would fit the cpu cache... 2014-10-01 11:22:54 LOL 2014-10-01 11:23:12 thats kinda cool :) 2014-10-01 11:23:13 I've actually upped the RAM on that box to 128 in the end 2014-10-01 11:23:19 But it worked nicely on 32 2014-10-01 11:23:27 you run the entire OS in the CPU cache :D 2014-10-01 11:35:54 morning.. 2014-10-01 11:36:06 ncopa did you saw my msg from yesterday here on channel 2014-10-01 11:37:23 crow: morning 2014-10-01 11:37:36 about setup-alpine using v3.0 repo instead of edge 2014-10-01 11:37:37 ? 2014-10-01 11:38:05 yes 2014-10-01 11:38:05 i think i know why that happened 2014-10-01 11:38:07 i messed up 2014-10-01 11:38:21 i should have updated release version in alpine-base 2014-10-01 11:38:30 i did it bu forgot to git commit 2014-10-01 11:38:33 and just tagged 2014-10-01 11:38:36 i was thinking edge is realy edge and no need to change apk repository from /3.0/main to /edge/main 2014-10-01 11:38:46 yes, it was a mistake 2014-10-01 11:38:51 i think it will work next time 2014-10-01 11:38:59 oh great, well good that i testet it. 2014-10-01 11:39:10 yes, thanks! 2014-10-01 11:39:57 next time is planned to be when? as it is for an SD card running from ram.. so if iso have everything no need to mess with repository once booted 2014-10-01 11:40:34 whenever linux-3.14.20 is released 2014-10-01 11:40:40 and there is a grsecurity patch available 2014-10-01 11:41:59 ok then i will need to make iso myself, but question is what must be done for that? 2014-10-01 11:44:05 git clone git://git.alpinelinux.org/alpine-iso; cd alpine-iso; make iso 2014-10-01 11:44:40 you'll need install some deps: apk add git make squashfstools 2014-10-01 11:44:57 hm... dont remember if there were more 2014-10-01 11:44:59 maybe rsync 2014-10-01 11:46:05 alpine-sdk is enough for me 2014-10-01 11:46:09 but it pulls in lots more 2014-10-01 11:58:56 ncopa but how to get the edge repository and have edge packages on it? because this is what you did yesterday if i get it right, and packages and repo wasnt changed 2014-10-01 11:59:33 oh 2014-10-01 11:59:35 hm 2014-10-01 12:02:34 i just tagged a new iso 2014-10-01 12:02:40 lets see if that works 2014-10-01 12:03:17 ncopa: is it possible to build a stable iso on edge? 2014-10-01 12:03:35 how do you mean? 2014-10-01 12:04:08 if i change my reposotory to stable and then run make 2014-10-01 12:04:30 i think that should work yes 2014-10-01 12:06:29 crow: please try: http://nl.alpinelinux.org/alpine/edge/releases/x86_64/ 2014-10-01 12:06:32 new snapshot 2014-10-01 12:07:53 \o/ 2014-10-01 12:10:34 ncopa thank you let me try new iso 2014-10-01 12:17:20 ncopa: i guess repo could be an option in the makefile. 2014-10-01 12:21:23 anyone know if the XSA-108 xen patches are coming to 3.0.x already? 2014-10-01 12:53:29 ncopa looks good: fetch http://mirror.bpiotrowski.pl/alpine/edge/main/x86_64/APKINDEX.tar.gz 2014-10-01 12:53:50 v3.0.0-1612-gf1f076a [http://mirror.bpiotrowski.pl/alpine/edge/main] 2014-10-01 13:40:31 hehehe, XSA-108 2014-10-01 13:40:46 someone from alpine should be on the xen security list 2014-10-01 13:42:04 i'm currently stuck in enterprise linux security wonders, OVAL xml hacking 2014-10-01 13:42:24 i think stuff like that is really close to skynet 2014-10-01 13:46:12 Hello 2014-10-01 13:46:30 I don't have an opinion on this, I try to avoid getting involved into security stuff 2014-10-01 13:51:10 Where can I find the source code for the binaries included in the Alpine Linux Vanilla 3.0.4 x86_64 stable release? 2014-10-01 13:53:38 http://git.alpinelinux.org/cgit/aports/tree/?h=3.0-stable&id=v3.0.4 2014-10-01 13:54:00 Thanks! 2014-10-01 13:57:55 ncopa when doing setup-bootable i see this msg in dmesg: [ 2517.969332] FAT-fs (sdc1): utf8 is not a recommended IO charset for FAT filesystems, filesystem will be case sensitive! 2014-10-01 14:35:07 how hard is it to setup alpine to boot from a fake raid? 2014-10-01 14:35:16 Intel ESRT2 to be specific 2014-10-01 14:35:28 basically it uses LSI's megasr drivers 2014-10-01 18:32:49 ncopa: rpcbind claims success but the process is gone by the time I can check for it, no messages in /var/log/messages 2014-10-01 18:33:42 ahills: i found if i restart it manually 2014-10-01 18:33:46 and then restart all of nfs 2014-10-01 18:33:55 then i can exportfs -va 2014-10-01 18:34:02 it is weird 2014-10-01 18:34:03 ish 2014-10-01 18:34:11 rpcbind process never shows in ps 2014-10-01 18:34:36 do it manually, running /sbin/rpcbind (or w/e location it is) 2014-10-01 18:34:40 not via the init script 2014-10-01 18:34:45 PLEASE DONT ASK ME WHY 2014-10-01 18:35:02 i am trying to understand it for months now 2014-10-01 18:35:03 it just exits right away 2014-10-01 18:35:09 no error or anything 2014-10-01 18:35:18 return code zero 2014-10-01 18:35:40 yes. now try a /etc/init.d/nfs restart, and then rpcbind restart 2014-10-01 18:35:57 and then see if suddenly exportfs -va does no longer say function not implemented. 2014-10-01 18:36:14 for me, in that order, it works 2014-10-01 18:36:38 you are a god among men, darkfader 2014-10-01 18:36:41 (except in the cases where i updated the kernel and don't have the nfs modules left) 2014-10-01 18:36:46 did it work? 2014-10-01 18:36:52 yes 2014-10-01 18:36:54 HAHAHA 2014-10-01 18:36:57 now I just have to solf NFS-specific issues 2014-10-01 18:37:02 i wish i could use that as a reference. 2014-10-01 18:37:10 but noone would understand it 2014-10-01 18:37:10 you can 2014-10-01 18:37:25 have anyone who is interviewing you for employment send me an email 2014-10-01 18:37:29 haha 2014-10-01 18:37:29 and I will reply: 2014-10-01 18:37:34 darkfader is a god among men 2014-10-01 18:37:57 thanks :) 2014-10-01 18:38:06 now, WHY doesn't this work... 2014-10-01 18:38:11 and if you ever understand _what the fuck_ happens there, let me know. 2014-10-01 18:38:21 yeah, time for some research 2014-10-01 19:29:06 ugh, NFS of course makes my whole system unstable 2014-10-01 22:10:10 is there any particular reason one should stick with busybox over coreutils xor toybox 2014-10-01 22:41:37 busybox is better tested than toybox. in the long term toybox will almost surely be better, at least for most of the applets 2014-10-01 22:58:37 is its interface as beautifully consistent as busybox's? 2014-10-01 23:05:43 ? 2014-10-02 00:28:50 dalias: thanks 2014-10-02 00:29:32 dalias: is there a glibc package, I can't seem to find it anywhere 2014-10-02 00:43:50 no :) 2014-10-02 00:46:53 dalias: would it somewhat simple to build it myself and run it locally on alpine 2014-10-02 00:46:58 dalias: that paxtest program 2014-10-02 00:47:03 requires a.out.h 2014-10-02 00:47:09 which isn't appart of MUSL 2014-10-02 00:47:29 and I couldn't find uClibc either 2014-10-02 00:49:17 i'm not familiar with this program 2014-10-02 00:49:28 it migth just be a matter of dropping in a header that looks like what it expects 2014-10-02 00:50:08 dalias: possibly, but I wouldn't be surprised that the header is tightly couple to glibc 2014-10-02 00:50:57 it's not. it's just random junk distributed with glibc because it happened to be present on some legacy systems 2014-10-02 00:51:08 oh okay 2014-10-02 00:51:23 well next time I get a spare moment i'll check that out 2014-10-02 00:51:24 it might have some dependencies on glibc's headers in it (like names of types, internal use macros, etc.) 2014-10-02 00:51:44 but it doesn't actually have anything to do with glibc code 2014-10-02 00:51:48 (or with any code) 2014-10-02 00:54:13 dalias: maybe I should go ask in the #musl channel :P 2014-10-02 00:54:51 dalias: do you know if someone has an existing Dockerfile for building alpine linux with docker? 2014-10-02 01:07:40 Isn't a.out.h part of kernel? 2014-10-02 01:08:33 (ie /usr/include/linux should link to linux arch includes) 2014-10-02 01:10:42 no 2014-10-02 01:10:47 it's not linux/a.out.h 2014-10-02 01:10:49 just a.out.h 2014-10-02 01:15:32 dalias katlogic were would be a good place to start to contribute to the wiki? 2014-10-02 01:47:30 systmkor: heads up, anything related to musl or alpine is not really a community project. it is just show off grounds for new libc implementations that barely work( and represent one or two mens egos). probably better off just building what you need with buildroot. 2014-10-02 02:10:50 barfod: ? 2014-10-02 02:11:09 The dude be trollin' 2014-10-02 02:11:22 sounded like it... 2014-10-02 02:27:50 "trolling" is something you do to catch fish. I simply shared my experience. If you have a different experience, I am sure you are welcome to share it. 2014-10-02 02:31:36 barfod: huh? 2014-10-02 02:34:39 it seems like what i said was pretty clear, if it does not suit your needs or your find it contrary to what you believe, you are welcome to abandon the information 2014-10-02 02:36:08 so alpine is not a community project 2014-10-02 02:36:36 i noticed they gave you no reply about the wiki, and happens on #musl 2014-10-02 02:36:42 "it is just a show off grounds for new libc implementations"? 2014-10-02 02:37:19 what part about wiki did I not rply to 2014-10-02 02:38:23 yeah, sadly this is true. In their haste to prove themselves they have abandoned most foss projects to create a new project with a new libc that has wasted 1000's if not easily tens of thousands of man hours learning to deal with and adjusting our build systems to it 2014-10-02 02:39:08 "In their haste" who is their/they? 2014-10-02 02:39:11 "they" being some members of the foss community, alpine is ok, but it is closely tied into this type of ego displaying work, with little regard to the foundations all of us have already used and built 2014-10-02 02:39:47 I don't know what this ego displaying work you speak of is. 2014-10-02 02:39:47 basically musl is an epic waste of everyones time 2014-10-02 02:39:57 okay... 2014-10-02 02:40:08 i am not saying anything about any of the libc implementations 2014-10-02 02:40:19 I was asking what would be a good start on the wiki 2014-10-02 02:40:28 like is there a page that needs spelling and grammar checking? 2014-10-02 02:41:55 you notice none of them chimed in to help you? 2014-10-02 02:42:04 perhaps you thought they were having dinner 2014-10-02 02:42:21 well this is the experience you will likely continue to have 2014-10-02 02:42:32 barfod: well i don't know the GMT of all the members 2014-10-02 02:42:40 and usually like 90% of people in irc channels just idle 2014-10-02 02:42:48 pulled that out of my ass 2014-10-02 02:42:50 yeah, i am curious. i really hope they let you help with the editing 2014-10-02 02:42:58 well i can edit pages already 2014-10-02 02:43:01 its open for editing 2014-10-02 02:43:16 anyway, you think they are idling 2014-10-02 02:43:17 i am just curious if there was anything that someone wanted worked on 2014-10-02 02:43:19 but they are not 2014-10-02 02:43:25 sure okay 2014-10-02 02:43:28 just my advice, use buildroot 2014-10-02 02:43:33 take it or leave it 2014-10-02 02:43:38 ehhhhh 2014-10-02 02:43:47 yah communities take a lot of work to build 2014-10-02 02:43:54 but usually save you way more time than going "lone wolf" 2014-10-02 02:43:59 so i am good 2014-10-02 02:44:11 um yeah, and you actually have to answer people like you who are asking questions 2014-10-02 02:44:29 yup 2014-10-02 02:45:35 sorry I don't think I am understanding you well. 2014-10-02 02:45:39 I may need to eat more food. 2014-10-02 02:45:42 What is your point? 2014-10-02 02:46:20 i think most of the main alpine devs are in europe 2014-10-02 02:46:28 so they are probably still asleep 2014-10-02 02:47:03 come on, we are nerds. we sleep for 7 hours, drink some milk and check IRC 2014-10-02 02:47:19 there is also #alpine-devel (although most of the devs talk in both channels) 2014-10-02 02:47:44 pnutzh4x0r: well I am pacific standard time 2014-10-02 02:47:50 any recommended best time to be on? 2014-10-02 02:47:56 meh, ive submitted some patches for new packages and some existing ones and ncopa and others have been pretty responsive 2014-10-02 02:48:18 systmkor: from my experience early in the morning 2014-10-02 02:48:23 im in central time 2014-10-02 02:48:33 and i see a lot of activity between 5-7 2014-10-02 02:48:36 my time 2014-10-02 02:48:41 so 3-5 your time... 2014-10-02 02:49:06 heh 2014-10-02 02:49:36 15:00 - 17:00? 2014-10-02 02:49:39 PST 2014-10-02 02:49:47 systmkor: http://irclogger.com/.alpine-linux/2014-10-01 2014-10-02 02:49:54 thanks 2014-10-02 02:49:56 you can look at the logs 2014-10-02 02:50:02 didn't know that existed 2014-10-02 02:50:06 a lot of it happens at 3-5 in the morning 2014-10-02 02:50:55 wow, they have you guys conditioned to only expect responses 2 hours out of the day? 2014-10-02 02:50:57 pnutzh4x0r: 03:00-05:00 PST? 2014-10-02 02:51:06 systmkor: basically 2014-10-02 02:51:10 that is more ludicrous than anything i could have imagined 2014-10-02 02:51:26 well, u can also email the mailing list 2014-10-02 02:51:41 im not saying that is the only time, just the most active time 2014-10-02 02:52:04 but i am amazed by peoples loyalty sometimes. albeit unfounded 2014-10-02 02:52:29 yep, you got us barfod 2014-10-02 02:52:36 pnutzh4x0r: m'kay sure thing 2014-10-02 02:52:37 naively loyal and conditioned 2014-10-02 02:52:47 totally brainwashed 2014-10-02 02:53:10 pnutzh4x0r: Locutus of Borg, I am Borg 2014-10-02 02:53:13 :P 2014-10-02 02:53:33 no, _we_ are borg 2014-10-02 02:53:38 not I 2014-10-02 02:53:48 there is no we in borg 2014-10-02 02:53:52 all are borg 2014-10-02 02:53:54 so I 2014-10-02 02:54:00 ok ok 2014-10-02 02:54:05 i have no idea if that is actually true 2014-10-02 02:54:10 I am borg too 2014-10-02 02:54:15 but just what I remember from that remix song 2014-10-02 02:54:20 just thought i would share my perspective 2014-10-02 02:54:52 barfod: Why u love us no more? :( 2014-10-02 02:55:36 cause you are dicks? 2014-10-02 02:55:42 is this a trick question? 2014-10-02 02:56:29 barfod: found it https://soundcloud.com/chrisjeade/darkmateria-the-picard-song 2014-10-02 02:57:34 systmkor: oh you are awesome for this find. 2014-10-02 02:59:17 systmkor: That would totally deserve fexomated intro - https://hearthis.at/fexomat/nsfstartrek-vs-sumonefexomat-mashup-2012 2014-10-02 03:01:59 katlogic: lol wtf 2014-10-02 09:36:33 greetings. which apks do i have to add to be able to compile the current kernel? 2014-10-02 09:37:45 your "documentation" so far left out bc too 2014-10-02 09:39:52 now its looking for crc32c 2014-10-02 09:39:53 great 2014-10-02 09:40:01 which apk is that part of? 2014-10-02 09:40:27 someone here must compile the linux kernel just for fun. pls. 2014-10-02 09:40:32 which apks?? 2014-10-02 09:42:21 fou i think alpine-sdk 2014-10-02 09:43:00 fou sure someone here compile kernel, just not everyone is always online and on irc 2014-10-02 10:07:17 lol 2014-10-02 10:07:27 hit and run 2014-10-02 10:07:31 just how we like it :) 2014-10-02 10:08:22 <^7heo_> yeah, people are used to 24/7 indian support lines... 2014-10-02 10:08:32 <^7heo_> maybe we should try to find some indian people to help. 2014-10-02 10:10:58 ;) 2014-10-02 10:11:01 Hi 2014-10-02 10:13:14 I'm trying to setup a test environment to investigate (and try to solve) the openvswitch segfault problem (it happens whith all flavors; 32 bits or 64 bits; regulare kernel or vanilla one). 2014-10-02 10:13:34 as ncopa suggested, I added the openvswith-dbg package 2014-10-02 10:15:07 but I dunno where to find the crash reports and in dmesg, I see no difference with or without ovs-dbg :( 2014-10-02 10:16:27 hi 2014-10-02 10:16:39 did you set ulimit -c unlimited? 2014-10-02 10:17:06 no... but will do and try again 2014-10-02 10:17:21 else the core will not be written 2014-10-02 10:18:13 ok 2014-10-02 10:18:25 how do you run ovs? 2014-10-02 10:18:25 ls 2014-10-02 10:18:38 from init i guess? 2014-10-02 10:19:19 /etc/init.d/ovs-modules start;/etc/init.d/ovsdv-server start;/etc/init.d/ovs-vswitch start 2014-10-02 10:19:49 the crash occurs when "ovs-vsctl add-br mylan" 2014-10-02 10:20:18 s/ovsdv-server/ovsdb-server/ 2014-10-02 10:20:36 you can debug the core after it crashes, or you can start it directly inside gdb. 2014-10-02 10:20:45 ls 2014-10-02 10:20:50 but im maybe not the best debugger around :) 2014-10-02 10:20:52 sorry, wrong window 2014-10-02 10:21:16 <^7heo_> what's better than gdb? 2014-10-02 10:21:26 I do'nt see any core 2014-10-02 10:22:46 try: find / -name core 2014-10-02 10:29:24 ok. I rebooted after installing gdb. then first comman issued was ulimit -c unimited. now launching ovs-vswitchd shows an error stating that a core was dumped. and it is in / ... 2014-10-02 10:39:18 Jean-Scotch: get a backtrace 2014-10-02 10:39:42 I'm reading doc about gdb as I'm not used to it 2014-10-02 10:40:18 gdb executable core 2014-10-02 10:40:21 bt full 2014-10-02 10:41:39 you can log to a txt 2014-10-02 10:44:08 gdb /usr/sbin/ovs-vswitchd /core respond "warning: can't read pathname for load map: I/O error." 2014-10-02 10:45:02 bt full ends with complaints about missing symbol table for /lib/ld-musl-i386.so.1 2014-10-02 10:45:34 set pagination 0 2014-10-02 10:45:39 I guess I have to install musl-dbg also... 2014-10-02 10:45:41 set logging on 2014-10-02 10:45:46 bt full 2014-10-02 10:45:55 and you can paste gdb.txt 2014-10-02 10:46:37 ok 2014-10-02 10:46:39 yes you probably need that. 2014-10-02 10:47:49 Ok. I do the full start again with musl-dbg 2014-10-02 10:52:36 that's better. 2014-10-02 10:56:02 I find a "cannot access memory at address 0x50692664" which seems related to ovsthread_wrapper at lib/ovs-thread.c:325 then start at src/thread/pthread_create.c:122 then __clone at src/threab/i386/clone.s:36 2014-10-02 10:56:35 the last two ones are related to musl (ad showned the try before without musl-dbg) 2014-10-02 10:57:24 Jean-Scotch: let ncopa look at it. he should be back shortly. 2014-10-02 10:57:30 I read something on bug.a.o about the size of the stack on musl which was to small. Could it be stille related? 2014-10-02 10:57:40 clandmeter: ok 2014-10-02 10:57:59 Jean-Scotch: can you sprunge the trace? 2014-10-02 10:58:07 apk add sprunge 2014-10-02 10:58:16 sprunge < gdb.txt 2014-10-02 10:58:46 :) never heard about sprunge ... 2014-10-02 10:59:12 http://sprunge.us/MWYi 2014-10-02 10:59:43 ncopa: ^ 2014-10-02 11:00:21 or maybe fabled can take a look at it. he is our musl expert. 2014-10-02 11:02:50 Jean-Scotch, yes, wrong assumption about default stack size is possible, and cause crash like that. need to look at the backtrace first. 2014-10-02 11:03:19 ok 2014-10-02 11:03:38 you still have the core? 2014-10-02 11:03:49 yes 2014-10-02 11:04:14 what does 'disassemble' and 'info registers' say? 2014-10-02 11:04:33 i thought we fixed the stack size issue 2014-10-02 11:05:11 yes, there's a patch for that 2014-10-02 11:05:22 and we do apply it 2014-10-02 11:05:33 :) 2014-10-02 11:05:43 disassemble gives some assembler lines output 2014-10-02 11:06:04 can you please sprunge them 2014-10-02 11:07:51 http://sprunge.us/VWAa 2014-10-02 11:11:53 smells like a race 2014-10-02 11:15:25 <^7heo_> burnt tires? 2014-10-02 11:15:33 ncopa, sounds like too small stack still 2014-10-02 11:18:21 hmm 2014-10-02 11:20:51 Jean-Scotch, can you check also: 2014-10-02 11:20:59 p sizeof(struct flow_miss) 2014-10-02 11:21:05 p sizeof(struct upcall) 2014-10-02 11:22:35 response are: $1 = 1136 for the first one and $2 = 604 for the second one 2014-10-02 11:22:52 hmm 2014-10-02 11:22:57 that's not still too much 2014-10-02 11:23:10 and you have the latest openvswitch with the stack size fix? 2014-10-02 11:23:40 yes downloaded from repo just before trying 2014-10-02 11:26:14 DUH 2014-10-02 11:26:20 ncopa, the stack size patch is incorrect 2014-10-02 11:26:44 and i was too blind to see it when reviewing it 2014-10-02 11:27:06 you set up nice 'attr' but pthread_create() call never uses it 2014-10-02 11:27:27 you want me to fix it? 2014-10-02 11:40:23 inside the gdb you can type: l to get code 2014-10-02 11:42:11 ups didnt scrol down the log :) 2014-10-02 11:47:31 doh 2014-10-02 11:47:41 i can clean up my own mistakes 2014-10-02 11:48:53 Jean-Scotch, ncopa: i pushed new openvswitch 2014-10-02 11:48:56 please try that 2014-10-02 11:49:04 i havent pushed it yet 2014-10-02 11:49:05 i will 2014-10-02 11:49:12 ncopa, i fixe it 2014-10-02 11:49:24 http://tinyurl.com/mhodjp2 2014-10-02 11:49:28 ok. I will build a new qemu alpine from scratch and test it 2014-10-02 11:50:46 lol 2014-10-02 11:54:02 fabled: i will send the patch upstream: https://github.com/ncopa/ovs/commit/9ff82b9a7cabea6fb367c1f37aef740b9f02b701 2014-10-02 11:54:55 I will have proceeded the new test in a couple of minutes... 2014-10-02 12:04:30 still broken. but this time not when launcing ovs-vswitchd first time but when trying to add-br.... 2014-10-02 12:06:25 maybe the openvswitch I just download was not from the right repo ? I took the fastest from my place... 2014-10-02 12:07:42 mirror.bpiotrowski.pl ... 2014-10-02 12:15:05 Jean-Scotch, yes, those sync once in an hour or so. nl.alpinelinux.org is the master. 2014-10-02 12:15:09 I suppose I have to use nl.alpinelinux.org 2014-10-02 12:15:10 ok 2014-10-02 12:15:12 lol 2014-10-02 12:15:36 I do it again (it's running). stay tuned ;) 2014-10-02 12:25:55 great! no crash (at same step at least). ovs-vsctl add-br mylan works this time 2014-10-02 12:29:15 thank you very much for you time and attention ;) I wil now make more tests 2014-10-02 12:32:27 nice! 2014-10-02 12:32:42 Jean-Scotch: thanks alot for helping us clean up our mess :) 2014-10-02 12:38:30 I'm new to alpine (I discovered it on pipedot.org) and it perfectly fits my current need as I'm migrating my servers out of debian. I plan to run alpine on all my baremetals (but the ones used for building a CEPH cluster and even there maybe I will and run the CEPH daemons in qemu...) and migrate the actual servers to some qemu-kvm. So my boxes will run alpine+ssh+openvswitch+qemu+screen from USB (or PXE). So my next move will be 2014-10-02 12:41:02 I guess you may change the status of https://bugs.alpinelinux.org/issues/3309 to resolved now ;) I will still follow this topic very closely 2014-10-02 12:41:37 (it's my real name there) 2014-10-02 12:49:45 people are used to 24 hour american support lines, instead we have a bunch of of europeans taking over who only want to work 4 hour days 200 days a year 2014-10-02 12:50:22 lol again with that story... 2014-10-02 12:51:33 interesting, i am not the person insulting Indians work ethic at the expense of mocking someone else. 2014-10-02 12:51:53 nor would you find me doing that 2014-10-02 12:56:48 i am glad that you all are finding community in your hate for Indians and americans. please do not be surprised if we take a bit of offense -- being that you europeans have robbed and stole your way into power over the last 1000 years. maybe have a little humility and give back more than you took. 2014-10-02 13:00:31 morning barfod, i'm sorry that we cannot live up to your support expectations 2014-10-02 13:01:51 Guten Tag all 2014-10-02 13:03:50 good morning 2014-10-02 13:04:02 good afternoon 2014-10-02 13:04:38 ncopa: it is ok, i still like you. just a batter system has to be derived than simply expecting "no one to state the obvious" 2014-10-02 13:04:39 ncopa, maybe try starting http://bugs.alpinelinux.org/issues/3401 2014-10-02 13:05:27 I just read up and are you seriously getting pissy because musl guys don't answer you instantly on IRC? 2014-10-02 13:05:56 don't be a whiny child and start paying someone for support if you need it that quick 2014-10-02 13:06:18 vkrishn: i saw it, i dunno if i want spend time on that though 2014-10-02 13:06:21 no, i am getting upset about routine character denegration that its affiliates and maker engage in 2014-10-02 13:06:30 the table need be a norm but just guiding 2014-10-02 13:06:51 I can create on wiki if ok ? 2014-10-02 13:07:08 vkrishn: i think a wiki page for that is ok 2014-10-02 13:07:15 and devs are not bound by it 2014-10-02 13:07:24 ok will do 2014-10-02 13:07:38 yes, i dont want think just because we are here, we are available for help everyone 2014-10-02 13:07:47 heh 2014-10-02 13:07:49 haha 2014-10-02 13:08:01 you are supposed to serve your community, that is what foss is about 2014-10-02 13:08:21 he is not SUPPOSED to do anything 2014-10-02 13:08:27 he is volunteering his time for this 2014-10-02 13:08:33 yes, haha. we are all responisble for our actions 2014-10-02 13:09:02 if you want a different response make it someone's worthwhile financially 2014-10-02 13:09:11 barfod: i will not stop you serving the community, please do if you have the capacity 2014-10-02 13:09:17 mhmm, "volunteer" i think the attitude is making it clear that it is just a transitionary stage to help procur employment 2014-10-02 13:09:52 we just code, put our stuff out for free there for anyone to reuse 2014-10-02 13:10:08 but we dont give any promise for anything 2014-10-02 13:10:39 "please do if you have the capacity" 2014-10-02 13:11:04 we help each other whenever possible 2014-10-02 13:11:05 speaks with a forked tounge 2014-10-02 13:11:32 i suppose you get what you pay for 2014-10-02 13:11:35 a mans words should be clear, and his intention made known 2014-10-02 13:11:47 barfod, why not make a list of alpinelinux related tech queries and post on ML 2014-10-02 13:12:13 barfod: what is your intention? 2014-10-02 13:13:06 to shift the communities focus back to uclibc so we can actually get some work done 2014-10-02 13:14:28 feel free to fork 2014-10-02 13:14:50 so you want us to move back to uclibc? 2014-10-02 13:15:31 do you have any good reasons for why we should do that? 2014-10-02 13:16:04 i have been very happy with musl 2014-10-02 13:16:42 yes, i think that musl represents a counter-productive hacker community which shares exploits instead of reporting them. And that it is an inferior software that does not work with valgrind. and has created 1000's of build failures. and wasted tens of thousands of man hours 2014-10-02 13:17:40 ncopa: i believe that what you are happy with is the connection to rich felkers special club of insiders. i know it seems fun making fun of me and all the members of foss with them, but it is childish 2014-10-02 13:18:18 as long as he spoon feeds you information, and mocks your competitors -- yes musl would be the best choice for you 2014-10-02 13:18:19 what exploits are shared around in cabal fashion? 2014-10-02 13:18:40 if that was to subsde, i believe that you would quikly find yourself on the other side of the fence 2014-10-02 13:19:55 look like i missed the start of something here... 2014-10-02 13:21:15 ginjachris startet last night CET time :) 2014-10-02 13:22:40 eh, special club of insiders 2014-10-02 13:23:05 that was interesting 2014-10-02 13:23:08 well 2014-10-02 13:24:31 i dont have any more relation with musl devs than i have with any other FOSS project coders 2014-10-02 13:24:46 ncopa can we get invites? 2014-10-02 13:26:13 crow: sure, its just to write beautiful code and help debug nasty bugs 2014-10-02 13:26:17 and youre in :) 2014-10-02 13:26:21 ncopa: understood. well alpine is all good with me then, as long as you can control your community and keep them from mocking a concerned citizen 2014-10-02 13:26:43 wow, this is the first english speaking person for which I need a dictionary to understand what he says. 2014-10-02 13:27:22 ah see, back with ad hominem attacks 2014-10-02 13:28:07 oh barfod, don't take it personally. It's more a compliment, you obviously speak eloquently! 2014-10-02 13:28:15 also; referring to people in third person is exceptionally rude 2014-10-02 13:28:36 (I'm more a short and simple kind of guy, Getting old, no time for long words! :S) 2014-10-02 13:28:51 oh, no problem 2014-10-02 13:29:08 ginjachris: you are all good. 2014-10-02 13:30:43 Reminds me of that time Comrade Che Guevarra and his fellows were sat around a table discussing things, and it soon got out of hand because everyone had a different opinion on what to do and so there were arguments 2014-10-02 13:31:08 so Che took out a matchbox, wrote something on the back of it and handed it around. After that there were no more arguments 2014-10-02 13:31:21 i love che, awesome person, though he probably killed to many innocent people 2014-10-02 13:31:22 all he wrote on the back of it was "Unity" 2014-10-02 13:31:28 too* 2014-10-02 13:32:33 I think there may have been some innocents killed, once they took power. I think it was mostly peeps who didn't want to give up their land for the community, and also refused to leave the country 2014-10-02 13:32:42 hmm 2014-10-02 13:32:52 you know he went to Africa for a while as well? 2014-10-02 13:33:13 But i think before then (from what I remember) when he was fighting as a guerilla, he only went for legitimate military targets 2014-10-02 13:33:14 at least mostly 2014-10-02 13:33:14 i ended up not having a clear picture I think, even after some anthro / humanties classes 2014-10-02 13:33:34 oh yeah he did, didn't he? Congo I think 2014-10-02 13:33:50 i seem to remember he wasn't impressed with his recruits 2014-10-02 13:34:13 i remember him writing something about them closing their eyes and holding down the triger when shooting 2014-10-02 13:34:26 which wasn't what he was hoping for, obviously 2014-10-02 13:34:52 wow, very interesting. an anthropological insight 2014-10-02 13:35:57 ok i like you guys, and i support alpine 2014-10-02 13:36:14 the base image size is too big though 2014-10-02 13:36:49 :) 2014-10-02 13:36:57 lolz 2014-10-02 13:37:26 you didn't accidentally download Ubuntu did you? ;¬) 2014-10-02 13:37:31 heh 2014-10-02 13:37:56 well my distro base is 8MB. but tbh my real problem is more embarassing 2014-10-02 13:38:18 it is that the driver i am using for this wifi USB on debian is terrible 2014-10-02 13:38:32 and i have been to lazy/scared to try and fix it 2014-10-02 13:39:38 i should have tried to fix the driver and sent a patch to debian. instead i did other stuff, because it was easier 2014-10-02 13:40:18 but i will download alpine after my family goes to sleep. 250MB is still pretty reasonable 2014-10-02 13:45:28 i am going to tell you guys what the connection speed drops to, but you can not tell anyone ok? 2014-10-02 13:45:58 it sometimes goes down to 30 KBs 2014-10-02 13:46:47 oh try Alpine mini, it's like 80MB 2014-10-02 13:46:56 ohh ok! 2014-10-02 13:47:07 need net connection for package installations of course :) 2014-10-02 13:47:13 http://alpinelinux.org/downloads 2014-10-02 16:13:20 re there is still a bug in the post-upgrade script of openvswitch. This is not a blocking bug, so very low priority I guess ;) 2014-10-02 16:14:43 after reading the script, it seems to want to make some change in the /etc/openvswitch/conf.db but at the time when the scripts run, the file is in use. I think taht the daemon ovsdb-server is already started. 2014-10-02 16:20:27 I fill a bug report on bugs.a.o 2014-10-02 16:42:12 On the other hand, while running from HDD (sys mode), even if "rc-update add ovs-modules;rc-update add ovsdb-server;rc-update add ovs-vswitchd" was previously issued, the OVS daemons do not start at boot time. One has to start those by hand... 2014-10-02 16:45:46 https://bugs.alpinelinux.org/issues/3417 2014-10-02 17:00:00 also the command halt doesn't power off the machine. reboot is fine. 2014-10-02 17:08:45 try poweroff - also reading up on busybox helps 2014-10-02 17:10:25 the wiki advice to use halt... see http://wiki.alpinelinux.org/wiki/Alpine_Linux_Init_System. I will try your advice and change my habits ;) 2014-10-02 17:16:04 I added poweroff to that page 2014-10-02 17:21:20 ! 2014-10-02 17:21:23 <3 2014-10-02 20:04:18 <^7heo> say, why is there no "SDL_Window" type declared in the SDL headers in Alpine? 2014-10-02 20:32:00 <^7heo> okay, the SDL available is the 1.2 2014-10-03 08:08:13 Morning all! 2014-10-03 08:08:58 morning ginjachris 2014-10-03 08:14:32 I am glad it is Friday. It's been a rough week! 2014-10-03 08:18:17 i agree 2014-10-03 08:35:32 morning 2014-10-03 08:49:38 hello again! 2014-10-03 08:49:59 gotta love Windoze, the stability is so "good" that it makes me want to cry 2014-10-03 09:40:45 yay. I'v just managed to tune my chroot-building scripts to support alpine. Now to see if there's something like profiles & custom metapackages I could use 2014-10-03 09:55:53 wading through the wiki, I can't seem to find clear instructions on how to set up custom package repo 2014-10-03 10:01:09 is there an equivalent to emerge --depclean? 2014-10-03 11:02:36 what does --depclean 2014-10-03 11:02:37 ? 2014-10-03 11:03:53 ncopa: removes packages that are not explicitly added and are not depended on by anything anymore 2014-10-03 11:04:38 apk upgrade should do that i suppose 2014-10-03 11:04:54 ok 2014-10-03 11:05:11 it should add new deps and remove unneeded deps 2014-10-03 11:05:50 so how could I make custom repository with metapackages? (packages that don't install anything just depend on other packages) 2014-10-03 11:06:31 CcxCZ: i suppose you create a dir with APKBUILDs 2014-10-03 11:07:05 that only depends on other packages 2014-10-03 11:07:42 example meta APKBUILD: http://git.alpinelinux.org/cgit/aports/tree/main/alpine-desktop/APKBUILD 2014-10-03 11:07:45 I find it rather convenient to manage custom portage overlay where I write down installation templates like that and keep it versioned so I can easily add/remove packages I need on different kinds of setups... 2014-10-03 11:08:18 yes 2014-10-03 11:08:28 can apk use apkbuilds directly? or do I need a repository with built packages? 2014-10-03 11:08:45 you can apk add /path/to/file.apk 2014-10-03 11:09:28 or you can set REPODEST in /etc/abuild.conf 2014-10-03 11:09:43 and it will automatically create the repo when you abuild -r 2014-10-03 11:10:21 what do I need to install to have abuild? alpine-sdk? 2014-10-03 11:10:41 note that I don't really need to compile anything... 2014-10-03 11:10:44 yes 2014-10-03 11:11:03 it automatically adds build-base as makedepend 2014-10-03 11:11:42 might be an idea to have some option to disable that though 2014-10-03 11:15:10 does apk upgrade detect a change in the /path/to/file.apk I installed? 2014-10-03 11:15:53 well, probably shouldn't since there's version in the filename, duh 2014-10-03 11:16:12 so I'd need proper repo 2014-10-03 11:22:44 ncopa: btw, did you look into packaging runit? 2014-10-03 11:27:46 CcxCZ: http://git.alpinelinux.org/cgit/aports/tree/testing/runit/APKBUILD 2014-10-03 11:33:06 neat! That was about the last package I wanted so I could use it as the host vserver system. grub would also be nice but it's not all that necessary 2014-10-03 11:51:52 <^7heo> ncopa: ping 2014-10-03 11:52:42 <^7heo> I have installed the intel drivers, have a HD4xxx something, am part of the group "video", and when I glx I get: MESA-LOADER: could not create udev device for fd 5 2014-10-03 11:52:54 <^7heo> have you had that problem already? 2014-10-03 11:53:06 <^7heo> also, glxgears fails w/out root rights 2014-10-03 11:58:49 ^7heo: it could be sysfs restrictions 2014-10-03 11:59:43 <^7heo> ncopa: okay. I'll add myself to sys and try again 2014-10-03 11:59:53 that wont help 2014-10-03 12:00:10 <^7heo> ah 2014-10-03 12:00:14 <^7heo> good. 2014-10-03 12:00:25 <^7heo> because I did it already on another machine and it didn't help. 2014-10-03 12:00:31 i think you need to do something like: find /sys -type d -print0 | xargs chmod o+rx 2014-10-03 12:00:32 <^7heo> I am not familiar with sysfs 2014-10-03 12:00:39 <^7heo> oh 2014-10-03 12:00:50 <^7heo> that's kinda a hack, isn't it? 2014-10-03 12:00:52 yes 2014-10-03 12:00:59 <^7heo> what would be the proper way then? 2014-10-03 12:02:45 http://forums.grsecurity.net/viewtopic.php?f=3&t=4012 2014-10-03 12:03:05 i think the proper fix would be to have a special group for sysfs 2014-10-03 12:03:11 but seems like spender does not agree 2014-10-03 12:03:29 grsec devs says that sysfs restriction is not for desktop 2014-10-03 12:03:39 so we might need disable that :-/ 2014-10-03 12:03:48 i'd love to have a runtime switch for it 2014-10-03 12:06:12 <^7heo> ncopa: ah yeah I found that page already. 2014-10-03 12:06:23 <^7heo> sorry for asking what the proper fix would be then... 2014-10-03 12:06:36 no, the question is valid 2014-10-03 12:06:47 i dont know what the proper way to fix it is really 2014-10-03 12:06:59 <^7heo> yeah but your time is precious, and I shouldn't ask for info I can easily find/already found 2014-10-03 12:07:26 thanks 2014-10-03 12:07:33 <^7heo> well, I don't know enough about device nodes to have an opinion 2014-10-03 12:07:56 ncopa: how do I invoke abuild to build an .abuild into .apk? (or whole directory of them) 2014-10-03 12:08:04 <^7heo> afaik, a device node is an inode where the device can be accessed. 2014-10-03 12:08:11 <^7heo> that's all I know 2014-10-03 12:10:28 CcxCZ: abuild -r 2014-10-03 12:10:41 will build the .apk 2014-10-03 12:10:58 where do I input the directory to look for .abuilds? 2014-10-03 12:11:32 you run abuild from the directory where the APKBUILD is found 2014-10-03 12:11:43 ah ok 2014-10-03 12:11:45 if you want buidl an entire repo, then there is a tool called buildrepo 2014-10-03 12:12:47 <^7heo> we definitely need a bot to replace you, ncopa :P 2014-10-03 12:12:59 <^7heo> at least on the channel 2014-10-03 12:13:42 or a wiki :P 2014-10-03 12:15:10 <^7heo> there IS a wiki. 2014-10-03 12:15:21 <^7heo> http://wiki.alpinelinux.org/wiki/Main_Page 2014-10-03 12:15:43 I /know/. toungue in cheek and all that 2014-10-03 12:15:44 <^7heo> I don't really like the choice of mediawiki but it's "standard" like most of the shit around. 2014-10-03 12:15:54 <^7heo> toungue? 2014-10-03 12:16:09 <^7heo> and I prefer tongue in chick... 2014-10-03 12:16:14 <^7heo> ACTION so hides... 2014-10-03 12:27:47 oh you. you should go out and do some sexing 2014-10-03 12:31:08 <^7heo> naaaah, nevaaar! 2014-10-03 12:40:48 ACTION .o( wondering if ^7heo actually knows what does that word mean ) 2014-10-03 13:04:35 ncopa: is there the default contents of /etc/runlevels listed somewhere? as that doesn't get set up when doing the chroot install 2014-10-03 13:10:58 http://git.alpinelinux.org/cgit/mkinitfs/tree/initramfs-init.in#n641 2014-10-03 13:11:25 thanks 2014-10-03 13:16:20 ah cool 2014-10-03 13:16:30 i did that with looking at rc-update from another box 2014-10-03 13:17:15 http://confluence.wartungsfenster.de/display/Adminspace/Reinstallation+log notes, just in case you hit something i also hit 2014-10-03 13:18:26 (that was installing to the beginning of a disk, setting up raid and preserving the live data in further partions on the diskon the disk 2014-10-03 13:51:27 I'll put it on the wiki 2014-10-03 13:52:32 thanks 2014-10-03 13:58:24 http://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot#Entering_your_chroot 2014-10-03 14:01:05 CcxCZ: which is your host system? gentoo? 2014-10-03 14:02:05 yes. for the most part I just go through what I have installed in which runlevel there and mimic it in alpine, but it's better to have something cut&pasteable 2014-10-03 14:02:19 plus there are differences in the bb/mdev area 2014-10-03 14:02:33 have you tried lxc? 2014-10-03 14:02:35 its like chroot 2014-10-03 14:02:43 like a smarter chroot 2014-10-03 14:03:03 ACTION is long-time linux-vserver user 2014-10-03 14:03:55 nope, did not see the appeal of lxc yet. if they patch it up security-wise then perhaps 2014-10-03 14:04:15 i'd say lxc is an alternative to chroot 2014-10-03 14:04:31 CsxCZ: I think likewise 2014-10-03 14:04:47 well yeah, the namespacing features are nice 2014-10-03 14:05:19 and upstream lxc has a template for alpine 2014-10-03 14:05:31 so you should be able to lxc-create -t alpine .... 2014-10-03 14:06:08 I don't really have use for it. either I just want to run commands in a different fs for which chroot entirely suffices or I want something properly locked down... 2014-10-03 14:07:28 yeah, that could simplify matters. though I believe I already wrote something equivalent enough 2014-10-03 14:07:39 lxc-create -t alpine ... is more or less the same as: http://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot 2014-10-03 14:08:26 and should work from gentoo/ubuntu/whatever, as long as lxc is new enough 2014-10-03 14:09:01 i think you can also configure lxc.network = none which will use the network stack/ip addr etc from host 2014-10-03 14:10:09 well, I do some extra fun as to provide repository version timestamped builds, aiming to have equivalent to automatic building of stage4 for gentoo 2014-10-03 14:12:41 ie. produce a tarball of alpine-x86-edge-v141001-33-gd5ead43-vsguest-unbound to have image that I can readily deploy as a vserver to work as dnssec recursor 2014-10-03 14:13:27 got you 2014-10-03 14:15:42 basically just adapting my gentoo-building scripts. It does more stuff there as builds can fail, what's used for compiler matters, I want to make binpkgs, ... 2014-10-03 15:04:28 binpkgs.....sounds like the excess packaging that you get with everything you buy in a shop nowdays! :) 2014-10-03 15:19:57 terrible news.....I have to go. Have a great weekend all! 2014-10-03 15:31:58 i like that guy 2014-10-03 15:33:18 in a way that is legal in russia ofc, as a comrade 2014-10-03 15:42:05 thnk i am going to study math stuff today. i feel asleep and forgot to hit my download button for alpine. also my (1) user is filing like 4 bug reports every few hours so my hands are full 2014-10-03 15:46:44 i am just going to go ahead and get this alpine-mini-3.0.4-x86.iso now 2014-10-03 19:02:35 hmm, undefined reference to strtol_l 2014-10-03 19:02:48 do I need special LDFLAGS for musl? 2014-10-03 21:50:14 which package has the man pages for libc functions? 2014-10-03 22:34:36 ahills: man-pages 2014-10-03 22:37:33 thanks! 2014-10-04 01:28:50 Hi, I think it's a bug in alpine's mksh binary. Does anyone who uses mksh can try « mksh -c "typeset -n testvar" » and tell me if it segfault ? 2014-10-04 01:29:03 There is a strace http://sprunge.us/fjiI 2014-10-04 01:29:33 ps: i'm on i686 2014-10-04 01:47:15 Goodbox: i get a segfault too 2014-10-04 01:48:27 Ok ! I recently tried on a glibc box, it works as expected 2014-10-04 01:48:32 yeah segfault for me too on x86_64 2014-10-04 01:50:53 I can't even apk-add it on my 3.0.4 x86_64 box, is that normal? 2014-10-04 01:51:19 jzono1: i think u need the testing/edge repo 2014-10-04 01:51:57 or edge/testing 2014-10-04 01:52:05 it's in main, as far as I can tell 2014-10-04 01:52:22 oh... it is 2014-10-04 01:52:25 nvm 2014-10-04 01:52:39 builds without any obvious errors, too 2014-10-04 01:53:37 it's weird, there's several packages that are in the tree that I have trouble with adding :/ 2014-10-04 01:54:45 looks like it does something that grsec doesn't like, Goodbox 2014-10-04 01:54:47 [706003.144298] grsec: From 192.168.1.131: Segmentation fault occurred at (nil) in /usr/bin/mksh[mksh:30693] uid/euid:1000/1000 gid/egid:1000/1000, parent /bin/busybox[ash:30692] uid/euid:1000/1000 gid/egid:1000/1000 2014-10-04 01:54:51 [706003.144310] grsec: From 192.168.1.131: denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/bin/mksh[mksh:30693] uid/euid:1000/1000 gid/egid:1000/1000, parent /bin/busybox[ash:30692] uid/euid:1000/1000 gid/egid:1000/1000 2014-10-04 01:55:14 did you try it on a vanilla kernel yet? 2014-10-04 01:55:32 it segfaults with vanilla kernel for me 2014-10-04 01:59:31 hmm tried in root, it fails too 2014-10-04 02:20:31 su -c "prlimit --core=unlimited ; prlimit" # seem does not affect actual resources 2014-10-04 02:30:18 Goodbox: I think you need a line in /etc/security/limits.conf 2014-10-04 02:31:09 ok, setting core unlimited for soft and hard ? 2014-10-04 02:31:12 Goodbox: * - core unlimited 2014-10-04 02:32:50 Goodbox: you also need a new login session 2014-10-04 02:33:34 this mksh code is pretty gross 2014-10-04 02:34:35 Modified, reloged, prlimit doesn't care of /etc/security/limits.conf 2014-10-04 02:34:52 gross ? dunno this adjective i'll check :) 2014-10-04 02:35:41 oh ok, like raw or crude 2014-10-04 02:35:43 i mean it in the sense of disgusting, not large :p 2014-10-04 02:35:44 yeah 2014-10-04 02:36:50 i've got a bit of time to look at this 2014-10-04 02:37:38 i think i'll build it with debugging flag 2014-10-04 02:38:02 i've got in in gdb now 2014-10-04 02:38:06 *it 2014-10-04 02:40:57 for b main, run -c "typeset -n lol" blocks all termination signals. However i'm a noob I'm not very familiar with gdb sorry :s 2014-10-04 02:46:37 moskvax: do you got "Error accessing memory address 0x157896aa: I/O error. 2014-10-04 02:46:40 " 2014-10-04 02:46:45 ? 2014-10-04 02:53:28 Goodbox: no 2014-10-04 02:54:20 Goodbox: for me it segfaults on a null pointer 2014-10-04 02:54:37 Goodbox: mksh that is 2014-10-04 02:58:22 ok 2014-10-04 02:58:35 There is a rlimits.opt in the source archive 2014-10-04 02:58:58 which have "FN("coredump(blocks)", RLIMIT_CORE, 512 2014-10-04 02:58:59 " 2014-10-04 03:23:11 so... after updating linux-grsec to 3.14.19-r3, I can't boot my xen guests anymore 2014-10-04 03:24:03 this is very exciting; does anyone have an older version (say, 3.14.17-r0) lying around anywhere? 2014-10-04 04:15:52 hi, after upgrade to linux-grsec-3.14.19-r3 xen reboots in a loop after "Loading Dom0" - no way to get any error-msg. booting without xen works expected the erros that it could not load modules / conecting to xen. Is there any reported error known? btw. 3.14.17 was running perfect! 2014-10-04 04:43:52 xen_roger: same issue here :) 2014-10-04 04:45:19 any idea? there was a huge change http://git.alpinelinux.org/cgit/aports/commit/?h=3.0-stable&id=64a3e70e97b42fdfa8caf6a252cbc3bae82c8388 2014-10-04 04:45:38 no, I'm trying to get an older version up 2014-10-04 04:45:59 you could boot by installing linux-virt-grsec 2014-10-04 04:46:11 still version 3.14.17 ;) 2014-10-04 04:48:20 xen_roger: thanks for the heads-up! 2014-10-04 04:48:37 np - but still not a solution ... 2014-10-04 04:48:48 no, but it gets my systems booting, so I can go to bed 2014-10-04 04:49:11 i would i could also ... my wife begins to hate me ;) 2014-10-04 04:49:17 good night dude 2014-10-04 07:08:30 hmm, linux-vanilla can't execute busybox's /sbin/init? 2014-10-04 07:13:19 do I need a different busybox for non-grsec? 2014-10-04 07:21:23 Um. Works for me 2014-10-04 07:22:02 I have a linux-vanilla alpine install right here on a vm 2014-10-04 07:22:05 Works like a charm 2014-10-04 07:29:49 hmm 2014-10-04 07:29:51 do anything special? 2014-10-04 07:30:12 I mounted and chrooted from the install CD and just apk-added linux-vanilla 2014-10-04 07:32:20 I've installed it from the iso 2014-10-04 07:33:09 is there a simple way to install it from the iso? disabling chroot_caps and chrooting seems like a hack 2014-10-04 07:33:53 Um. setup-disk 2014-10-04 07:34:07 ... which installs linux-grsec 2014-10-04 07:34:19 Not if you do it from the vanilla iso 2014-10-04 07:34:29 oh, haha 2014-10-04 07:34:33 Yes, there is an iso with the vanilla kernel :) 2014-10-04 07:34:50 well, will you look at that... 2014-10-04 07:46:17 installer problems in xen on that image, too tired to debug, trying again tomorrow 2014-10-04 08:41:49 i have problems to restart with "service networking restart" 2014-10-04 08:41:57 it won't bring up interface lo back 2014-10-04 08:42:11 should i only "ifdown/ifup" use? 2014-10-04 08:51:58 hi, isn't service tool in openrc, try rc-service instead 2014-10-04 08:52:37 ugh ok there's a symlink 2014-10-04 08:53:34 sorry, there is /etc/network/interfaces which support debian-like networking syntax 2014-10-04 08:54:19 I guess it is readed by open-rc from multi-user stage 2014-10-04 08:56:21 Otherwise, you'll find the init script in /etc/init.d/networking 2014-10-04 08:57:57 you can take a look in /etc/rc.conf and modify if needed, rc_use="net." 2014-10-04 08:58:56 when I restart the iface lo, then I get always: "RTNETLINK answers: File exists" 2014-10-04 09:00:14 'k, did you tried to release your interface ? 2014-10-04 09:00:26 dhcpcd -r 2014-10-04 09:02:11 I don't use dhcp 2014-10-04 09:03:22 ah yeah for iface lo, sorry again i'm tired 2014-10-04 09:03:30 strange for lo 2014-10-04 09:06:19 what about your network device ? 2014-10-04 09:11:57 they are working great 2014-10-04 09:12:04 I have only br0 2014-10-04 09:12:10 its starting up and all okay 2014-10-04 09:12:16 just lo makes problems 2014-10-04 09:12:20 -s* 2014-10-04 09:23:38 Ok, i guess if there's no overlapping config, do you use alpine in a chroot or in para-virt environment ? 2014-10-04 09:24:12 If isn't, no ideas what append to your loopback network 2014-10-04 10:02:18 no 2014-10-04 10:02:24 alpine is my kvm host 2014-10-04 10:39:25 how can I help on this topic ? http://bugs.alpinelinux.org/issues/3241 2014-10-04 10:40:04 I would love to launch qemu instance on alpine hosts from a CEPH cluster 2014-10-04 10:43:01 maybe preparing a new package for rbd client? but the kernel module should be added also. I guess this is not done in a client package... 2014-10-04 12:09:09 I added http://bugs.alpinelinux.org/issues/3418 to request a new compile option for qemu (--enable-rbd) 2014-10-04 14:51:29 any news about linux-grsec in dom0 at xen -> crashing after "loading Dom0" 2014-10-04 15:05:22 hi people 2014-10-04 15:06:42 <^7heo> moin 2014-10-04 16:55:50 hello all, is it possible to upgrade 32 alpine to 64? 2014-10-04 16:56:07 without reinstalling 2014-10-04 16:56:28 It's as possible as with any other distro. 2014-10-04 16:56:54 I've converted my arch installs (two of them at the time) with the help of a live iso 2014-10-04 16:57:22 so, just switch 32 to 64 in /etc/apk/repositories and apk upgrade ? 2014-10-04 16:57:31 Take a list of explicitly installed packages, remove all packages from the system, install the 64 bit versions 2014-10-04 16:58:13 It's not esaier not is it faster than reinstalling 2014-10-04 16:58:16 nor* 2014-10-04 16:58:23 easier* 2014-10-04 16:58:37 Those typing skills man 2014-10-04 16:58:48 has anyone tried to build gc and run into a 'struct sigcontext' conflict between musl-dev and linux-headers? 2014-10-04 16:59:06 can't figure out why libgc is including both signal.h and asm/sigcontext.h 2014-10-04 16:59:55 k0r10n, you could also boot with a 64bit kernel 2014-10-04 17:00:05 And then install 64bit packages on top 2014-10-04 17:00:15 Should work 2014-10-04 17:00:53 Just don't blame me for it braking everything :D 2014-10-04 17:00:56 breaking* 2014-10-04 17:01:01 openfbtd, ok will try 2014-10-04 17:01:06 tyvm 2014-10-04 17:01:14 :))) 2014-10-04 17:03:05 k0r10n, https://wiki.archlinux.org/index.php/Migrating_Between_Architectures_Without_Reinstalling 2014-10-04 17:03:05 This may help 2014-10-04 17:03:05 It's for arch, but basic principles are the same 2014-10-04 17:03:29 openfbtd, great!! 2014-10-04 18:05:54 one more question, is there way to change network interface for lxc without recreating container? 2014-10-05 07:24:30 I'm curious why does alpine's hwdetect script run the /sys scan twice? 2014-10-05 07:30:16 Ah. Wait. Some devices may appear only after loading a module 2014-10-05 08:01:53 Also. Is the decision not to parse init= from the kernel cmdline in initrd deliberate? 2014-10-05 08:02:13 It makes testing alternatives a bit of a pain 2014-10-05 08:02:55 I've already rebuilt my initrd to allow for that, but I'm curious if that's a conscious decision or not 2014-10-05 09:51:06 hello, i have a problem with Alpine - it won't boot from my flash drive 2014-10-05 10:20:45 vitamin, that is a very vague description of your problem 2014-10-05 10:22:17 yes, but what can i say more? it just won't boot 2014-10-05 23:40:19 haha 2014-10-06 07:19:26 newest kernel 3.14.19-r3 crashes XEN at boot while loading Dom0 => http://pastebin.com/AQcCwjv5 --- Is there any fix around / already reported anyone this problem? 2014-10-06 08:56:30 xen_roger: first time i hear about it 2014-10-06 09:37:28 bah, alpine doesn't have an xcp-xapi/project khronos for it... 2014-10-06 09:37:40 wonder if khronos supports musl 2014-10-06 09:45:56 <^7heo_> damn 2014-10-06 09:46:03 <^7heo_> wpa_supplicant isn't starting at boot or default 2014-10-06 09:46:06 <^7heo_> but it starts manually... 2014-10-06 09:46:09 <^7heo_> what the heck? 2014-10-06 09:51:47 <^7heo_> what package to install to have the service "logger"? 2014-10-06 09:56:36 <^7heo_> okay, installing rsyslog makes wpa_supplicant work. 2014-10-06 09:56:44 <^7heo_> maybe rsyslog should be in the deps of wpa_supplicant? 2014-10-06 09:56:46 <^7heo_> ncopa: ^ 2014-10-06 09:58:07 shouldt busybox syslog do it? 2014-10-06 09:58:48 <^7heo_> I dunno, I had no syslog. 2014-10-06 09:59:08 <^7heo_> fresh install, from automated deployment 2014-10-06 09:59:13 <^7heo_> I can show you the packages I install 2014-10-06 09:59:25 <^7heo_> it's from 3.0.4 2014-10-06 09:59:32 should be provided by /etc/init.d/syslog from busybox-initscripts 2014-10-06 09:59:43 <^7heo_> okay lemme check if that service exists 2014-10-06 10:00:53 <^7heo_> you're right, syslog is started by that 2014-10-06 10:00:56 <^7heo_> lemme try to replicate. 2014-10-06 10:28:41 <^7heo_> okay, can't replicate the bug, seems like something went wrong during the install... 2014-10-06 10:28:45 <^7heo_> (the first one) 2014-10-06 14:31:58 <^7heo_> any idea what package to install to get the fonts on firefox? 2014-10-06 14:39:41 ^7heo_: Have you tried font-adobe-100dpi already? 2014-10-06 14:40:51 <^7heo_> jomat: I did install everything ttf-*, did the trick. 2014-10-06 14:40:56 apk search ttf 2014-10-06 14:41:05 Ah, ok :-) 2014-10-06 15:09:31 ncopa - using vanilla kernel seems to work ... sorry for late response. had to build iso first. so I think the changes in grsec are the reason. you could just boot a alpine-xen.iso build with 3.14.19-r3 to see the problem. Hardware / VM doesn´t matter. 2014-10-06 15:20:39 ncopa: so error is related to grsec :( 2014-10-06 15:42:34 ncopa: vanilla is also still linux-vanilla-3.14.17-r0 - linux-grsec-3.14.17-r0 also works fine! 2014-10-07 05:15:50 Does anyone know a way to delete from disk the modules that are not being used? 2014-10-07 05:20:20 jseutter: once you know which they are, they're in /lib/modules 2014-10-07 05:20:27 be careful you don't erase one you need ;) 2014-10-07 05:21:28 nod. I was thinking they'd be in initramfs so I deleted all the drivers from /lib/modules. The install still booted, but networking no longer worked and I reverted to a snapshot 2014-10-07 05:21:51 use lsmod to determine which ones are active 2014-10-07 05:22:47 I need some better bash fu to be able to delete modules that aren't in the lsmod listing 2014-10-07 05:25:02 lsmod | tail +2 | cut -d\ -f1 2014-10-07 05:25:06 that gives you a list of the active ones 2014-10-07 05:25:22 dump that in a file and use grep with find to figure out which ones aren't in the file 2014-10-07 05:25:50 find /lib/modules -type f -name '*.ko' -exec grep ... 2014-10-07 05:26:21 or maybe convert that list to full paths first, for safety 2014-10-07 05:27:18 or just build a custom kernel with build in drivers you need ... 2014-10-07 05:44:41 hm yeah, deleting the modules isn't working. perhaps a module is being unloaded after use earlier in the boot sequence 2014-10-07 06:21:47 I was something in net/. At any rate, I remove several MBs of unused modules 2014-10-07 06:22:00 s/I/it 2014-10-07 06:22:08 need sleep 2014-10-07 08:31:03 Hi 2014-10-07 08:31:16 Guys, I'm confused... 2014-10-07 08:31:29 can I use the gpg version from the repo ? 2014-10-07 08:31:45 it's in version 2.1 "BETA" 2014-10-07 08:32:08 moreover, there is no "gpg" binary, only "gpg2" is available 2014-10-07 08:43:35 maybe we ship gpg with gnupg1 package? 2014-10-07 08:43:49 the gnupg-2.0.x does not work with musl due to pth 2014-10-07 08:44:06 so we added gnupg1 for alpine v3.0 2014-10-07 08:44:27 gnupg-2.1 will work and the beta should work 2014-10-07 08:44:32 thats why we added it back 2014-10-07 09:31:08 Ah, yeah. Did not see the gnupg1 package 2014-10-07 09:54:12 mornign all 2014-10-07 09:54:17 morning* 2014-10-07 09:54:45 anyone with any experience of osmocom on alpine? 2014-10-07 10:40:48 osmocom on alpine anyone? 2014-10-07 10:40:53 stuck as a stuck thing 2014-10-07 11:17:22 maybe if you wrote what the exact problem is, someone would respond :] 2014-10-07 11:18:47 well, that's teh thing. 2014-10-07 11:19:54 I'm pretty new to all of this, and have a few problems with no real clue as to which direction to even look in :P 2014-10-07 11:20:21 so at the moment, I'm trying to build a gnu-arm-toolchain 2014-10-07 11:20:28 which claims it needs gcc 4.5.2 2014-10-07 11:20:34 I have gcc 4.8.X installed 2014-10-07 11:20:39 do you intend to to crosscompile? 2014-10-07 11:21:09 ncopa - apparently the build script will produce an arm-elf cross compiler 2014-10-07 11:21:15 we were able to bootstrap a native arm toolchain 2014-10-07 11:21:33 which means you can build for arm with it, but it will have to be done on an arm machine 2014-10-07 11:21:36 or in qemu 2014-10-07 11:21:41 both are slow... 2014-10-07 11:21:51 hmm. 2014-10-07 11:22:01 seems like I migh tbe on a hiding to nothing here 2014-10-07 11:22:47 http://bb.osmocom.org/trac/wiki/GnuArmToolchain 2014-10-07 11:22:56 is what I'm looking at 2014-10-07 11:24:24 but /why/ are you building arm toolchain? 2014-10-07 11:25:34 ah, it's supposed to run on a bare arm chip right? the baseband implementation is for some kinds of ICs found on older phones 2014-10-07 11:26:48 as I understand it, yes. 2014-10-07 11:26:57 and therefore you need bare-metal target, not the linux-gnu that is commonly provided 2014-10-07 11:28:08 perhaps you would be better served by http://crosstool-ng.org or similar toolchain builder 2014-10-07 11:30:03 in this case the gcc you use for building system packages and the one for bare-hw development can be completely different versions. but I don't see any advantage in building outdated gcc unless something in osmocom BB specifically requires it 2014-10-07 11:30:24 maybe. Before lookign at other options, I woul dlike to explore the osmocom method, as that may be the best chance for success. 2014-10-07 11:30:41 if I' go too far off-piste, it will go *really* wrong 2014-10-07 11:31:59 ok so your question is "I can't compile bare-metal target for gcc" - so where exactly does it fail, with which error message? 2014-10-07 11:32:05 plus, it's also exposing me to areas of linux I have absolutely no experience with, so the learning curve is pretty much vertical. 2014-10-07 11:32:20 so, in advance, thank you for your efforts CccCZ :P 2014-10-07 11:32:28 crosscompiling is non-trivial 2014-10-07 11:33:05 so I have downloaded gcc-4.5.2, bunutils 2.21.1a and newlib 1.19.0 2014-10-07 11:33:16 setting up a toolchain is also not that easy 2014-10-07 11:33:22 what is your ARM target? 2014-10-07 11:33:40 and run the script that osmocom have provided, gnu-arm-build.2.sh 2014-10-07 11:33:49 it runs quite happily for a while but then errors with: 2014-10-07 11:34:16 Building GCC requires GMP 4.2+, MPFR 2.3.1 and MPC 0.8.0+ 2014-10-07 11:34:19 crosscompiling is only non-trivial as far long as GNU software is involved :P http://aiju.de/rant/cross-compiling 2014-10-07 11:34:47 if I acquired the pacakges from the vendor, make sure I have installed both the libraries and header files 2014-10-07 11:35:00 Guest3557: try apk add gmp-dev mpfr-dev mpc-dev 2014-10-07 11:35:01 are the errors that the script ends with ^^ 2014-10-07 11:36:45 I have gmp-dev 5.1.3-r0, mpfr-dev-3.1.2-r0 2014-10-07 11:36:55 have also install the apk package mpc1-dev 2014-10-07 11:36:59 is that the correct one? 2014-10-07 11:38:59 yeh, I get an error with apk add mpc-dev 2014-10-07 11:39:14 can't find it on the package search page, only mpc1-dev 2014-10-07 11:40:12 ncopa: by "your arm target" you mean http://bb.osmocom.org/trac/attachment/wiki/GnuArmToolchain/gnu-arm-build.2.sh#L105 or the actual chip? 2014-10-07 11:41:02 arm target (as far I understand it atm) is arm-elf 2014-10-07 11:43:00 the supported chips are listed on http://bb.osmocom.org/trac/wiki/Hardware/Phones 2014-10-07 11:43:33 Guest3557: mpc1 seems to be the correct package; I guess it's named that due to name collision with "mpd client" 2014-10-07 11:43:59 mpc1-dev rather 2014-10-07 11:44:28 ok, ty. I was hoping it would be correct, as there was no other, but I shouldn't assume these things 2014-10-07 11:44:46 we have a supported phone, ready to plug in. 2014-10-07 11:45:18 as long the 0.8.0+ means any bigger version including major it should be the one 2014-10-07 11:45:18 waiting on mr. slowcoach over here to build the flippin' software. Which is going well, so far. Haha 2014-10-07 11:46:36 righto 2014-10-07 11:46:42 I was wondering the acutal chip it is supposed to run on 2014-10-07 11:47:05 we have arm stuff compiled and we have an arm build server 2014-10-07 11:47:18 so if we build it for x86, we'd get the arm package for free 2014-10-07 11:47:23 so, osmocom's advice that texinfo5 won't work is probably my next step, I think. I need to downgrade that at least. 2014-10-07 11:47:33 just to rule it out, if anything 2014-10-07 11:47:38 but we dont support older arm chips 2014-10-07 11:48:55 ncopa - is that re: the machine that you're building the cross-compiler on, or the chip that the cross-compiled code should run on? 2014-10-07 11:49:02 ncopa: bare-metal and linux-gnu targets are something completely different, as far as gcc is concerned afaik 2014-10-07 11:49:31 Guest3557: the chip that the crosscompiled code should run on 2014-10-07 11:50:02 http://bb.osmocom.org/trac/wiki/Hardware/Calypso is the primary target 2014-10-07 11:51:21 ok. I'm not sure which phone exactly it will be, but the fact that alpine doesn't support the chip will not matter, will it? 2014-10-07 11:51:43 building this cross-compiler is the first step to comile the osmocom open baseband software 2014-10-07 11:51:58 what i was thinking was that if it is compatible with the alpine arm build, you could just use those 2014-10-07 11:52:06 ah I see, apols. 2014-10-07 11:52:32 no linux can run on those chips, they are single-purpose control chips, not something you can just throw OS on 2014-10-07 11:52:50 aha, ok 2014-10-07 11:53:02 so you kinda need a crosscompiler 2014-10-07 11:55:37 there's the deepend, then there's building a crosscompiler for s first project 2014-10-07 11:55:42 still. Could be kinda fun. 2014-10-07 11:59:06 that's better. I feel more.... *identified* now 2014-10-07 12:08:24 I'm going to have to duck out for an hour or so soon. Thanks for the help so far, will speak in a bit 2014-10-07 12:35:00 <^7heo> Does alpine support the freescale i.MX6 quad? 2014-10-07 13:21:35 hello, i have a question regarding 'install to disk' document 2014-10-07 13:21:56 hi 2014-10-07 13:23:26 could somebody tell me please how this process looks? 2014-10-07 13:23:43 i mean, how my hdd will be partitioned? 2014-10-07 13:24:11 iirc, there will only be a single partition 2014-10-07 13:24:20 maybe two for /boot 2014-10-07 13:25:02 but you can partition it yourself, mount your partitions and then, run setup-disk 2014-10-07 13:26:55 ok, but how does it work actually? 2014-10-07 13:27:21 i need to run 'setup-alpine' first? 2014-10-07 13:27:35 or 'setup-disk'? 2014-10-07 13:28:07 setup-disk will install the files within a mountpoint 2014-10-07 13:32:12 i'm sorry but i don't understand it. does it mean that i need to run either 'setup-disk' or 'setup-alpine' to install it onto my hdd? 2014-10-07 13:33:22 vitamin: I believe you can run the scripts in either order 2014-10-07 13:34:43 setup-disk takes your current run-from-ram system and installs it to disk 2014-10-07 13:36:30 so it means that once i run 'setup-disk' i cannot use 'setup-alpine' as the second script will overwrite the previous one. am i correct? 2014-10-07 13:38:49 after running setup-disk, reboot 2014-10-07 13:39:03 then you can run any script you want and it will update the disk install 2014-10-07 13:39:55 sorry, I just remembered that setup-alpine now _includes_ the option to setup-disk 2014-10-07 13:40:00 ACTION never uses that option 2014-10-07 13:41:06 so it means that i can set up partitions manually within 'setup-alpine' script? 2014-10-07 13:41:38 setup-disk DO NOT partition/format your disks 2014-10-07 13:42:00 http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually 2014-10-07 13:42:40 vitamin: if you want to manually set up partitions, you should run setup-disk separately from setup-alpine 2014-10-07 13:42:42 oh, might be mistaken in fact ^^ 2014-10-07 13:43:03 within setup-alpine, use 'none' for disk, and it will not run setup-disk 2014-10-07 13:43:30 setup-alpine limits the options for setup-disk, so you must run them separately 2014-10-07 13:47:34 so to use custom partition scheme i need to run 'setup-disk' first, then reboot, and after that use 'setup-alpine' to finish my installation. is that right? 2014-10-07 13:47:50 no 2014-10-07 13:48:01 use fdisk to partition your disk 2014-10-07 13:48:03 format them 2014-10-07 13:48:11 mount them to /mnt 2014-10-07 13:48:37 and then, run "setup-disk /mnt" to install alpine using your custom scheme 2014-10-07 13:50:25 ok, now question about file system - can i choose any file system supported by linux for root partition? 2014-10-07 13:50:47 like what ? 2014-10-07 13:50:56 xfs for example 2014-10-07 13:51:00 totally 2014-10-07 13:51:17 my current install uses xfs / btrfs over LVM 2014-10-07 13:51:26 (and ext2 for /boot) 2014-10-07 13:51:30 works like a charm :) 2014-10-07 13:52:28 ok, thank you very much for this information. some people adviced my against choosing xfs or btrfs for root partition 2014-10-07 13:52:32 even btrfs for / with no seperate boot works 2014-10-07 13:52:40 yep 2014-10-07 13:53:18 you're limited by what linux supports, and by what your bootloader supports, the bootloader is most likely going to be your limiting factor (and a seperate /boot can work around that) 2014-10-07 13:53:33 ^ 2014-10-07 13:54:49 btw, is there another way to know which packages will be updated, other than using --simulate ? 2014-10-07 13:54:54 (just curious) 2014-10-07 13:55:41 i'm going to use partitions like: /boot (ext2), / (xfs), /home (xfs), and swap - is it good choice? 2014-10-07 13:55:57 it's not a bad one, actually :) 2014-10-07 13:58:03 it's going to be pretty standard desktop system - would you recommend any changes to this scheme, like btrfs instead of xfs or jfs maybe? 2014-10-07 13:58:48 btw, i will use standard sata2 hard drive - not ssd 2014-10-07 13:59:22 nah, should be fine 2014-10-07 13:59:47 but, keep in mind that alpine uses musl, not glibc 2014-10-07 14:00:05 so some programs you're used to might not copile on alpine 2014-10-07 14:00:28 vitamin: once again, you could run setup-alpine before setup-disk, or run setup-disk and reboot before running setup-alpine 2014-10-07 14:00:33 either way works 2014-10-07 14:00:57 i'm not used to any programs on linux - it's been over 10 years since i got linux installed:) 2014-10-07 14:01:34 tdtrask: ok, i got it:) 2014-10-07 14:01:45 ACTION would consider having a minimal regular distro tucked away in a chroot, for things that don't like musl 2014-10-07 14:02:55 I though about it fisrt 2014-10-07 14:03:13 But I like the idea to force myself to use well-coded alternatives :) 2014-10-07 14:03:34 thank you guys for help. i will try to install alpine today. 2014-10-07 14:03:43 have fun :) 2014-10-07 14:03:58 thank you:) 2014-10-07 14:05:36 ACTION wonders if alpine is the right solution for such user. 2014-10-07 14:10:01 yep 2014-10-07 14:10:09 I warned him, though 2014-10-07 14:14:34 maybe he'll be happier learning about unix than learning about gconf2-editor and similar :) 2014-10-07 14:27:44 you know we were talking about osmocom dependencies earlier? Well they said I could try building the libraries by hand 2014-10-07 14:27:48 would that work with alpine? 2014-10-07 14:27:58 for, say, libncurses5 and libmpfr4 2014-10-07 14:43:46 setup-disk is the final step in setup-alpine, right? 2014-10-07 15:16:54 NorfolkNClue: ncurses-dev should provide libncurses5(.9) 2014-10-07 15:17:25 CcxCz: chers dude 2014-10-07 15:17:28 Cheers* 2014-10-07 15:17:45 mpfr4 seems you'd have to build yourself, since I see only v3 in the repos 2014-10-07 15:17:47 I have thrown in the towel with Alpine - I'm going to build an ubuntu box 2014-10-07 15:18:08 you can just make debian chroot for building 2014-10-07 15:18:22 or debian* 2014-10-07 15:18:28 should be pretty straightforward. 2014-10-07 15:18:28 there's debootstrap in alpine repos 2014-10-07 15:18:38 going ot get it running on debian, see what's what, then tackle on alpine 2014-10-07 15:19:07 so just apk add debootstrap and use that to make a chroot with all necessary dependencies 2014-10-07 15:19:28 or perhaps use lxc and/or docker if you're into that 2014-10-07 15:23:57 careful with debian, if you use anything more recent than stable, you'll be getting systemd 2014-10-07 15:29:06 :) 2014-10-07 15:31:12 heh. I generally only use stable 2014-10-07 15:31:20 as I need it to be, erm, stable :P 2014-10-07 15:42:59 something went wrong during my first try 2014-10-07 15:43:21 after rebooting i get error: no such partition 2014-10-07 15:44:00 and when running 'setup-disk /mnt' i also get error 2014-10-07 15:44:22 syslinux missing: required by: world[syslinux] 2014-10-07 15:44:56 and sed: /etc/update-extlinux.conf : no such file or directory 2014-10-07 15:45:43 where did i make mistake? 2014-10-07 16:59:06 I did pull the aports from git but I cannot find for instance libuuid in it even if I found it on http://nl.alpinelinux.org/alpine/edge/main/x86_64/ (I'm trying to prepare a ceph package) 2014-10-07 16:59:42 are the repo and the aport git tree different ? 2014-10-07 17:20:49 Jean-Scotch: http://git.alpinelinux.org/cgit/aports/diff/main/e2fsprogs/APKBUILD?h=1.9&id=642737c0f027123a566abb8c7c6058aa346ac0aa 2014-10-07 18:17:00 xen_roger: Does it mean that the repo and the git are out of sync? because libuuid is removed in the package toyr linked but is still in the repo. I can apk add it bur when trying to lanch a abuilr -r xhich depend on it, it fails miserabily 2014-10-07 20:29:21 Jean-Scotch: sry for late response ... look at here http://git.alpinelinux.org/cgit/aports/tree/main/util-linux/APKBUILD?id=80f5ac62c11a29201922a51d941e60b4dfa104c2 => subpackage libuuid 2014-10-07 20:45:16 Jean-Scotch: building util-linux relsuts in libuuid-2.24.2-r4.apk 2014-10-07 21:53:38 strange, because inside the build() we find --disable-uuidd ... 2014-10-07 22:06:00 Jean-Scotch: --disable-uuidd do not build the uuid daemon 2014-10-07 22:06:51 i guess it doesnt effect libuuid 2014-10-07 22:07:13 ha! ok. 2014-10-07 22:07:38 but it does not solve my problem. I guess I still have a lot to learn about builing a package for alpine 2014-10-07 22:08:19 Jean-Scotch: what is the problem you are having? 2014-10-07 22:08:37 I would like to build qemu with RBD support, 2014-10-07 22:09:22 SO I'm trying to build ceph ant from them to have some subpackage with librbd and librados to be able to build qemu... 2014-10-07 22:10:01 s/ant from them/and from there/ 2014-10-07 22:12:55 unfortunately ceph needs a lot of different elements to build and a lot of them do net seems to be already packaged for alpine. So I'm trying to identify them all and if possible prepare package for the missing ones 2014-10-07 22:13:46 currently you are trying ceph and afterwards qemu? 2014-10-07 22:13:59 yes 2014-10-07 22:14:33 when adding --enable-rbd in qemu; it complains for missing librbd and librados. Both are part of ceph. 2014-10-07 22:24:06 Is it a easy way to know in wich package a specific tool belongs? I tried with the package web front-end but with no results 2014-10-07 22:25:10 I'm trying to itarate building ceph, each try blocking for some missing file or dependencies. Then add those to my APKBUIL file and run it again... 2014-10-07 22:28:45 Jean-Scotch: that sounds ok. you can also take a look at other distro's how they packaged it. 2014-10-07 22:29:27 I did ;) I'm following the arch linux package. but some dependencies are missing in alpine 2014-10-07 22:33:03 yeah, some pkg's are a b*tch 2014-10-07 22:37:47 what I find strage is that the package libuuid exists but when lanching the "abuild -r" for ceph it complains that libuuid is missing even if it is the first in depends... 2014-10-07 22:38:29 I suppose I need someting like libuuid-dev which does not exist 2014-10-07 22:40:35 ou need util-linux-dev 2014-10-07 22:40:50 it will have the headers for libuuid 2014-10-07 22:41:38 ok. let's try that. 2014-10-07 22:46:33 clandmeter: is right, util-linux-dev has header and .a should solve your problem 2014-10-07 22:46:47 bingo. for this one. 2014-10-07 22:47:12 wasn´t that hard ;) 2014-10-07 22:47:54 not really. but the difficulty is to find the right package when ./configure complains that something is missing 2014-10-07 23:57:57 for instance I do not find libresolv even if resolv.h is present. From what I read, this should be part of the libc. 2014-10-08 00:08:04 good night, I'm falling asleep... 2014-10-08 05:58:56 any progress on linux-grsec in xen? 2014-10-08 06:06:18 morning 2014-10-08 06:06:32 ahills: yes, i got pipacs attention and he has a proposed fix 2014-10-08 06:06:37 which I am testing now 2014-10-08 06:06:39 awesome! 2014-10-08 06:06:49 I would love to test it as well if you need an extra hand, etc 2014-10-08 06:07:06 we also discovered a config difference with vanilla kernel 2014-10-08 06:07:24 similar config difference to linux-virt-grsec, or no? 2014-10-08 06:07:40 dunno yet 2014-10-08 06:08:04 is the latter recommended for xen guests? 2014-10-08 06:08:08 (generally) 2014-10-08 07:40:54 ahills: yes, i think the main motivator for linux-virt-grsec is xen guests 2014-10-08 07:41:02 ok 2014-10-08 07:41:03 amazon cloud etc 2014-10-08 07:41:41 running xen here, if I install a guest and, at the end of setup-alpine, mount my new / and /boot, then apk -p /mnt add linux-virt-grsec 2014-10-08 07:41:52 the kernel panics because it can't find /init 2014-10-08 07:47:06 \o/ 2014-10-08 07:47:20 the patch i got from pipacs seems to fix the issue 2014-10-08 07:47:42 now THAT is some great news! 2014-10-08 07:48:25 did we have an issue for it on bugs.a.o? 2014-10-08 07:48:46 nope 2014-10-08 07:48:50 sorry, I don't know 2014-10-08 07:48:58 did not do my due diligence in reporting the issue 2014-10-08 07:50:07 might have been good to have it for history 2014-10-08 07:50:23 (ncopa: last night i tried to prepare a bootable alpine install usb stick on windows without any success :( 2014-10-08 07:50:37 kocka: unetbootin? 2014-10-08 07:50:53 yep 2014-10-08 07:51:26 bbl 2014-10-08 07:51:28 I had the same issue on one PC but not on another 2014-10-08 07:51:41 :( 2014-10-08 07:51:51 hum, weird, i'm running with grsec in domU, is that a recent breakage? 2014-10-08 07:52:17 katlogic: yes, hold off on updating linux-grsec 2014-10-08 07:52:26 it was just -r3 of the most recent 2014-10-08 07:52:28 I think .19 2014-10-08 07:52:33 ah 2014-10-08 07:52:51 ncopa: you got a patch? sounds great. did that fix our problem ? 2014-10-08 07:53:11 xen_roger: pipacs> who's the original reporter i should credit? 2014-10-08 07:53:25 xen_roger: yes, it did fix the problem 2014-10-08 07:53:40 I think its me ... but you could get it also ;) 2014-10-08 07:53:42 i told him to credit xen_roger 2014-10-08 07:53:49 ok 2014-10-08 07:53:58 but i dunno if i should give full real name? 2014-10-08 07:54:01 i suppose xen_roger works 2014-10-08 07:54:28 it´s enough I just reported it 2014-10-08 07:55:09 are you going to commit the patch soon ? 2014-10-08 07:56:39 the plan is: rebuild 3rd party modes (running) now 2014-10-08 07:56:46 running as we speak 2014-10-08 07:56:48 hi 2014-10-08 07:56:54 push to edge 2014-10-08 07:56:57 hi 2014-10-08 07:57:03 verify it works from edge 2014-10-08 07:57:11 ncopa: sounds great! 2014-10-08 07:57:14 push to 3.0-stable 2014-10-08 07:57:19 and tag v3.0.5 2014-10-08 07:57:51 does it make sense to try apline as a system for a desktop ? 2014-10-08 07:58:05 wotwot: i use it with xfce 2014-10-08 07:58:15 ncopa, sweet 2014-10-08 07:58:26 it works ok, unless you need closed source stuff 2014-10-08 07:58:41 eg adobe flash, nvidia/ati closed source drivers 2014-10-08 07:58:53 not so much, i want to switch from debian to something without systemd 2014-10-08 07:59:05 skype does not work 2014-10-08 07:59:12 nor spotify 2014-10-08 07:59:12 no prob 2014-10-08 07:59:21 wotwot: you will probably run into a few things that can't be built without libc, but generally alpine is a very complete distro 2014-10-08 07:59:28 wotwot: maybe gennto / funtoo + openrc is an option 2014-10-08 07:59:48 yeah that might work too. what about audio/jack ? 2014-10-08 07:59:50 *gentoo 2014-10-08 08:00:22 i dont think pulseaudio will work with alpine 2014-10-08 08:00:25 but i havent tried 2014-10-08 08:00:27 :D 2014-10-08 08:00:40 i have tried rosegarden 2014-10-08 08:00:45 hydrogen 2014-10-08 08:00:46 i usually remove pulseaudio in favour of jackd 2014-10-08 08:00:59 i think i build jackd, but i dont know if i have tested it 2014-10-08 08:01:13 in general, alpine is not as polished as other distros when it comes to desktop 2014-10-08 08:01:47 i'll give a try, maybe its just what i want 2014-10-08 08:02:03 we dont have chrome/chromium atm 2014-10-08 08:02:22 i'm mostly on firefox 2014-10-08 08:03:06 what i use mostly is: hexchat, firefox, xfce-terminal, claws-mail, virt-manager+qemu/kvm 2014-10-08 08:03:38 for me, terminals/ssh, firefox, thunderbird 2014-10-08 08:03:48 we dont have thunderbird :-( 2014-10-08 08:03:55 hard to build? 2014-10-08 08:03:57 oh :) 2014-10-08 08:03:59 oh right, mozilla 2014-10-08 08:04:15 last i looked at it it could not build against shared xulrunner 2014-10-08 08:04:21 which was the dealbreaker at that time 2014-10-08 08:04:28 ok 2014-10-08 08:04:36 i got a box at work i wanted to try for xen/docker, maybe i'll try alpine for that first 2014-10-08 08:05:05 works great as lxcontainers 2014-10-08 08:05:18 due to OS overhead beeing minimal 2014-10-08 08:05:29 also great on xen 2014-10-08 08:05:51 nothing is faster ... using it for networking stuff 2014-10-08 08:06:49 cool, gotta go to work, ttyl and ty 2014-10-08 08:07:22 xen_roger: pushed fixed kernel to edge/main 2014-10-08 08:07:31 ok 2014-10-08 08:07:38 + the vanilla description fix 2014-10-08 08:07:44 i'll go look for coffe 2014-10-08 08:10:43 possible push to mirrors today? 2014-10-08 08:12:31 yup 2014-10-08 09:02:49 Hi. Does libresolv exists in some package? My current iteration for trying to build a ceph package fails because "no resolv library found". I already search for more than 2 hours where I could find this lib :( The test in ./configure is looking to link res_nquery 2014-10-08 09:03:27 Mu real goal is to build qemu with RBD support and that needs librbd and librados which are part of ceph 2014-10-08 09:04:37 Jean-Scotch: i think thats a part of libc, which is not fully implemented in musl 2014-10-08 09:05:04 the libresolv stuff is generally a bit ugly so i dont think musl wanted implement it fully 2014-10-08 09:05:05 from what I read, you are right. It's part of libc 2014-10-08 09:05:28 but its partiallly implemented 2014-10-08 09:05:45 and it might be its not that far off from making it work 2014-10-08 09:05:53 ACTION found yesterday that mosh (ssh for lossy connections) shits on unix resolvers and ignores /etc/hosts. that was disgusting 2014-10-08 09:06:29 darkfader: is that alpine/musl only? or all? 2014-10-08 09:06:41 ncopa: was freebsd yesterday, so i suppose "all" 2014-10-08 09:06:47 i'll recheck, good point 2014-10-08 09:07:16 ok. So I will not be able to build ceph without going into some source code, I guess... 2014-10-08 09:07:45 Jean-Scotch: you might wanna ask the musl people about that resolver function 2014-10-08 09:08:37 I tried to go to the musl web site for the last few days but always got a nginx error... 2014-10-08 09:08:51 oh they didnt fix musl website yet 2014-10-08 09:08:55 thats bad 2014-10-08 09:09:47 talking about http://www.musl-libc.org/ ? 2014-10-08 09:09:50 Jean-Scotch: #3241 2014-10-08 09:10:02 xen_roger: the wiki i suppose 2014-10-08 09:10:06 ncopa: that's mine ;) 2014-10-08 09:10:55 Jean-Scotch: you can post the work-in-progress APKBUILD to #3241 if you want 2014-10-08 09:11:04 the "rbd" util is enough for xen tinkerers like me, and kvm users also need qemu-rbd, which possibly needs the backend library librados 2014-10-08 09:11:35 librados is from ceph-common? 2014-10-08 09:11:42 or is it separate lib? 2014-10-08 09:11:55 on debia it is a separate package 2014-10-08 09:12:14 but comes from same source package? 2014-10-08 09:12:19 yes 2014-10-08 09:12:21 ok 2014-10-08 09:12:39 we probably need get that ceph package built then 2014-10-08 09:12:55 Jean-Scotch: can you please sprunge < APKBUILD 2014-10-08 09:12:56 librados and librbd are subpackages of ceph 2014-10-08 09:13:15 i can try look at it 2014-10-08 09:13:37 http://sprunge.us/KgZF 2014-10-08 09:13:56 but it's far from being ok... 2014-10-08 09:14:29 ncopa, this is my ceph APKBUILD so far: http://sprunge.us/KaOT (in case you need it) 2014-10-08 09:15:35 thanks 2014-10-08 09:16:35 ah, rnalrd sed out the libresolv issue from configure script 2014-10-08 09:17:08 I was seeing just that ;) 2014-10-08 09:17:16 yes, but does not build yet :) 2014-10-08 09:18:01 i also prefer do the autoreconf thingy in prepare() 2014-10-08 09:20:56 the sed line is wrong too 2014-10-08 09:21:08 conftest.c:36:1: error: too many arguments to function 'res_query' 2014-10-08 09:21:09 res_query(0, 0, 0, 0, 0, 0); 2014-10-08 09:21:09 ^ 2014-10-08 09:21:20 and since res_nquery is used, it might need refactor 2014-10-08 09:24:26 Morning all 2014-10-08 09:24:34 well, it here here, anyway :P 2014-10-08 09:24:39 it is here* 2014-10-08 09:24:58 sheesh. I'm pretty fupping wound up dealing with insurance companies. apols 2014-10-08 09:27:56 this is what they use it for: https://github.com/ceph/ceph/blob/master/src/rgw/rgw_resolve.cc 2014-10-08 09:28:03 i think i'll ask in #musl 2014-10-08 09:37:29 brb 2014-10-08 09:49:12 I am pushing xen dom0 fix to v3.0 now 2014-10-08 09:49:25 will most likely do release 3.0.5 after lunch 2014-10-08 09:49:25 woohooo! 2014-10-08 10:01:09 ncopa: what toolstacks does alpine have? 2014-10-08 10:03:35 Diftraku: for xen? 2014-10-08 10:04:37 i know atleast xl and libvirt 2014-10-08 10:06:37 the old xm also works if you add xen-xend but will be removed soo as i know 2014-10-08 10:07:07 ncopa: cryptsetup task is done so far 2014-10-08 10:07:19 xend (xm) has been removed from xen 4.5 2014-10-08 10:07:59 alpine is on 4.3.3 2014-10-08 10:08:22 but you are right using xl is the way to go 2014-10-08 10:12:06 I'd love to have xcp-xapi 2014-10-08 10:12:20 ncopa: solution for unlocking luks devices at boot was quite easy. I just use the way gentoo does. could you take a look? 2014-10-08 10:12:20 been pondering what to run on 2014-10-08 10:21:29 ncopa: solution is just this - http://sprunge.us/VGOU 2014-10-08 10:34:07 tze: also has some crypto-in-initrd stuff 2014-10-08 11:44:28 darkfader: excuse me? 2014-10-08 11:45:37 xen_roger: there's one more guy that built some luks integration , i think he has it in the initramfs somehow but i don't exactly know all parts 2014-10-08 11:47:26 yes, i think it was for encrypted root 2014-10-08 12:31:40 haha! I just received a new development machin efrom my partner especially to prepare the ceph and qemu-rbd packages ;) He totally agrees that we need those package on Alpine ;) 2014-10-08 12:33:11 :) 2014-10-08 12:33:20 i dont think it will be too hard to port it to musl 2014-10-08 12:33:35 should I always quote in reply on mailinglist? 2014-10-08 12:34:01 might be good for the archives. sometimes we link to older discussion in mailing list 2014-10-08 12:34:24 full-quote? 2014-10-08 12:34:26 then its easy to understand what it is all about without clicking yourself through N older posts 2014-10-08 12:34:34 quote only only relevant 2014-10-08 12:34:45 ok. 2014-10-08 12:35:39 i will try get ceph working before v3.1 2014-10-08 12:36:46 ncopa: that's great news! 2014-10-08 12:37:16 the list is long though: http://bugs.alpinelinux.org/projects/alpine/issues?utf8=%E2%9C%93&set_filter=1&f[]=status_id&op[status_id]=o&f[]=fixed_version_id&op[fixed_version_id]=%3D&v[fixed_version_id][]=83&f[]=&c[]=project&c[]=tracker&c[]=status&c[]=priority&c[]=subject&c[]=assigned_to&c[]=updated_on&group_by= 2014-10-08 12:37:35 wget -qO- http://dl-3.alpinelinux.org/alpine/edge/main/x86_64/|grep openssh-client <-- says there is a openssh-client-6.6_p1-r5 2014-10-08 12:37:52 wget -qO- http://dl-3.alpinelinux.org/alpine/v3.0/main/x86_64/APKINDEX.tar.gz|tar -zxOf- APKINDEX|grep P:openssh-client -A1 <-- tells my apk to use V:6.6_p1-r3 2014-10-08 12:38:37 How can this be fixed? 2014-10-08 12:38:38 edge vs v3.0 2014-10-08 12:38:39 Oh 2014-10-08 12:38:41 My failt 2014-10-08 12:38:49 I've realized it right now 2014-10-08 12:38:53 m( 2014-10-08 12:39:10 alrighty :-) 2014-10-08 12:39:34 openssh 6.6 hole was about internal-sftp without chroot and doesn't affect grsec hehe 2014-10-08 12:40:02 hmm... I kinda forgot to follow up on that targetcli bug, ncopa 2014-10-08 12:40:12 jzono1: i just closed it :) 2014-10-08 12:40:18 i think it should be all ok now 2014-10-08 12:41:00 fabled: do you think we should cherrypick the main/openssh: curve25519pad patch added 2014-10-08 12:41:00 to v3.0? 2014-10-08 12:41:17 probably -- though, there was new openssl realease 2014-10-08 12:41:23 err 2014-10-08 12:41:24 openssh* 2014-10-08 12:41:33 6.7, right 2014-10-08 12:41:48 yeah, http://lwn.net/Articles/609321/ 2014-10-08 12:42:27 our HPN patches needs rebase 2014-10-08 12:42:36 is why didnt push it already 2014-10-08 12:42:43 we wont do 6.7 for v3.0 2014-10-08 12:44:08 the question is, should i add the curve25519pad patch to be included for v3.0.5 or not 2014-10-08 12:44:48 i suppose i should 2014-10-08 12:45:09 please do 2014-10-08 12:46:00 mmm 2014-10-08 12:46:13 should i rebase HPN? or do we wait for hpn upstream? 2014-10-08 12:46:18 dunno 2014-10-08 12:46:25 they havent been very quick 2014-10-08 12:46:34 and we probably want it before v3.1 2014-10-08 12:46:43 we can wait a week but not more 2014-10-08 12:47:41 btw: https://lists.mindrot.org/pipermail/openssh-unix-dev/2014-April/032494.html writes that 0.2 % of the connection attempts will fail - But it fails everytime when "KexAlgorithms curve25519-sha256@libssh.org" is specified in sshd_config 2014-10-08 12:49:08 jomat, what you mean fails everytime? 2014-10-08 12:49:37 oh. you mean patch applied, and KexAlgo like that -> nothing works 2014-10-08 12:49:46 fabled: try connecting i. e. to bender.greatnet.de without patch apploed 2014-10-08 12:49:50 *aplied 2014-10-08 12:50:10 it's because they can't negotiate kex 2014-10-08 12:50:15 OpenSSH_6.6 will result in a "Connection closed" 2014-10-08 12:50:24 OpenSSH_6.6.1 should tell you that you don't have a key 2014-10-08 12:50:27 the patch effectively disables curve25519 with older stuff 2014-10-08 12:50:58 Yeah 2014-10-08 12:51:00 it should disable it only if there's other acceptable kex options 2014-10-08 12:51:01 tagged 3.0.5 2014-10-08 12:52:28 or even better, automatically retry kex negotiation if one hit the 0.2% failure case 2014-10-08 14:25:48 I just updated http://bugs.alpinelinux.org/issues/3241 2014-10-08 14:26:01 with my work-in-progress 2014-10-08 14:34:20 if i can help with something, let me know, i.e. if you need docs for the wiki written or so 2014-10-08 14:49:44 Hello ! How may-I do cdrecord recognize SCSI driver ? Actually he says «Cannot open SCSI driver!» 2014-10-08 15:00:49 Goodbox: wild guess, but could you do "modprobe sr" 2014-10-08 15:01:30 yeah i loaded scsi_mod , i got sg but no sr 2014-10-08 15:01:34 ah ok 2014-10-08 15:01:37 hmmpf 2014-10-08 15:01:48 that is the only thing i could think of 2014-10-08 15:02:26 idem, strange thing is that i got sr_mod 2014-10-08 15:02:38 but cdrecord can't anywal list scsci devs 2014-10-08 15:02:44 anyway * 2014-10-08 16:27:13 Is there a reason I should upgrade a alpine-xen-2 something to alpin-xen-3? besides the potential fix for (null) domain? 2014-10-08 16:38:05 Frosh: tbh if you don't feel you're missing out on some feature... 2014-10-08 17:49:12 it's safe to assume that prebuilt static libraries are not candidates for inclusion in aports, yes? 2014-10-08 19:24:35 nah, I'm sorta happy with it, I don't even update, not sure what feature I could be missing, I just start up a os and I'm happy. 2014-10-08 19:24:54 ahills: we have a directory called non-free, which has some of those. 2014-10-08 19:25:10 but i think we prefer to keep it as small as possible. 2014-10-08 19:25:22 it's free, it's just built against libc because I can't build it against musl 2014-10-08 19:25:34 building gc from aports fails every single test 2014-10-08 20:40:19 oh cool, so xen devs didn't even backport the patches for booting rhel7 to 4.3 or 4.4.0 2014-10-08 20:40:30 i think it's enough, i'll just go buy a vmware license 2014-10-08 20:40:44 pays off in 2 months of not-working-around-shit 2014-10-09 09:44:43 hi 2014-10-09 10:03:36 Hello all 2014-10-09 10:39:48 hey all 2014-10-09 10:40:21 working on getting alpine working on linode 2014-10-09 10:40:30 used http://uggedal.com/journal/alpine-linux-on-linode/ 2014-10-09 10:41:26 as a beginning, I changed the version from 2.7 to 3.0, and the apk-tools-static to the current main for 3.0 2014-10-09 10:41:48 however it looks like for grub the boot process is out of date 2014-10-09 10:46:52 which seems to be similar to http://wiki.alpinelinux.org/wiki/Install_Alpine_on_Amazon_EC2 2014-10-09 10:47:11 however installing version 3.0 /boot/grsec & /boot/grsec.gz 2014-10-09 10:47:12 don't exist 2014-10-09 10:50:33 useless and unhelpful fact of the day: I didn't have any joy getting Alpine working on my VPS, which uses Citrix Xen server :( sad times :'( 2014-10-09 10:51:57 ginjachris: well the point I was getting to 2014-10-09 10:52:16 is that what is boot process for alpine 2014-10-09 10:52:38 so vmlinuz-grsec as far as I can tell is the kernel 2014-10-09 10:53:24 but I am not sure exactly how initrd was used for version 2.7 but now no longer exists for current alpine linux version 2014-10-09 10:53:32 so what is missing to get alpine to boot from grub 2014-10-09 10:54:58 well, I've never tried using Grub, Don't suppose this helps? http://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting 2014-10-09 10:55:18 ginjachris: http://bit.ly/1oTKtvN :P 2014-10-09 10:56:34 lol 2014-10-09 10:57:39 ginjachris: well I don't think 'pv-grub' is grub 2 2014-10-09 10:57:45 which the wiki is using 2014-10-09 10:57:57 but lemme go test 2014-10-09 10:59:09 I dunno, but under "Installing bootloader manually " it mentions grub 2 config 2014-10-09 10:59:28 Works fine here with grub2 2014-10-09 11:10:05 as i understand pv-grub != grub 2014-10-09 11:10:32 i'm experimenting with xen and openvswitch btw 2014-10-09 11:10:41 i must say that xen is pretty nice 2014-10-09 11:14:23 pv-grub is grub ported to run inside of a PV container, for grub2 supports is upstream, for grub1 I think there's a fork 2014-10-09 11:21:21 royger: so from pv-grub 2014-10-09 11:21:45 other than setting kernel to be /boot/vmlinuz-grsec 2014-10-09 11:22:16 is there something else to do do to specify the initrd 2014-10-09 11:24:21 else than specifying it ? 2014-10-09 11:25:41 quinq: well as far as I can tell 2014-10-09 11:25:52 for the grsec package on version 3.0 2014-10-09 11:25:58 there is no initrd image 2014-10-09 11:26:31 unless the grub command initrd creates it and doesn't read a file 2014-10-09 11:26:49 there is initramfs-grsec 2014-10-09 11:27:28 not sure what i was doing wrong :P 2014-10-09 11:27:33 ACTION feels like an idiot 2014-10-09 11:27:43 but it is 4:30 in the morning where i am :) 2014-10-09 11:27:55 quinq: thank you very much let me test out 2014-10-09 11:28:24 Maybe you should go to sleep and be back to it in the morning 2014-10-09 11:30:01 systmkor: no, to use pv-grub you have to set kernel = "/path/to/pv-grub" 2014-10-09 11:30:23 royger: for linode pv-grub is what you are initially booted into 2014-10-09 11:30:29 and trying to get pv-grub to boot up alpine 2014-10-09 11:30:46 ah, then you just need a proper grub.config that pvgrub can parse 2014-10-09 11:31:02 yup 2014-10-09 11:31:21 i though I couldn't find an init ramfs in /boot fro the linux grsec package 2014-10-09 11:31:26 but quinq pointed out that it exists there 2014-10-09 11:35:42 btw, thank you all :) 2014-10-09 12:38:48 quinq: royger wooo, got it booted and running 2014-10-09 12:48:31 systmkor: maybe help us update the wiki pages? 2014-10-10 06:17:57 how can I see wich files are inside a *.apk package ? 2014-10-10 07:00:06 ncopa: hello sir, I would commit some new package I need ... could you explain me wich would the best way to do ? 2014-10-10 07:12:53 Good Morning all 2014-10-10 07:14:42 morning 2014-10-10 07:14:59 xen_roger: git add APKBUILD && git commit? 2014-10-10 07:15:06 i'm not I understand 2014-10-10 07:15:51 yes, fo sure, but can I commit to alpine? 2014-10-10 07:18:17 http://krautchan.net/files/1412925180001.png 2014-10-10 07:18:23 tfw alpine not mainstream culture yet 2014-10-10 07:35:54 <^7heo_> is there any vnc server in alpine? 2014-10-10 07:50:57 xen_roger: no, you cannot 2014-10-10 07:51:08 you could have your own forked git tree, and we can pull from it 2014-10-10 07:51:18 or you could send patches one by one 2014-10-10 07:51:24 yes, my fail ... i just put it into mailing list again ;) 2014-10-10 07:51:36 if you have many packages, i prefer to do a git pull 2014-10-10 07:51:48 no at the moment a single on 2014-10-10 07:51:50 +e 2014-10-10 07:52:17 you can also if its trivial sprunge it here 2014-10-10 07:52:33 with: git format-patch -1 --stdout | sprunge 2014-10-10 07:53:31 ^7heo_: i dont think so. i think i tried to package it ones. but it was not that easy. 2014-10-10 07:55:19 on done 2014-10-10 07:59:10 katlogic: :D 2014-10-10 08:01:46 <^7heo_> clandmeter: thanks for the info. 2014-10-10 13:36:21 <^7heo_> does anyone here know how to configure IPv6 with DHCPv6 on busybox? 2014-10-10 13:37:46 I'm afraid I've only ever done static IPv6 addresses :( 2014-10-10 13:38:20 is DHCPv6 the same as SLaaC? That seemed to work very well when I last used it (although that wasn't alpine) 2014-10-10 13:40:06 <^7heo_> no, SlaaC is another beast. 2014-10-10 13:41:47 so i guess you want udhcpc6? 2014-10-10 13:42:55 i can't say I've used it, I don't suppose man pages help? http://linux.die.net/man/8/dhcp6c and http://linux.die.net/man/5/dhcp6c.conf 2014-10-10 13:43:56 oh dhcp6c is the dhcp client 2014-10-10 13:43:58 oops 2014-10-10 13:46:34 the isc ipv6 dhcp client 2014-10-10 13:48:22 ^7heo_, it seems that busybox doesn't support that, you'll have to stick to isc dhc6c 2014-10-10 13:55:44 <^7heo_> okay thanks :) 2014-10-10 14:05:40 <^7heo_> dhcpcd installed, thanks :) 2014-10-10 14:08:55 ACTION misses the time he had a real ISP giving an IPv6 prefix 2014-10-10 14:09:15 <^7heo_> I'm not sure if I like IPv6 2014-10-10 14:09:25 <^7heo_> it can work, but it's a mess. 2014-10-10 14:11:00 It sure isn't 2014-10-10 14:11:19 It's well less a mess than having to deal with NAT 2014-10-10 14:13:00 <^7heo_> I love archlinux. problem: No IPv6 Routers available. Fix? Remove this line: eth0=dhcp 2014-10-10 14:13:19 <^7heo_> quinq: the DHCPv6 protocol is a real mess. 2014-10-10 14:13:23 <^7heo_> quinq: SLAAC is okay tho. 2014-10-10 14:14:21 You mean it's a mess as it's internal protocol or as using a dhcp6 client ? 2014-10-10 14:14:38 <^7heo_> also, some very clever people think and say it's harmful, and I am not sure if it's much better than IPv4 2014-10-10 14:14:49 <^7heo_> we would need a better alternative imho 2014-10-10 14:15:01 <^7heo_> quinq: the complexity of it. 2014-10-10 14:15:46 Well, I haven't looked into it, but I had no problem setting my dhcp6 client (although I would have prefer not to use it) 2014-10-10 14:15:58 How harmful is it ? 2014-10-10 14:17:43 (I always heard people not knowing anything in networking saying that ipv6 is bad, because they feared its novelty, other saying it's really good) 2014-10-10 14:18:11 But if you have some paper or article about that, I would gladely read it 2014-10-10 14:18:16 gladly 2014-10-10 14:20:16 <^7heo_> okay, moment, I try to find that. 2014-10-10 14:20:57 On the other hand, I can search for myself 2014-10-10 14:21:17 <^7heo_> yeah, but I know someone who can know where to find the info, so I just asked. 2014-10-10 14:21:21 <^7heo_> it can come later :) 2014-10-10 14:21:27 I see a pdf from defcon in 2007 2014-10-10 14:21:38 thanks 2014-10-10 14:21:43 <^7heo_> quinq: link? 2014-10-10 14:22:08 https://www.defcon.org/images/defcon-15/dc15-presentations/Lindqvist/Whitepaper/dc-15-lindqvist-WP.pdf 2014-10-10 14:23:16 Oh well, it says that autoconf based on MAC address is a privacy issue 2014-10-10 14:23:18 Nevermind. 2014-10-10 14:23:31 <^7heo_> yeah =/ 2014-10-10 14:23:50 <^7heo_> like the MAC address isn't broadcasted in WiFi connections... 2014-10-10 15:00:43 the mac address problem really isn't one, you can a) modify your mac address or (b) generate an IPv6 address that is 'random' (ie. not related to your MAC 2014-10-10 15:00:45 ) 2014-10-10 15:01:17 Yes, thank you ginjachris 2014-10-10 15:01:33 :D sometimes I can be useful 2014-10-10 15:02:49 <^7heo_> quinq: okay, the main problem that my buddy reports is the fact that IPv4 w/ NAT hides everything by default (in the local network) 2014-10-10 15:02:57 <^7heo_> while IPv6 can expose everything by default. 2014-10-10 15:03:09 haha 2014-10-10 15:03:22 Yes, that's the one thing that everybody talks about 2014-10-10 15:03:27 ie not knowing networking 2014-10-10 15:03:56 Thinking having a router nating a connexion is a security 2014-10-10 15:04:25 People must be using firewalls, that's a fact 2014-10-10 15:04:31 <^7heo_> This guy is paid by the biggest German bank as a security consultant, so I wouldn't say he knows nothing about security. 2014-10-10 15:04:32 i have some random ipv6 books i have acculmulated but not read yet, if anyone wants a copy? 2014-10-10 15:04:42 With a firewall, this pseudo problem you're talking about (which isn't) is solved 2014-10-10 15:04:49 <^7heo_> He perfectly knows it can work afaik. 2014-10-10 15:04:56 <^7heo_> He just plans for the worst 2014-10-10 15:04:56 nat is not security, for sure 2014-10-10 15:05:00 Yeah, paygraid is always a sign of knowledge 2014-10-10 15:05:21 <^7heo_> quinq: no, but banks spend a lot of money getting the right people, especially for security. 2014-10-10 15:05:24 <^7heo_> Anyway. 2014-10-10 15:05:41 <^7heo_> Also, he never said that NAT was a security. 2014-10-10 15:05:42 apparently alot of things (like ipsec vpn's) were originally designed for ipv6 and have effectively been backported to ipv4 2014-10-10 15:05:50 <^7heo_> He just said that IPv6 was a potential problem. 2014-10-10 15:06:36 Well, I don't know what he said exactly, but it's implied in saying that saying 2014-10-10 15:06:48 gnnn 2014-10-10 15:06:52 -saying 2014-10-10 15:07:36 <^7heo_> You imply whatever you want. 2014-10-10 15:07:47 <^7heo_> Truth is that NAT hides the internal network. 2014-10-10 15:07:55 <^7heo_> It's not a security feature, but it happens. 2014-10-10 15:07:56 Hey, I don't want to criticise your friend here, I don't care :) 2014-10-10 15:08:05 <^7heo_> I'm not defending him either. 2014-10-10 15:08:13 <^7heo_> I try to be clear about what is and what is not. 2014-10-10 15:08:28 <^7heo_> NAT isn't a security feature, nobody said that here or on the other chan I asked on. 2014-10-10 15:08:50 <^7heo_> but NAT effectively hides the internal network, period. 2014-10-10 15:08:56 I dont really understand how hidden is a computer connecting to the internet 2014-10-10 15:09:00 NATted or not 2014-10-10 15:09:00 <^7heo_> Now, it's how it WORKS. 2014-10-10 15:09:08 <^7heo_> it's not a configuration, it's the mechanic of it. 2014-10-10 15:09:11 <^7heo_> so it hides by design 2014-10-10 15:09:20 That's why I said not understanding how networking works 2014-10-10 15:09:38 <^7heo_> Explain to me why I don't understand then :) 2014-10-10 15:09:42 <^7heo_> I'm all ears. 2014-10-10 15:09:59 When you connect to a computer on the Internet, from a « local » IPv4 computer, is it hidden ? 2014-10-10 15:10:16 You make a connection to that computer ie not hidden anymore 2014-10-10 15:10:28 The same is for IPv6 2014-10-10 15:10:39 <^7heo_> The remote server does not know more than the NAT router. 2014-10-10 15:10:42 <^7heo_> if it's what you ask. 2014-10-10 15:10:55 So you think that is hiding 2014-10-10 15:11:02 <^7heo_> If you have 2 devices behind a NAT router, the server can't tell if you're one, two, or one hundred. 2014-10-10 15:11:14 <^7heo_> behind IPv6, this is not necessarily true. 2014-10-10 15:11:15 You think in term of address 2014-10-10 15:11:17 Not of connexion 2014-10-10 15:11:19 gotta be careful here because ther is more to identification than IP's... 2014-10-10 15:11:20 <^7heo_> Ofc. 2014-10-10 15:11:22 Who cares about address ? 2014-10-10 15:11:26 <^7heo_> ginjachris: true. 2014-10-10 15:11:39 <^7heo_> quinq: because most of the time, one device == one address. 2014-10-10 15:12:19 Well, I think that's where you're confused 2014-10-10 15:12:44 <^7heo_> I must be confused, you must be right. 2014-10-10 15:12:49 when ipv6 properly gets taken up (which isn't a simple thing for alot of companies, we're talking at least a year of prep and change for alot of them) networking is going to have to be looked at from anther angle....it's gonna have to be done properly and not just bodged with a firewall on the perimeter 2014-10-10 15:13:05 <^7heo_> < ginjachris> "when ipv6 properly gets taken up" 2014-10-10 15:13:08 <^7heo_> no need to quote more. 2014-10-10 15:13:12 <^7heo_> That's EXACTLY the point here. 2014-10-10 15:13:17 Yeah, transition 2014-10-10 15:13:26 <^7heo_> security can't rely on things being properly done. 2014-10-10 15:13:26 Moving things 2014-10-10 15:13:30 The most difficult thing 2014-10-10 15:13:45 <^7heo_> security must be enforced by design. 2014-10-10 15:13:58 <^7heo_> otherise a sign "please don't enter" is enough to replace a bank vault. 2014-10-10 15:14:07 lol 2014-10-10 15:14:40 <^7heo_> I mean, in germany, a sign "please don't enter" would work as well, pretty much everywhere, excepted in big cities... 2014-10-10 15:14:44 <^7heo_> but it's not the point. 2014-10-10 15:14:46 17:13 < ^7heo_> security must be enforced by design. 2014-10-10 15:14:47 haha 2014-10-10 15:14:54 <^7heo_> laugh all you want. 2014-10-10 15:15:05 <^7heo_> I don't really care. 2014-10-10 15:15:10 <^7heo_> However, I have work to do. :) 2014-10-10 15:15:11 And before that, you said that IPv6 is less humm secure ? than IPv4 2014-10-10 15:15:13 come on 2014-10-10 15:15:17 <^7heo_> 'nuf time lost here. 2014-10-10 15:15:51 <^7heo_> quinq: I never said it's less secure, I said it wasn't the proper expected evolution. 2014-10-10 15:16:04 <^7heo_> IPv4 has an excuse, it's damn old. 2014-10-10 15:16:09 <^7heo_> IPv6 is just a failure. 2014-10-10 15:16:12 <^7heo_> anyway, as said 2014-10-10 15:16:15 <^7heo_> too much time lost. 2014-10-10 15:16:16 <^7heo_> laters. 2014-10-10 15:16:28 :( 2014-10-10 15:16:32 <^7heo_> wait, horrible doubt. 2014-10-10 15:16:46 <^7heo_> okay, can't know. 2014-10-10 15:18:45 let's leave it at IPv6 being a subject very much under discussion for which everyone has a different opinion :) getting away from IPv4 will do doubt be a good thing.....here is the UK IPv6 uptake is slooooooooow which is...annoying 2014-10-10 15:22:32 I have to go soon, so I wish everyone a gloriously happy weekend! 2014-10-10 15:22:47 Here is some crazy drum and bass to get you in the party mood: http://grooveshark.com/s/Flash+Fire/3qtCb2?src=5 2014-10-10 15:50:12 Anyone have vagrant boxes for alpine? A quick search came up empty 2014-10-10 16:51:34 don't usually keep vagrants in my home 2014-10-11 04:17:00 hi there, I've been doing quite alot of research on how to Update and upgrade Alpine, and all I get is this: 2014-10-11 04:17:36 "4462 packages available" "OK: 254 MiB" but nothing installs 2014-10-11 04:17:46 then you're up to date! 2014-10-11 04:18:21 i was used to see the downloading and installing like in Debian-based distros 2014-10-11 04:18:33 it won't install anything if there's nothing to install 2014-10-11 04:19:03 I see, i thought it was listing the updates that could be installed 2014-10-11 04:19:10 thanks 2014-10-11 04:19:19 apk update will get the list of packages 2014-10-11 04:19:34 but it won't display it for you 2014-10-11 04:19:48 4462 is the total number of packages available to you 2014-10-11 04:20:06 so it means they aren't installed yet? 2014-10-11 04:20:14 no, generally available 2014-10-11 04:20:32 it means the repositories have that many packages to install 2014-10-11 04:20:40 you have probably installed several as part of installing your system 2014-10-11 04:20:49 like linux-grsec, busybox, etc 2014-10-11 04:21:52 i c, how can install them? i've tried the update, upgrade commands among others, same message 2014-10-11 04:22:01 install what? 2014-10-11 04:22:11 install ALL the packages?? 2014-10-11 04:23:37 let's say I needed to install them all 2014-10-11 04:23:44 how would i go about that? 2014-10-11 04:23:47 ok, you probably shouldn't try to do that 2014-10-11 04:23:53 some of them will conflict with one another 2014-10-11 04:24:38 ok, so do they just need to be installed as needed? 2014-10-11 04:24:56 yes, apk add any packages you want to use 2014-10-11 04:25:25 it works the same as apt-get update/apt-get upgrade 2014-10-11 04:25:36 apk add is like apt-get install 2014-10-11 04:25:48 I say it works the same... actually it works much better 2014-10-11 04:25:55 apk doesn't trap itself in endless dependency conflicts 2014-10-11 04:26:33 really? heck I spent hours of research, good to know i can use debian like commands 2014-10-11 04:27:13 you can't use apt-get, but apk works the same as most other package managers 2014-10-11 04:27:27 if you want to see all the commands, just type apk 2014-10-11 04:28:04 for e.g. help on the add command, apk add --help 2014-10-11 04:28:05 etc, etc 2014-10-11 04:28:14 i see 2014-10-11 04:28:57 and unlike aptitude, there aren't 100 different commands with no clear feature separation 2014-10-11 04:29:15 interesting 2014-10-11 04:29:39 I read the main purpose of Alpine was/is networking stuff, 2014-10-11 04:29:49 how stable in your experience? 2014-10-11 04:29:53 really? 2014-10-11 04:30:06 very stable in my experience; I use it in xen 2014-10-11 04:30:18 I use it for firewall, lots of VMs, etc 2014-10-11 04:30:43 only issue I've run into is trying to build libgc, which requires gnu libc 2014-10-11 04:30:50 oh wow, i'm a newbie to alpine 2014-10-11 04:31:11 I liked the fact that i took mins to install 2014-10-11 04:31:21 yeah 2014-10-11 04:31:29 that, and it takes almost no resources to run 2014-10-11 04:31:35 so I can spin up tons of VMs 2014-10-11 04:31:38 very quickly 2014-10-11 04:32:01 i c, do you have any GUI desk, let's say Gnome, or other 2014-10-11 04:32:20 I use dwm 2014-10-11 04:32:31 http://dwm.suckless.org/ 2014-10-11 04:33:24 nice, pretty descriptive name 2014-10-11 04:33:52 I'm sure someone else in here uses a desktop environment 2014-10-11 04:35:46 here is what im planning on doing, list the updates/upgrades available and then just download the ones 2014-10-11 04:36:20 apk update downloads an index of available packages, apk upgrade updates any packages that have a different version than the index 2014-10-11 04:36:39 if you want to find new packages to install, you can use apk search 2014-10-11 04:37:49 cool 2014-10-11 04:40:20 so if i want to install "samba-libs..." should i do: apk upgrade samba-libs..? or any other package name? 2014-10-11 04:40:34 no, use "add" 2014-10-11 04:40:37 apk add samba-libs 2014-10-11 04:40:54 upgrade will upgrade to new versions of what you already have installed, if they are available 2014-10-11 04:41:47 aahh k k 2014-10-11 04:46:13 thanks for your help ahills! 2014-10-11 04:46:19 no problem 2014-10-11 10:02:52 Hi, there! Is there an package containing man pages of the linux/posix c programming interface (like man pages for malloc, mmap, open, ...)? 2014-10-11 10:05:39 Hi maribu, yes like everywhere else, in man-pages 2014-10-11 10:28:50 cool, thanks :-) 2014-10-11 10:30:18 quinq: Thank you very much :-) 2014-10-11 10:31:37 You're welcome :) 2014-10-12 19:17:13 hi 2014-10-12 22:13:12 how do I boot root-on-lvm? 2014-10-12 22:13:32 or rather, force initrd to have lvm support 2014-10-13 00:15:32 what package contains user management tools 2014-10-13 00:15:50 like useradd or usermod 2014-10-13 00:16:38 in the apk repo 2014-10-13 00:17:56 There isn't useradd or usermod, however there is adduser from busybox. 2014-10-13 00:19:45 ah 2014-10-13 00:19:55 i take it is a coreutils vs busybox thing 2014-10-13 00:20:30 Goodbox: thanks 2014-10-13 00:21:22 np, you can install user management tools you need, or there's coreutils packaged. 2014-10-13 00:22:10 the ones busybox provides should more than suffice 2014-10-13 00:22:24 i thought I tried addgroup but clearly I messed that up 2014-10-13 00:22:27 yeap probably 2014-10-13 06:03:15 does anyone know about alpine and snmp? can´t get informations from agent... response is mostly "noSuchObject" - snmpd is configured. 2014-10-13 06:09:22 it querys the host - got some reponse and just stops 2014-10-13 06:20:30 I was wondering if there is some bug with xen and centos under pygrub? it seems to stay in blocked state during centos7 bootup, not sure how to do pvgrub on centos 7 without grub-legacy, issue happens from time to time. This is on alpine-3, will there be some pvgrub2? 2014-10-13 07:21:03 good morning all 2014-10-13 07:37:47 <^7heo_> moiin 2014-10-13 08:55:57 CcxCZ i wrote this http://wiki.alpinelinux.org/wiki/Installing_on_GPT_LVM maybe you can use it for your task. 2014-10-13 17:10:57 hey all, when I try booting Alpine from the installation CD using the alpine-mini.iso 2014-10-13 17:11:15 for version 3.0.4 & 3.0.5 on a Dell 1850 2014-10-13 17:11:18 server I am working on 2014-10-13 17:11:29 when it's booting from the CD 2014-10-13 17:11:35 i get Mounting boot media failed 2014-10-13 17:11:59 initramfs ermergency recover shell launched. 2014-10-13 17:12:08 sh: can't access tty; job control turned off 2014-10-13 17:15:17 when I boot the CD using VirtualBox it works correctly 2014-10-13 17:21:01 and I have been able to install an other Linux distro on this machine before 2014-10-13 17:23:31 well I think i found my solution http://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB 2014-10-13 21:52:38 Hi people, i'm messing with an install using qemu vm in a qemu vm, i have no resources for a full blown solution at this time, this should work 2014-10-13 21:53:29 BUT when i boot the VM with the -nographic option the ssh and vnc session show '1024x768 mode' which i cannot escape 2014-10-14 01:49:51 hey all is anyone compiling sdparm 2014-10-14 01:49:56 since there is hdparm in the repo 2014-10-14 01:50:02 I am thinking about running it on a server I have 2014-10-14 08:59:28 does anyone know if there is a way to combine acl-names in squid? say there are two acls "acl foo" and "acl bar" - could I do something like "set baz foo bar" so I can use baz in the rest of the config? 2014-10-14 09:04:12 xen_roger: afaik no 2014-10-14 09:05:30 damn ... it makes it difficult when it comes to exclude diffrent acls from long rule set ... have to add every acl ... :( 2014-10-14 10:01:35 . 2014-10-14 11:19:02 Greetings all 2014-10-14 11:19:41 <^7heo_> moin 2014-10-14 14:06:11 Does anybody run KVM virtualization on Alpine? 2014-10-14 14:06:39 I do. with openvswitch. 2014-10-14 14:07:30 with alpine as kvm host? 2014-10-14 14:07:44 yes with alpine on the bare metal 2014-10-14 14:08:31 thats good. I'm looking for qemu-kvm, but can't find it :-/ 2014-10-14 14:09:21 I use "qemu-systems-x86_64 -enable-kvm" 2014-10-14 14:09:57 Perfect! That was the info i needed, i think now I have everything together :-) 2014-10-14 14:10:20 great! ;) 2014-10-14 14:11:33 ... well... almost everything, still waiting for the confirmation mail of my hoster %) 2014-10-14 14:16:56 <^7heo_> You need your hoster to run KVM at home? 2014-10-14 14:17:02 <^7heo_> or do your KVM on your dedicated server? 2014-10-14 14:19:36 ^7heo_: I'm currently running debian on most of my servers, I want to migrate to Alpine 2014-10-14 14:19:51 <^7heo_> wise move. 2014-10-14 14:19:59 <^7heo_> also 2014-10-14 14:20:01 <^7heo_> you can use docker. 2014-10-14 14:20:02 jomat: I'm in the same mood ;) 2014-10-14 14:20:27 <^7heo_> what I do is run anything on docker on alpine on kvm on alpine. 2014-10-14 14:20:42 <^7heo_> anything can contain ubuntu or debian or whatever. 2014-10-14 14:20:54 <^7heo_> 'cause you have real virtualization when you want 2014-10-14 14:21:02 <^7heo_> as well as disposable Linuxes when you want. 2014-10-14 14:21:26 ^7heo_: I know. Probably I'll use lxc, but there are some "special" machines which need hw virtualization... so I wanted to know if it's possible :-)( 2014-10-14 14:21:29 :-) 2014-10-14 14:21:35 <^7heo_> ah yeah 2014-10-14 14:21:42 <^7heo_> KVM is better than docker for mostly anything 2014-10-14 14:21:46 <^7heo_> but it's also heavier 2014-10-14 14:22:14 <^7heo_> Also 2014-10-14 14:22:22 <^7heo_> with KVM you can actually run NetBSD :P 2014-10-14 14:22:50 I don't need an own kernel for most of my stuff 2014-10-14 14:23:30 jomat: i use KVM on my desktop, virt-manager 2014-10-14 14:24:27 ncopa: Does your virt-manager use the ssh-agent? This worked with debian, i have to start it with --no-fork on alpine so I can enter the passphrase... 2014-10-14 14:25:11 i use polkit, no ssh atm 2014-10-14 14:25:29 NetBSD \o/ 2014-10-14 14:25:41 I also lost all my hosts as the alpine virt-manager doesn't use the gconf-config-stuff here... but actually i hate virt-manager >:-( 2014-10-14 14:25:43 i think you can make ssh prompt for password witih askpassword 2014-10-14 14:26:04 i dont think i would use virt-manager for a server setup 2014-10-14 14:26:05 I don't have password auth activated on my servers 2014-10-14 14:26:24 its okish for development desktop 2014-10-14 14:26:29 I use virt-manager on my desktop to connect to libvirt-servers 2014-10-14 14:26:37 ok 2014-10-14 14:26:55 i think you can set an env var ASKPASS or soo 2014-10-14 14:27:10 but ... it takes ages to connect to a server with more than... let's say 30 VMs on it 2014-10-14 14:27:13 and make ssh client look for an executable named askpass 2014-10-14 14:27:51 jomat: why don't you use ssh-add ? 2014-10-14 14:28:05 does it ask for ssh password or the ssh key password? 2014-10-14 14:28:13 you can use ssh-add for key password 2014-10-14 14:28:38 i use ssh-add 2014-10-14 14:29:08 if i open a terminal and type "ssh root@someserver" I'm on it without entering anything else 2014-10-14 14:29:29 If I open virt-manager in this terminal, it wants to ask for the passphrase of the key 2014-10-14 14:30:02 ssh-add -l lists my keys 2014-10-14 14:30:19 SSH_AGENT_PID and SSH_AUTH_SOCK are set 2014-10-14 14:30:27 virt-manager just ignores it... 2014-10-14 14:30:50 hmm 2014-10-14 14:30:52 weird 2014-10-14 14:32:13 i actually only use virt-manager to connect to the vnc-displays because i'm still to lazy to write a two-liner that connects to the port returned by "virsh vncdisplay 1.2.3.4" 2014-10-14 14:35:56 jomat: if you have DISPLAY set and SSH_ASKPASS set and there are no terminal with the process, it will execute $SSH_ASKPASS 2014-10-14 14:36:33 i think you can try apk add openssh-askpass 2014-10-14 14:37:01 make sure that SSH_ASKPASS is set from env 2014-10-14 14:37:06 yeah, but it just should use the ssh-agent-keys :-| 2014-10-14 14:37:29 yes 2014-10-14 14:37:30 hm 2014-10-14 14:37:34 I don't want to enter passphrases all the time... it will ask again once you select a vm on that host 2014-10-14 14:37:53 how do debian solve that? 2014-10-14 14:38:16 dunno, it just works if i add my keys with ssh-add :-) 2014-10-14 14:38:53 s/works/worked/;s/add/added/ 2014-10-14 15:56:57 reading today's chatter, folks use kvm and not xen? 2014-10-14 15:58:44 I use Xen :) 2014-10-14 16:04:41 do you have any hardware virtualized guests with xen? 2014-10-14 16:05:04 or just paravirtualized 2014-10-14 16:06:28 <^7heo_> Xen too here, but not on Alpine. 2014-10-14 16:07:20 <^7heo_> (NetBSD can run Xen) 2014-10-14 16:07:26 <^7heo_> and it's MUCH better 2014-10-14 16:08:38 ScrumpyJack: yes, mostly HVM 2014-10-14 16:12:16 ^7heo_: what is your Dom0 for xen? 2014-10-14 16:16:56 <^7heo_> NetBSD 6 2014-10-14 16:17:05 <^7heo_> ScrumpyJack: but it's EOB here. 2014-10-14 16:17:06 <^7heo_> brb. 2014-10-14 16:17:07 <^7heo_> o/ 2014-10-14 16:43:58 xen on alpine is wonderful 2014-10-14 16:51:54 ahills: it is! ;) 2014-10-14 16:54:17 roger that 2014-10-14 18:45:27 hi there, is there a way to get wicd running on alpine? 2014-10-14 18:55:57 the answer to "can I run X" is usually "probably, as long as it can be built against musl" 2014-10-14 18:58:32 I may have missed it but with apk how do you list installed packages 2014-10-14 18:58:57 found it my b 2014-10-14 19:05:53 ncopa: when running gradme -P the error "Could not open /dev/grsec." is printed 2014-10-14 19:06:12 Online there are some recommendations using mknod or makedev 2014-10-14 19:06:41 is there a prefered method on alpine to do this, possibly with openrc on boot up 2014-10-14 19:07:11 and should the gradm package or the linux-grsec package create this device for you? 2014-10-14 19:14:17 systmkor: list install pkg with "apk info | sort" 2014-10-14 19:20:38 xen_roger1: thanks, yah I ended up finding it on the wiki 2014-10-14 19:21:10 xen_roger1: that said, that feels like a very counter intuitive way of searching for packages 2014-10-14 19:21:45 systmkor: it took me a few months to find out where kpartx was 2014-10-14 19:22:11 yet, it's a very random thing to bug me 2014-10-14 19:22:15 *rare 2014-10-14 19:22:37 i guess othe question is how do you list the installed files of a package? 2014-10-14 19:24:39 yeah afaik there's nothing to list uninstalled packages' files. but i think someone worked on that, if i remember correctly 2014-10-14 19:26:41 systmkor: to put in perspective, once you saw that alpine can back out from a botched upgrade because the package manager works so well happiness is the more common mood :) 2014-10-14 19:26:59 systmkor: list files by installed package -> "apk info -La pkgname" 2014-10-14 19:27:32 xen_roger1: thanks 2014-10-14 19:28:07 sorry, hadn't understood you only needed to look at installed ones 2014-10-14 19:37:18 xen_roger: just noticed that i have ~1GB ram with dom0_mem=2048, do you have an idea why noone is poking at that? 2014-10-14 19:40:16 darkfaded: running from RAM ( LIVE BOOT ) ? 2014-10-14 19:40:21 no 2014-10-14 19:40:35 this is a bug many people have 2014-10-14 19:40:40 it's not any stupid issue 2014-10-14 19:40:58 i just upgraded to 3.0 on this one and saw it's still around 2014-10-14 19:41:15 ? dont know ... running dom0 at 512MB 2014-10-14 19:41:27 let me ask the other way round: is your dom0 mem exactly as you set it? 2014-10-14 19:41:44 and are you running a stable version 2014-10-14 19:41:54 because, it really doesn't matter if it works in edge lol 2014-10-14 19:41:56 no .... as I run a ram-system it´s not 2014-10-14 19:42:21 ok, for ram boot it's understandable 2014-10-14 19:42:25 << running @3.0.5 2014-10-14 19:42:43 i'll open a bug for it in a few days so it's at least trackable 2014-10-14 19:42:56 ah :) same version :) 2014-10-14 19:43:15 it doesn't seem to panic with flashcache any longer 2014-10-14 19:45:16 i had bought sas ssds to use the writecache mode and write caching caused a nice, immediate panic 2014-10-14 19:46:43 << does not use any heavy-disk envs ... mostly works on ram ;) 2014-10-14 19:54:51 that way definitely faster :) 2014-10-14 19:55:11 seems the crash is gone so i could resume where i left off in april 2014-10-14 20:03:31 long time ago ... 2014-10-14 20:24:19 xen_roger or darkfaded do either of you happen to have experience with grsec on alpine? 2014-10-14 20:24:57 systmkor: it's there any works, i don't have much insight 2014-10-14 20:25:00 *and it 2014-10-14 20:25:16 systmkor: running all my systems with grsec --- now again ;) 2014-10-14 20:27:19 xen_roger: cool 2014-10-14 20:27:29 so what is with running gradm -P on first try 2014-10-14 20:27:35 says that /dev/grsec doesn't exist 2014-10-14 20:27:50 I am running the grsec kernel with alpine 2014-10-14 20:29:04 I glanced at this http://forums.gentoo.org/viewtopic-t-984246-start-0.html 2014-10-14 20:29:10 but it seems that nothing was really resolved 2014-10-14 20:44:07 systmkor: look at this 2014-10-14 20:44:23 localhost:~# ls /dev/grsec -la 2014-10-14 20:44:23 -rw--w--w- 1 root root 0 May 15 09:45 /dev/grsec 2014-10-14 20:44:24 localhost:~# gradm -P 2014-10-14 20:44:24 Setting up grsecurity RBAC password 2014-10-14 20:44:24 Password: 2014-10-14 20:57:46 cya 2014-10-14 20:59:48 nite 2014-10-15 00:49:03 do you guys more commonly use GCC or Clang for building packages? 2014-10-15 02:40:01 gcc 2014-10-15 02:40:17 tcc 2014-10-15 02:40:38 ACTION →[] 2014-10-15 02:40:44 ... 2014-10-15 03:47:22 ncopa, @alpinelinux's profile still says uclibc :-p 2014-10-15 03:47:31 "A security-oriented, lightweight Linux distribution based on uClibc and Busybox." 2014-10-15 06:10:59 morning 2014-10-15 06:11:07 dalias: where did you read that? 2014-10-15 06:12:18 hi 2014-10-15 06:13:02 ncopa. twitter account 2014-10-15 06:13:15 who runs it? 2014-10-15 06:13:38 i think its alacerda who set it up 2014-10-15 06:15:49 here too: http://wiki.xen.org/wiki/Template:Distro_Resources 2014-10-15 07:08:53 Morning all! 2014-10-15 07:09:06 morning 2014-10-15 07:30:33 morning 2014-10-15 07:30:54 <^7heo_> moin clandmeter 2014-10-15 07:33:28 ncopa: I want to move a pkg from testing to main, but this package has lots of deps of which im not sure they are in main yet. what would be the best way to check this? 2014-10-15 07:33:50 I guess i would need something that can do recursive deps and show which repo its installed from. 2014-10-15 08:08:54 hello everybody 2014-10-15 08:09:49 clandmeter: currently the only way is manual 2014-10-15 08:09:57 hi luigi_345 2014-10-15 08:10:05 I'm working on a failover cluster manager written in posix shell script. I'm looking for buddies that want to play with it and contribute in the development. 2014-10-15 08:11:12 while it's not focused on a particular linux distro I was wondering if the alpine linux community can be interested in it, maybe by taking a look at the code and write with me a wiki on how set up a cluster or.. i don't know ... 2014-10-15 08:11:16 :) 2014-10-15 08:11:32 thaat sounds awesome 2014-10-15 08:11:42 unfortunately I am not a coder 2014-10-15 08:11:45 :( 2014-10-15 08:11:50 ome not too :D 2014-10-15 08:11:58 I'm a sysadmin with some coding skill 2014-10-15 08:12:24 that's more coding than me! 2014-10-15 08:12:31 what kind of cluster? 2014-10-15 08:12:37 virtual machien cluster? 2014-10-15 08:12:38 anyway, first step, the link :D https://code.google.com/p/back-to-work/ one big missing is the documentation 2014-10-15 08:12:44 failover cluster 2014-10-15 08:12:52 high availability 2014-10-15 08:13:21 I inspire to HP ServiceGuard my preferred (unfortunally proprietary) cluster manager 2014-10-15 08:13:45 but I will not implement all the feature and options of service guard while there are some new. 2014-10-15 08:14:02 I want to make it run (and for now it run) on ESXi 2014-10-15 08:14:09 to implement a failover for virtual machines 2014-10-15 08:14:49 I just need to learn how to create the VIB package to install on ESXi otherwise when you reboot the ESXi host all the files get wiped out 2014-10-15 08:15:32 and I need an ESXi lab :D for now I have only one virtualized in kvm but nested virtualization is a little hard on my old laptop 2014-10-15 08:17:04 anyway, my goal now is to find someone interested in it, I'm alone for now, I have a friend that will join in the future but busy at work at present 2014-10-15 08:17:23 i dont have time or energy for it, sorry 2014-10-15 08:17:36 ncopa: thanks anyway 2014-10-15 08:18:03 I tried to talk about this on #linux-ha but I din't find a warm welcome :D 2014-10-15 08:18:11 :) 2014-10-15 08:18:46 I know it sound crazy to rewrite everything in shell but with less than 1000 line of code I'm about at 90% of the code needed :D 2014-10-15 08:19:15 now it's just about cleaning, fixing, improving, rewriting again :D 2014-10-15 09:42:24 Hi ! 2014-10-15 09:43:57 luigi_345: you're making a clone of service guard? nice :) 2014-10-15 09:44:24 i ran into a non-kindergarten but basic OSS cluster manager yesterday 2014-10-15 09:45:25 darkfader: I started with the idea of a clone of serviceguard but while still implementing a failover cluster manager my project took some different direction 2014-10-15 09:45:28 for example 2014-10-15 09:46:04 in back-to-work (my project) you have a set of lock servers (ideally a minimum of 3) 2014-10-15 09:46:22 and those server can have only the lock server role, no package running 2014-10-15 09:46:43 then you can have one (but ideally more then one :)) server where you run the package 2014-10-15 09:46:56 then you want failover? run the package on another node 2014-10-15 09:47:30 you want more failover? run the package (back-to-work ) on a third server 2014-10-15 09:48:01 what a like of service guard is that it's simple to set up. mine cluster is even simpler :D 2014-10-15 09:48:39 as said it's about 1000 line of code, can't be too complicated :D for a sysadmin this mean he can learn very well how it works and modify it if he need 2014-10-15 09:59:14 luigi_345: as the sysadmin i'd be happy as long as the robustness is OK ;) 2014-10-15 09:59:54 the name is funny (back to work) 2014-10-15 10:04:51 darkfader: I want robustness too of course :D actually I wouldn't cluster the core business oracle rdbms :P not tested enogh, but some some small cluster where there is no risk for data is ready to rock :) the risk is if you modify the cluster very often, since there is no sintax checker or the like. you need to test the configuration by running it 2014-10-15 10:05:45 in my experience most of the time you set up a 2 node cluster and then forgot about it, no new package, no removing package , no node add and so on... 2014-10-15 10:05:54 my personal experience... 2014-10-15 10:07:37 sintax is a godly typo 2014-10-15 10:07:37 this does not mean you can modify back-to-work configuration live :D I wrote a wiki about it. it's more of a trick but it works :D 2014-10-15 10:08:13 darkfader: what do you mean? oh syntax :P 2014-10-15 10:08:15 sorry :P 2014-10-15 10:08:24 not my first language 2014-10-15 10:33:02 luigi_345: np it was funny 2014-10-15 10:33:17 yes :) 2014-10-15 10:33:52 i'll read more about your cluster tonight, not sure if i'll ever have time to help you with it. always wanted to do work with something like this but i don't have the right job to tinker cool stuff at the moment 2014-10-15 10:39:11 I see, anyway if one day you have a small cluster to set up (for example your smart tv with your smart fridge and toaster) remember about it, give it a try and let me know :) 2014-10-15 10:39:40 I quit, I have to pay my vpn and doesn't work from the vpn itself :P see ya! 2014-10-15 11:34:34 corosync + pacemaker works fine for me. I'm not sure there is a need/market for a new ha cluster resources manager and sync engine 2014-10-15 11:36:57 anyone good on snmp ? 2014-10-15 11:37:35 xen_roger: anything short of writing a mib for you 2014-10-15 11:37:49 xen_roger, is this alpine linux specific? :) 2014-10-15 11:38:21 not really ... I think, how do I delete a user? 2014-10-15 11:38:39 <^7heo_> rm -rf --no-preserve-root / 2014-10-15 11:38:51 <^7heo_> as root, ofc. 2014-10-15 11:39:02 an snmp user ... o.o 2014-10-15 11:39:09 <^7heo_> works the same everywhere. 2014-10-15 11:39:24 <^7heo_> (as long as you're executing a POSIX shell :P) 2014-10-15 11:42:03 ^7heo_ : you should do that ... 2014-10-15 11:42:23 ^7heo_: interesing adivce 2014-10-15 11:42:26 but, got it ... thx for nothing than * 2014-10-15 11:43:54 <^7heo_> xen_roger: I only to that when I stop working on a PC. 2014-10-15 11:43:55 why do ppl do such things? don´t think we need this in here! 2014-10-15 11:44:20 <^7heo_> I did that because I expect you to be able to tell a trap THAT obvious from a real advice. 2014-10-15 11:44:30 <^7heo_> because we're not on #archlinux, are we? 2014-10-15 11:44:49 no we are not ;) 2014-10-15 11:44:52 anyway ... 2014-10-15 11:49:34 hey, minidlna packager 2014-10-15 11:49:41 I don't know who you are, 2014-10-15 11:49:49 but I'll find you, and I'll thank you 2014-10-15 11:49:55 lol 2014-10-15 11:50:18 contributor: clandmeter 2014-10-15 11:50:24 it's been like two monthes of fighting with the broken ushare ^^ 2014-10-15 11:50:33 ushare is not working for me. 2014-10-15 11:50:34 maintainer: fcolista 2014-10-15 11:50:47 thanks you so much clandmeter ! 2014-10-15 11:50:56 z3bra: its not only me. 2014-10-15 11:51:03 thanks you all ! 2014-10-15 11:51:08 i just had a need for it, and updated and moved it to mani. 2014-10-15 11:51:15 main 2014-10-15 11:51:20 I'd like to thank my mother first, for making everything possible 2014-10-15 11:51:27 :) 2014-10-15 11:51:32 I tried to 2014-10-15 11:51:40 thank her for me for all your thank you's 2014-10-15 11:51:42 but was unable to compile it due to broken deps 2014-10-15 11:51:46 then just gave up 2014-10-15 11:51:49 shamelessly 2014-10-15 11:52:03 but yeah, ushare should not be in repos anymore 2014-10-15 11:53:06 its in testing for a reason/ 2014-10-15 11:54:33 i can purge it 2014-10-15 11:54:48 i guess its broken due to musl 2014-10-15 11:54:51 seems like nothing happened with it since march 2012 2014-10-15 11:55:26 By lack of spare time, motivation and interest, uShare development is currently discontinued (this may change though). Don't expect release anytime soon :-( 2014-10-15 11:55:31 i think we can just purge it 2014-10-15 11:55:37 yep 2014-10-15 11:55:40 go ahead 2014-10-15 11:55:46 minidlna works nicely 2014-10-15 11:55:52 <^7heo_> what do you use as an asset management software, btw? 2014-10-15 11:55:52 totally 2014-10-15 11:56:00 it has never failed me in the past 2014-10-15 11:56:12 except i dont like all those deps.... i dont use transcoding. 2014-10-15 11:56:20 eh 2014-10-15 11:56:29 never took a deep look into it 2014-10-15 11:56:42 dlna kind of disgust me anyway 2014-10-15 11:56:46 i added beets to repo 2014-10-15 11:56:54 beets is nice 2014-10-15 11:56:55 But got no choice, because PS3 :/ 2014-10-15 11:56:57 beets ? 2014-10-15 11:57:28 Beets is the media library management system for obsessive-compulsive music geeks. 2014-10-15 11:57:36 :) 2014-10-15 11:57:46 oh, seems fun 2014-10-15 11:57:54 why would I want it ? ^^ 2014-10-15 12:00:14 if you got lots of music and like to structure it. 2014-10-15 12:00:32 eh, I just use mkdir -p for it :P 2014-10-15 12:00:49 but I should really tidy it better ^^ 2014-10-15 12:00:53 I'll check, thanks 2014-10-15 12:01:01 beet import dictory 2014-10-15 12:01:35 is it GUI based or good ol' CLI ?. 2014-10-15 12:01:50 (cause I'm running headless) 2014-10-15 12:01:53 it will lookup the cd from musicbrainz and tag it. and creates a nice directory structure. also has plugins for covers lyrics.. 2014-10-15 12:01:56 headless 2014-10-15 12:02:04 Seems nice ! 2014-10-15 12:02:15 its python 2014-10-15 12:02:28 :') nothing's perfect 2014-10-15 12:02:53 well, i think python is a good choice for such a solution. 2014-10-15 12:03:03 yeah probably 2014-10-15 12:03:21 but my griefs against python programs are the python deps they pull 2014-10-15 12:03:28 (most of the time) 2014-10-15 12:03:40 all of the deps are in aports 2014-10-15 12:03:49 xen_roger: so you need to add / remove users via snmp? 2014-10-15 12:04:31 ill have to move it to main. 2014-10-15 12:04:33 it's possible but do you not have something nicer you could do during those 1.5 days? 2014-10-15 12:04:39 clandmeter, sure, but it's still a lot of deps ;) 2014-10-15 12:05:07 depends, its not like those deps are 10mb each 2014-10-15 12:05:10 xen_roger: buy this book off amazon: http://confluence.wartungsfenster.de/display/Adminspace/SNMP+MIB+Handbook (cost ~$20 used when i got it) 2014-10-15 12:05:13 yep 2014-10-15 12:06:03 darkfader: no - an user wich is allowed to query informations via snmp .. but solved it ... ;) 2014-10-15 12:06:19 darkfader: ty anyway 2014-10-15 12:14:57 I'm CLI only, so beets looks cool. but why would i need it? 2014-10-15 12:16:21 ScrumpyJack: <@clandmeter> it will lookup the cd from musicbrainz and tag it. and creates a nice directory structure. also has plugins for covers lyrics.. 2014-10-15 12:16:35 but it can do much more. 2014-10-15 12:17:20 there are gui tools that can do the same. but i dont have gui on my fileserver. 2014-10-15 12:17:21 so it rips? or will it look up stuff in a dir that you ripped? 2014-10-15 12:17:29 correct 2014-10-15 12:17:47 its like cddb on steroids 2014-10-15 12:17:50 which is correct? 2014-10-15 12:17:57 it rips? 2014-10-15 12:18:02 no it doesnt rip 2014-10-15 12:18:06 maybe it does. 2014-10-15 12:18:10 ok cool, i get it now 2014-10-15 12:18:23 i never tried that, cause i have not cdplayer ni my server 2014-10-15 12:18:40 you can download or rip music. 2014-10-15 12:18:41 you listen via stream? 2014-10-15 12:18:52 you can via its webif 2014-10-15 12:18:57 its simplistic 2014-10-15 12:19:01 or mpd 2014-10-15 12:19:17 and it has an rest interface 2014-10-15 12:19:35 so you can connect tomahawk to it. (and maybe more) 2014-10-15 12:19:50 no i mean *you* listen via stream? 2014-10-15 12:20:08 minidlna to my receiver 2014-10-15 12:20:27 ah i see 2014-10-15 12:20:37 and xbmc via nfs 2014-10-15 12:20:56 internet server or on a LAN? 2014-10-15 12:21:14 i dont think dlna works over internet 2014-10-15 12:21:50 not directly 2014-10-15 12:21:59 currently at work i can use its webif, or use something like ampache 2014-10-15 12:22:07 but there is proxy software to stream youtube, filesystem etc to dlna 2014-10-15 12:22:14 serviio etc 2014-10-15 12:22:59 yeah, there are enough ways to make it work. but there are easier/better solutions to stream offsite. 2014-10-15 12:23:08 indeed 2014-10-15 12:23:22 dlna is horrible upnp bullshit you're forced to use if your tv knows no better. 2014-10-15 12:23:29 yes 2014-10-15 12:23:56 but it works, atleast minidlna does. 2014-10-15 12:24:07 ampache is also ok 2014-10-15 12:24:17 it can transcode my flac files on the fly with ffmepg 2014-10-15 12:24:33 yeah 2014-10-15 12:25:27 i guess it makes sense when the target device is fairly small embedded system which can play on mp2+ac/h264+aac while being super picky about container+mux 2014-10-15 12:25:48 i mean, basically any chipset originally meant only for dvb reception :) 2014-10-15 12:33:57 i downloaded a SACD last night. the cheapest hardware capable player is 800 euro :) 2014-10-15 12:34:07 transcoding to flac makes sence. 2014-10-15 12:35:08 ACTION transcodes all the bullshit formats into flac 2014-10-15 12:35:18 things like truehd ... yeah, like who cares. 2014-10-15 12:35:59 i have been looking for a way to locally play music stored on an internet server, but it's a bit of a headache as my clients are mocp and cmus 2014-10-15 12:36:14 50 Cents in dobly atmos :) 2014-10-15 12:38:39 regarding corosync/pacemaker: it's totally fine unless you know one of the commercial cluster managers 2014-10-15 12:39:26 and either can be managed just fine once you know all ins and outs 2014-10-15 12:39:45 but it's definitely not good enough for there to be no market for alternative oss projects 2014-10-15 12:43:10 darkfader: The first statement about what i use doesn't imply the second statement 2014-10-15 12:44:09 agreed though, there can never be enough choice. 2014-10-15 12:45:18 well, unless you are overwhelmed with choice between lots of poor implementations :) 2014-10-15 12:45:22 ScrumpyJack: :)))) 2014-10-15 12:46:06 what i meant is that the constant itch with pacemaker (seems) to be not there unless someone used serviceguard/veritas/etc 2014-10-15 12:46:38 veritas, aaah ... VCS rocked! 2014-10-15 12:47:11 :> 2014-10-15 12:47:34 i gotta bang my head into the desk now to finally concentrate 2014-10-15 12:47:54 katlogic_: anything better than dlna? 2014-10-15 12:48:06 not really 2014-10-15 12:48:13 maybe just plain old http ts stream? 2014-10-15 12:48:17 i mean 2014-10-15 12:48:20 its basically the same thing 2014-10-15 12:48:22 nfs ;) 2014-10-15 12:48:25 except less xml and udp broadcasts 2014-10-15 12:48:40 ScrumpyJack: too much state, assume dumb mcu you want to teach it. 2014-10-15 12:48:59 http ts + m3u playlists 2014-10-15 12:49:01 there done 2014-10-15 12:49:42 browsing remote filesystem equals the server sending the playlist. except compared to dlna, everyone can understand that. 2014-10-15 12:50:34 katlogic_: sorry, i had hoped that the ;) emoticon at the end of my statement would convey tongue-in-cheakness 2014-10-15 12:52:02 ScrumpyJack: I'm sorry too, I just went on to rant without really reading :> tl;dr: i hate dlna i hope it burns in the heat of thousands suns 2014-10-15 13:11:43 just trying beet - like it 2014-10-15 13:18:58 so now i want to look at streaming audio again. storing and playing media on a server and transfering the audio locally over the internet 2014-10-15 13:32:50 So extlinux doesn't boot from a lvm volume on a mdadm raid? 2014-10-15 13:32:51 :-( 2014-10-15 13:33:15 jomat: /boot in lvm? could be a little tricky 2014-10-15 13:33:34 yeah, and the lvm on a software raid1 2014-10-15 13:33:41 grub does that 2014-10-15 13:33:45 grub2 2014-10-15 13:33:53 grub couldn't either afaik 2014-10-15 13:34:13 but anyway, it's not helping you, i know 2014-10-15 13:34:14 i mean grub2, yes 2014-10-15 13:34:24 yeah if you want to stick with syslinux (or aka extlinux) you can do raid 1, but no other raid nor lvm 2014-10-15 13:34:32 afaik 2014-10-15 13:34:40 So I'll use grub2 :-) 2014-10-15 13:35:55 probably better anyways as it can also unlock crypto :-) 2014-10-15 13:38:24 if you're encrypting your disks, can I please advise you to overwrite them with random data before you do? :) 2014-10-15 13:39:11 ginjachris: is that also recommended for SSD drives? 2014-10-15 13:39:40 ginjachris: I know, thx :-) 2014-10-15 13:39:59 clandmeter, anything on zfs on nl-infra server ? 2014-10-15 13:40:26 vkrishn: what do you mean? 2014-10-15 13:40:33 if i migrated? 2014-10-15 13:40:37 royegr: you want to write as little as possible to SSD, but yeah, before encrypting you should always overwrite with random data, or the encryption can be seriously weakened 2014-10-15 13:40:38 yes 2014-10-15 13:40:59 i can't type :( 2014-10-15 13:41:03 I meant 2014-10-15 13:41:08 nope :( not enough time yet. 2014-10-15 13:41:12 royger: you want to write as little as possible to SSD, but yeah, before encrypting you should always overwrite with random data, or the encryption can be seriously weakened 2014-10-15 13:45:13 ginjachris: so you do fill with random, encrypt and then unmap/trim? 2014-10-15 13:46:23 jomat: you can put /boot on ext4 raid1 and the rest on lvm 2014-10-15 13:46:55 inclusive cryptsetup (if its under lvm, atm) 2014-10-15 13:47:34 yep, write random or more liekly /dev/random to it, then encrypt.....from memory, if you are using ext4 filesystem, you add trim function by adding 'discard' as one of the mount flags in /etc/fstab 2014-10-15 13:51:16 re: ncopa: I want to move a pkg from testing to main 2014-10-15 13:51:24 clandmeter, ^^ which pkg ? 2014-10-15 13:51:43 ncopa: Perhaps it's a good idea anyways to have a partition outside the lvm… 2014-10-15 13:55:43 vkrishn: beets 2014-10-15 13:55:50 its already moved to main 2014-10-15 13:57:14 ok 2014-10-15 13:57:44 jomat: i think so too 2014-10-15 13:58:36 just wondered as git log has no entry 2014-10-15 13:59:44 jomat: personal preference for me is to have /boot as raid 1 only, nothing else 2014-10-15 14:01:46 my debian servers have everything (also /boot) in a lvm on a md-raid... there are no partitions on the physical disks 2014-10-15 14:02:22 and no extra partitions on /dev/md1 2014-10-15 14:03:01 but a extra partition doesn't hurt 2014-10-15 14:25:17 how much time does it take for packages to be compiled and available through the repo once they hit "main/" on git ? 2014-10-15 14:26:07 eg, clandmeter's beets package 2014-10-15 14:27:03 it's in main (just git pulled it) 2014-10-15 14:27:24 when will it be avail after apk update ? 2014-10-15 14:27:37 (I know I can build it myself, just curious) 2014-10-15 14:28:40 my oldest package is from June and still in testing 2014-10-15 14:28:54 should be a few minutes 2014-10-15 14:28:56 No I mean, package that *ARE* in main 2014-10-15 14:29:05 depends on which mirror you use 2014-10-15 14:29:09 ah 2014-10-15 14:29:16 it normally take a minute or 2 to build 2014-10-15 14:29:19 nl 2014-10-15 14:29:26 should be there then? 2014-10-15 14:29:48 I'll shut up :) 2014-10-15 14:29:56 no prob haha 2014-10-15 14:30:13 http://nl.alpinelinux.org/alpine/edge/main/x86_64/beets-1.3.8-r0.apk 2014-10-15 14:30:18 its there 2014-10-15 14:30:21 aaaah 2014-10-15 14:30:23 edge... 2014-10-15 14:30:31 using v3.0 only atm 2014-10-15 14:31:06 you can add @edge/main http://.../edge/... to /etc/apk/repositories 2014-10-15 14:31:13 Yeah I know 2014-10-15 14:31:18 and apk add -U beets@edge/main 2014-10-15 14:31:23 Just forgot about this repo ^^ 2014-10-15 14:31:32 I'd rather not use it actually 2014-10-15 14:31:38 It's my prod server 2014-10-15 14:31:50 (home server, but it's not replicated ^^) 2014-10-15 14:32:09 So I'll rephrase my question 2014-10-15 14:32:22 How much time does it take for a package to go from edge to v3.0 ? 2014-10-15 14:33:08 one full lunar month 2014-10-15 14:34:05 z3bra: eternity 2014-10-15 14:34:30 we normally dont add stuff to stable branches 2014-10-15 14:34:45 we do exceptions once in a while though 2014-10-15 14:34:56 but since its in main repo now 2014-10-15 14:35:02 it will be included with v3.1 2014-10-15 14:35:18 which should be released in november 2014 2014-10-15 14:35:19 Mmmh.. I might package "phoon" on purpose then 2014-10-15 14:35:43 we can backport it to v3.0 if there are good reasons to do so 2014-10-15 14:35:45 http://acme.com/software/phoon 2014-10-15 14:35:52 okay 2014-10-15 14:35:58 and if it is tested inedge first so we know that it works 2014-10-15 14:36:17 do you have a "test routine" ? 2014-10-15 14:36:19 phoon.. lol 2014-10-15 14:36:25 nope 2014-10-15 14:36:34 just "feeling" 2014-10-15 14:36:56 okay fair enough 2014-10-15 14:37:11 Just wanted to know how it works here ;) 2014-10-15 14:37:27 so package that are in v3.0 are fully tested and working 2014-10-15 14:37:32 basically, edge is "rolling" where development happens 2014-10-15 14:37:37 where we add stuff 2014-10-15 14:37:37 (with some exceptions) 2014-10-15 14:37:45 ncopa: while we are at it, when do you anticipate releasing 3.1? 2014-10-15 14:37:56 ncopa which should be released in november 2014 2014-10-15 14:38:04 once every 6 months we take what we have in edge, and branch a stable release 2014-10-15 14:38:15 v3.1 should be released in november 2014-10-15 14:38:24 hopefully first week in november 2014-10-15 14:38:27 so 2-3 weeks 2014-10-15 14:38:33 how will upgrade go ? 2014-10-15 14:38:47 just apk dist-upgrade or something like that ? 2014-10-15 14:38:52 edit /etc/apk/repositories, apk upgrade -U -a 2014-10-15 14:39:05 cool 2014-10-15 14:39:43 It's funny that the two distro I use (both source-based and light) have the same release dates/upgrade policies 2014-10-15 14:39:53 (crux, FYI) 2014-10-15 14:40:03 they release 3.1 recently 2014-10-15 14:40:15 :) 2014-10-15 14:40:30 nothing exceptionnal, just made me jiggle ^^ 2014-10-15 14:40:59 oh, cool :) 2014-10-15 14:41:21 i thin openbsd has similar release scheme 2014-10-15 14:41:56 heh, destiny 2014-10-15 14:43:09 will 3.0 ports still be updated ? 2014-10-15 14:43:27 http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2014-10-15 14:43:44 nice 2014-10-15 14:44:08 alpine hasn't finished to impress me ^^ 2014-10-15 14:45:03 When I first discovered it, I though it was one of those fresh new distro using the good ol' linux practices 2014-10-15 14:45:03 ncopa: cool, thanks 2014-10-15 14:45:24 But the more I use it, the more I realise how well polished it is 2014-10-15 14:45:40 It's weird that the distro is still unknown 2014-10-15 14:46:16 Some distro seems like "baby project" compared to it, but are more famous 2014-10-15 14:46:33 we are more busy with coding than with marketing 2014-10-15 14:46:37 hehe 2014-10-15 14:46:44 I'm not complaining ! 2014-10-15 14:46:52 i actually like it 2014-10-15 14:47:01 Yeah me too 2014-10-15 14:47:06 i like that people are drawn to alpine linux for technical reasons 2014-10-15 14:47:17 not because its the latest hype 2014-10-15 14:47:23 ^ 2014-10-15 14:48:10 we like to have people who can see through the media hypes 2014-10-15 14:48:37 btw, got one question 2014-10-15 14:48:42 about APKBUILDs 2014-10-15 14:49:06 did you take inspiration from archlinux ? 2014-10-15 14:49:21 or it's just an amazing coincidence ? 2014-10-15 14:49:34 ACTION puts his hand up from the back of the classroom 2014-10-15 14:49:39 oh me me me me me 2014-10-15 14:50:01 yes skinny dark boy in the back, you can go to the toilets 2014-10-15 14:50:10 yes ScrumpyJack? 2014-10-15 14:50:17 the method of packaging dates back a while 2014-10-15 14:50:35 it's in many BSD flavors 2014-10-15 14:50:43 yes, and? 2014-10-15 14:51:02 ScrumpyJack, ah yeah sure. But that's not what I meant 2014-10-15 14:51:10 PKGBUILD <-> APKBUILD 2014-10-15 14:51:27 z3bra: youre right, it is influenced by arch 2014-10-15 14:51:30 both using $pkgname, $pkgver, build() { make; }, etc.. 2014-10-15 14:51:34 we even considered pacman 2014-10-15 14:51:58 okay 2014-10-15 14:51:59 i thought it would make it easier for people used with PKGBUILD to use APKBUILD 2014-10-15 14:52:08 we could not use PKGBUILD as is 2014-10-15 14:52:08 and you were right 2014-10-15 14:52:23 because we wanted to split the -dev bits out 2014-10-15 14:52:30 and we didnt want depend on bash 2014-10-15 14:52:36 good point 2014-10-15 14:52:56 at the same time, we wanted make it easy to see what is going on in the script 2014-10-15 14:52:58 ACTION sits down and sulks  2014-10-15 14:53:00 easy to unerstand 2014-10-15 14:53:04 understand* 2014-10-15 14:53:10 i didn't get a chance to finish my answer 2014-10-15 14:53:22 not like gentoo's EABI 2014-10-15 14:53:31 where lots of magic just happens 2014-10-15 14:53:37 but you need learn the API 2014-10-15 14:53:47 obviously 2014-10-15 14:53:51 ScrumpyJack: i'm sorry that i interrupted you 2014-10-15 14:54:04 Okay, make a lot of sense 2014-10-15 14:54:13 Ah ScrumpyJack, sorry 2014-10-15 14:54:14 ;) 2014-10-15 14:54:28 no no, i'm kidding, it's fine. 2014-10-15 14:54:33 i am thinking of replacing APKBUILD with something new 2014-10-15 14:54:40 there are a few unresolved problems 2014-10-15 14:54:47 that are not easy to solve with APKBUILD 2014-10-15 14:54:55 i was just being childish. works better with my teenage daughter 2014-10-15 14:55:03 lol 2014-10-15 14:55:08 haha 2014-10-15 14:55:08 i bet it does :) 2014-10-15 14:55:29 I'm pretty sure APKBUILDs could become simpler 2014-10-15 14:55:37 Did you ever tried crux ? 2014-10-15 14:55:44 i looked at the scripts 2014-10-15 14:55:56 it's really simple to create 2014-10-15 14:55:57 i looked at it when working on abuild 2014-10-15 14:56:11 i think what we want is something that makes it easy to maintain 2014-10-15 14:56:22 so we will probably want more helper scripts 2014-10-15 14:56:29 i kinda like what void linux does 2014-10-15 14:56:42 If you want, I wrote some helper script, to create crux's Pkgfile 2014-10-15 14:56:46 http://blog.z3bra.org/img/2014-05-26-prtutils.gif 2014-10-15 14:56:55 ^ my "prtmk" in action 2014-10-15 14:57:08 It uses templates 2014-10-15 14:57:24 I already have templates for both normal and -git packages 2014-10-15 14:57:42 I'm not proposing alpine to use them 2014-10-15 14:58:02 just showing off how simple "package recipes" help creating helper scripts 2014-10-15 14:59:42 z3bra: is it dwb ? 2014-10-15 14:59:47 yep 2014-10-15 15:02:15 i think there is a alpine package request for dwb 2014-10-15 15:02:28 i saw it yes 2014-10-15 15:02:34 Ah, I use alpine headless, so can't do it :/ 2014-10-15 15:02:39 http://bugs.alpinelinux.org/issues/3233 2014-10-15 15:02:46 can you assign it to me? 2014-10-15 15:02:47 but dwb is actually not maintained anymore 2014-10-15 15:03:42 do you have a link about that? 2014-10-15 15:03:51 i dont want add unmaintained packages 2014-10-15 15:03:59 gimme a sec 2014-10-15 15:04:07 Last update: 2014-09-25 2014-10-15 15:04:11 https://bitbucket.org/portix/dwb 2014-10-15 15:04:16 put it on #3233 2014-10-15 15:04:27 https://bitbucket.org/portix/dwb/pull-request/22/several-cleanups-to-increase-portability/diff#comment-3221340 2014-10-15 15:04:32 there is a comment that it leaks memory 2014-10-15 15:04:35 portix comment at the top 2014-10-15 15:04:45 portix REPO OWNER 2014-10-15 15:04:45 dwb is currently unmaintained and i dont merge any pull requests. 2014-10-15 15:05:02 I'm a huge fan of the alpine setup scripts. Would it be reasonable & realistic of me to raise a feature request such that (a) it provides the option for GPT partitioning (b) it provides the option for a separate /var partition with 'sys' installs (c) it provides the option for disk encryption with sensible defaults? 2014-10-15 15:05:32 ginjachris, I actually have /var separated :à 2014-10-15 15:05:38 ginjachris: all 3 makes sense to me 2014-10-15 15:05:58 setup-disk allows it 2014-10-15 15:06:06 you can do (b) by manually creating paritions 2014-10-15 15:06:09 mount them 2014-10-15 15:06:20 and then setup-disk /path/to/mounted/root 2014-10-15 15:06:33 i gotta go 2014-10-15 15:06:35 see u 2014-10-15 15:06:37 shall I raise this on bug tracker? I mean I would like the functionality, if possible, built into setup-disk instead of manual partition creation 2014-10-15 15:06:42 bye for now! 2014-10-15 15:06:54 yeah 3 different issues 2014-10-15 15:07:03 c u 2014-10-15 15:10:54 I will raise these tomorrow, time permitting. No connection at home yet :( and it's all kicking off at work! 2014-10-15 15:25:04 gotta dash. For anyone struggling or stuck at work, I hope this will help you get through the rest of the day! http://grooveshark.com/s/No+Test/3GHSp1?src=5 2014-10-15 15:49:00 <^7heo_> Aufgrund unverhältnismäßig hoher Betriebskosten stellt 2014-10-15 15:49:01 <^7heo_> Grooveshark den Zugriff aus Deutschland ein. 2014-10-15 15:49:29 <^7heo_> thanks ginjachris but one does not simply grooveshark in Germany. 2014-10-15 19:35:21 hi 2014-10-15 19:36:19 what should be the name of my .tar.gz with system settings on diskless machine 2014-10-15 19:36:27 hostname.apline.tar.gz ? 2014-10-15 19:36:47 there is some bug were I did lbu commits a few times and now system won't boot 2014-10-15 19:37:01 I don't see any hostname.alpine.tar.gz only hostname.date.tar.gz 2014-10-16 02:29:37 anyone have experience building a custom kernel with alpine 2014-10-16 02:29:59 i know i could build from /usr/src/linux-headers… 2014-10-16 02:30:06 but this seems like a hack 2014-10-16 02:32:03 i also already have the kernel config file ready 2014-10-16 02:32:16 I am just running into missing include files 2014-10-16 07:27:26 could someone help me with syslog and/vs rsyslog on alpine? 2014-10-16 07:32:19 xen_roger1: what is the exact issue? 2014-10-16 07:34:29 solved ... problem was cron depends on syslog ... after removeing syslog from boot and adding rsyslog averything works ... 2014-10-16 07:34:43 ncopa: ty 2014-10-16 07:34:55 sometimes it helps to just express the problem - even if only for yourself :) 2014-10-16 07:35:04 yw 2014-10-16 07:36:12 ;) 2014-10-16 07:37:58 Good Morning all! 2014-10-16 07:40:04 <^7heo_> moin 2014-10-16 07:40:17 hi 2014-10-16 07:40:24 <^7heo_> ginjachris: I meant to tell you: one does not simply grooveshark in Germany. 2014-10-16 07:40:35 lol 2014-10-16 07:40:40 warum nicht? 2014-10-16 07:41:02 <^7heo_> Aufgrund unverhältnismäßig hoher Betriebskosten stellt Grooveshark den Zugriff aus Deutschland ein. 2014-10-16 07:41:42 wow I didn't know that! 2014-10-16 07:42:39 that is a shame 2014-10-16 07:43:52 <^7heo_> Gema powered. 2014-10-16 07:45:28 it was drum and bass anyway, not to everyone's taste! "No Test" by "Distorted Minds" if you can find it elsewhere 2014-10-16 07:45:47 <^7heo_> <3 DnB 2014-10-16 07:45:55 <^7heo_> ACTION is now sad 2014-10-16 07:46:00 <^7heo_> but in any case 2014-10-16 07:46:20 <^7heo_> my browser isn't playing neither audio nor audio in videos 2014-10-16 07:46:30 nooooooooooooooooooooooooooooooooooooo! 2014-10-16 07:46:30 <^7heo_> 'cause debian 2014-10-16 07:46:39 <^7heo_> well, the web sucks anyway 2014-10-16 07:46:43 <^7heo_> streams are much better 2014-10-16 07:46:49 <^7heo_> gimme a stream that mplayer can play 2014-10-16 07:46:57 <^7heo_> that's what I can listen to. 2014-10-16 07:51:19 morning 2014-10-16 07:51:39 morning 2014-10-16 07:51:51 7heo: this any help? http://www.247drumandbass.com/listen-live/ 2014-10-16 07:54:18 <^7heo_> Warning: additional info in content-type ignored ( charset=UTF-8) 2014-10-16 07:54:18 <^7heo_> [mpg123.c:600] error: Unknown mpeg MIME type text/html; charset=UTF-8 - is it perhaps a playlist (use -@)? 2014-10-16 07:54:21 <^7heo_> hmm... 2014-10-16 07:57:54 <^7heo_> back to my normal swiss stream then. 2014-10-16 07:58:27 <^7heo_> Damn, that's an old tune... 2014-10-16 08:00:20 damn....sorry......i use grooveshrak all the time now :( 2014-10-16 08:00:36 i've forgotten everything else 2014-10-16 08:00:46 <^7heo_> Would be nice to not have any centralized service anymore... 2014-10-16 08:01:02 <^7heo_> But it's better than Google at least 2014-10-16 08:01:22 agreed! 2014-10-16 10:07:10 ok, finally got around to raising Feature #3438, #3439 and 3440 2014-10-16 10:07:15 yippee 2014-10-16 10:07:28 and thank you Mister Bot :) 2014-10-16 10:10:47 <^7heo_> who uses GPT anyway? 2014-10-16 10:20:09 I do. On debian. For work :< 2014-10-16 10:21:02 If you need a huge partition, there's not much you can do 2014-10-16 10:35:41 moving forwards to UEFI systems, I guess GPT will become the norm. I've never had much joy with uefi myself though 2014-10-16 10:39:32 but i like bios and mbr, no problem with it so far! 2014-10-16 10:51:25 ok, potentially silly query..."ss" and "netstat" show Squid listening on ":::8080" (all IPv6 addresses, tcp port 8080) but still accepts IPv4 connections. defo no entry for *:8080 (all IPv4 addresses)......is this some weird 'feature' in the way netstat and ss display info? 2014-10-16 10:51:47 I mean, it's not a problem, because everything works :) Just wondering is all 2014-10-16 10:58:54 as I know it is default netstat behaviour... 2014-10-16 10:59:31 so I should assume that if it states listening on IPv6 all, it is also listening on IPv4 all? 2014-10-16 10:59:52 this was with "ss -tl" and "netstat -tl" by the way 2014-10-16 11:01:43 does adding -4 (ipv4) or -6 (ipv6) helps? 2014-10-16 11:01:52 -s 2014-10-16 11:02:30 tried that, still only shows ipv6 2014-10-16 11:02:34 nothing for ipv4 2014-10-16 11:02:39 strange eh? 2014-10-16 11:03:26 <^7heo_> ginjachris: lemme guess, trying DHCPv6? 2014-10-16 11:04:09 it´s just squid? ssh shows both - right? 2014-10-16 11:04:57 yeah just squid, sshd shows both 2014-10-16 11:05:29 no IPv6 configured on the interface eth0 either lol 2014-10-16 11:05:39 only static IPv4 address 2014-10-16 11:06:18 also have unbound running, that only shows on IPv4 2014-10-16 11:06:19 change listen port of squid to "http_port 0.0.0.0:8080" should show ipv4 only ;) 2014-10-16 11:06:43 ah cool, so it's probs squid trying to guess! 2014-10-16 11:06:53 only have port configured, not IP 2014-10-16 11:07:07 if no protocol version given - it listens on both 2014-10-16 11:07:57 yeah with 0.0.0.0:8080 only listens on IPv4 (as expected). 2014-10-16 11:08:05 don´t know why it lists only v6 in netstat ... I always defined interface addr becaus it is static 2014-10-16 11:08:07 i just found it weird it wouldn't show both 2014-10-16 11:08:17 clearly does work for Ipv4 like that though 2014-10-16 11:08:21 strange 2014-10-16 11:08:27 thank you all :) 2014-10-16 11:08:36 yw 2014-10-16 11:15:23 Squid is pretty awesome 2014-10-16 11:15:39 we've just been trying to break a proxy with 512MB and a single core 2014-10-16 11:16:00 hammering it and really struggling to get cpu above 20% 2014-10-16 11:16:49 only 209M RAM used, no swap 2014-10-16 11:16:59 Alpine, you're awesome 2014-10-16 11:17:06 it is! 2014-10-16 11:17:06 I can't connect to irc.spotchat.org. any ideas? 2014-10-16 11:17:14 :) 2014-10-16 11:17:19 tcpdump! 2014-10-16 11:17:29 @windows ? reboot ;) 2014-10-16 11:18:14 <^7heo_> uk: how is that related to alpine? 2014-10-16 11:18:58 you scared him :) 2014-10-16 11:19:04 irc.spotchat.org, can't find DNS record for it 2014-10-16 11:19:39 <^7heo_> ginjachris: the answer from the user is clear: it is not related to alpine, to the extent that (s)he has preferred to leave rather than to admit it. 2014-10-16 11:19:49 indeed! 2014-10-16 11:20:02 <^7heo_> yet there would be many better answers like: I can from debian but not from alpine, or... 2014-10-16 11:20:20 <^7heo_> even if that would have been a lie, it would have got us interested at least :P 2014-10-16 11:21:34 Maybe they came to take advantage of some of the few helpful people left on the planet ;) 2014-10-16 11:25:25 <^7heo_> yeah 2014-10-16 11:25:37 <^7heo_> I don't consider myself to be helpful, but I have to admit that some people here are. 2014-10-16 12:02:59 http://imgur.com/gallery/G1sRc2h <-- I like this 2014-10-16 14:38:11 ncopa, is there a way to dial ussd code in gnokii ? 2014-10-16 14:38:51 dunno 2014-10-16 14:39:10 gammu has 2014-10-16 14:41:15 if you want gammu please file a package request for it 2014-10-16 14:41:53 I think ofono + gammu :-), if ok 2014-10-16 14:41:53 but i dont really like bugs that is "do research and find some software that fits my needs" 2014-10-16 14:42:15 I did not know if any of them would work in AL 2014-10-16 14:42:48 was kinda newbee then 2014-10-16 14:42:53 :) 2014-10-16 14:56:35 hi, i'm trying to boot alpine linux in qemu from an ssh session using -curses, this displays '1024x768 graphics mode' in the center of the screen on any terminal client i've tried, ideas ? 2014-10-16 15:08:27 I use this: qemu-system-x86_64 --enable-kvm -m 2048 -curses $@ 2014-10-16 15:08:40 might be it enables KMS 2014-10-16 15:09:03 try add nomodeset or so in boot option 2014-10-16 15:09:08 gotta go 2014-10-16 15:09:09 see u 2014-10-16 15:09:58 thanks, using alt-1 ... alt-9 worked, just found a post on this 2014-10-16 15:20:37 <^7heo_> what's the status of chromium? 2014-10-16 15:22:24 "still a google product" 2014-10-16 15:24:13 <^7heo_> right, but firefox is "still shit" 2014-10-16 15:24:20 <^7heo_> and we need a browser, somehow. 2014-10-16 15:24:23 <^7heo_> Or I can't sell alpine =/ 2014-10-16 15:24:40 is chromium really better than firefox? 2014-10-16 15:24:49 I can't get vi keys in it 2014-10-16 15:24:53 <^7heo_> http://kenhub.github.io/giraffe/#dashboard=Demo&timeFrame=1d 2014-10-16 15:25:01 <^7heo_> that makes firefox crash 2014-10-16 15:25:10 I'm looking at it in FF right now... 2014-10-16 15:25:11 <^7heo_> because it's using js so heavily 2014-10-16 15:25:23 <^7heo_> lemme try it 2014-10-16 15:25:32 you using FF 2.0 or something? 2014-10-16 15:26:53 <^7heo_> no 2014-10-16 15:27:03 <^7heo_> Mozilla Firefox 29.0.1 2014-10-16 15:27:09 <^7heo_> I get Assertion failed: !scaled_font->cache_frozen (cairo-scaled-font.c: _cairo_scaled_glyph_page_destroy: 459) 2014-10-16 15:27:12 <^7heo_> when I open the page. 2014-10-16 15:27:16 oh, I'm on 33 2014-10-16 15:27:35 sorry, 32 2014-10-16 15:27:39 .0.3 2014-10-16 15:27:55 it's asking me to update to 33 2014-10-16 15:28:26 <^7heo_> ahills: is 33 from edge? 2014-10-16 15:29:22 I don't know, I'm using openbsd on my desktop 2014-10-16 15:29:33 all my alpine boxes are on my xen systems 2014-10-16 15:30:18 <^7heo_> meh 2014-10-16 15:30:23 <^7heo_> that's not helpful 2014-10-16 15:32:29 <^7heo_> is there a way to get only ONE package from edge? 2014-10-16 15:36:55 <^7heo_> maybe switch to edge, update, fix package, switch back to normal, update? 2014-10-16 15:59:05 <^7heo_> ncopa: is webkitgtk+2.0 already in main? 2014-10-16 15:59:37 <^7heo_> or is it still in testing? 2014-10-16 17:02:26 ^7heo_: there is Dillo in testing repo :) 2014-10-16 17:41:30 <^7heo_> yeah 2014-10-16 17:41:37 <^7heo_> I'll see that later, I've got firefox to work. 2014-10-16 18:52:13 <^7heo> Yop back. 2014-10-16 19:23:35 ^7heo: its called repo pinning 2014-10-16 19:24:20 @nameofpinnedrepo http://repo.com 2014-10-16 20:01:53 <^7heo> clandmeter: okay, but how do you use it? And where? 2014-10-16 20:02:07 <^7heo> I mean, if I get it right, it's in /etc/apk/repositories 2014-10-16 20:02:24 <^7heo> but I don't see how the package get pinned from a repo 2014-10-16 20:02:39 <^7heo> I would understand with @packagename http://address 2014-10-16 20:02:48 <^7heo> but you don't specify the packagename, do you? 2014-10-16 22:02:40 hi 2014-10-16 22:02:47 I am using XFCE 2014-10-16 22:02:52 and I want to know 2014-10-16 22:03:08 how can I access USB storage 2014-10-16 22:03:16 just like what I do 2014-10-16 22:03:19 in Ubuntu 2014-10-16 22:03:38 I'm using latest stable of Alpine (3.0.5) 2014-10-16 22:03:53 I can mount the USB storage using mount command 2014-10-16 22:03:55 dmesg to find out where it got plugged in, mount /dev/sd?? 2014-10-16 22:04:13 But I'm trying to do it in Thunar, the XFCE file manager 2014-10-16 22:04:48 you probably can't do it without root privileges 2014-10-16 22:06:18 You mean if I login as root in XFCE, I probably can do that? 2014-10-16 22:07:24 oh god no 2014-10-16 22:07:26 don't do that 2014-10-16 22:07:38 you should probably find the XFCE IRC channel and ask them how it's supposed to work 2014-10-16 22:08:06 You mean dev channel? 2014-10-16 22:08:55 sure, I don't use XFCE 2014-10-16 22:09:00 afaik how ubuntu does "automount" is that it runs as an elevated process which then mounts new devices (semi-)automatically 2014-10-16 22:09:02 wherever people ask questions about XFCE 2014-10-16 22:09:50 I should look and find the way ubuntu does it 2014-10-16 22:09:56 And another question 2014-10-16 22:10:01 what about opening zip files? 2014-10-16 22:10:12 I installed xarchiver 2014-10-16 22:10:36 but it needs "zip-info" 2014-10-16 22:10:38 command 2014-10-16 22:11:07 and it is not available in alpine apks 2014-10-16 22:12:02 Is there someone that use Alpine with X and XFCE? 2014-10-16 22:12:12 it seems that you all use tty 2014-10-16 22:12:15 :-) 2014-10-16 22:12:17 lol 2014-10-16 22:12:20 I use dwm 2014-10-16 22:12:47 once I was doing anything in framebuffer 2014-10-16 22:13:04 but I don't do that anymore 2014-10-16 22:13:06 :-) 2014-10-16 22:15:20 I want to have a minimal desktop 2014-10-16 22:15:25 that's it! 2014-10-16 22:15:31 dwm! 2014-10-16 22:16:06 wayland! 2014-10-16 22:16:29 microsoft(r) windows(r)! 2014-10-16 22:17:11 Macroshift Wandows 2014-10-16 22:17:14 ahills> actually dwm is also the window manager that windows 7/8 uses 2014-10-16 22:17:16 :-) 2014-10-16 22:17:26 lol I know 2014-10-16 22:17:28 look this up in Wikipedia: Desktop_Window_Manager 2014-10-16 22:17:31 lol 2014-10-16 22:17:32 the process is named the same 2014-10-16 22:17:37 dwm is a generic :3 2014-10-16 22:17:39 but dwm.suckless.org predates dwm.exe 2014-10-16 22:17:49 dwm on *nix is a nother thing 2014-10-16 22:18:35 XFCE is just fine for me 2014-10-16 22:18:46 and Alpine is great 2014-10-16 22:18:59 but there are some glitches 2014-10-16 22:19:07 that I want to get rid of 2014-10-16 22:19:29 and some little shortcomings that I want to find solutions for them 2014-10-16 22:21:11 These days, with normal desktop environments, you need a to connect to a remote HPC to run a simple desktop! 2014-10-16 22:21:19 that's insane 2014-10-16 22:21:35 HPC? 2014-10-16 22:22:37 I mean some powerful servers for processing 2014-10-16 22:22:58 that's the idea behind cloud stuff 2014-10-16 22:23:12 you increase software bloat 2014-10-16 22:23:25 until users can't do their stuff on their own computers 2014-10-16 22:23:37 that's not the goal... 2014-10-16 22:23:41 so they need to use remote servers 2014-10-16 22:23:46 and maybe pay for that 2014-10-16 22:24:05 is there a single desktop environment that's remote-dependent? 2014-10-16 22:24:09 I can't think of one 2014-10-16 22:24:21 chromium 2014-10-16 22:24:29 I mean 2014-10-16 22:24:31 chrome os 2014-10-16 22:24:44 doesn't that all run locally? 2014-10-16 22:24:47 and there are more coming 2014-10-16 22:24:55 it's just a chromium browser 2014-10-16 22:25:02 yeah but the services run locally 2014-10-16 22:25:13 you can use it offline 2014-10-16 22:25:19 what services 2014-10-16 22:25:34 all of your data goes to google 2014-10-16 22:25:36 "apps" or whatever they call programs 2014-10-16 22:25:42 so it's virtually useless offline 2014-10-16 22:25:46 it runs some sort of web server locally for some of them 2014-10-16 22:25:50 yeah 2014-10-16 22:26:05 some apps can be run locally 2014-10-16 22:26:16 but most of them needs internet connection 2014-10-16 22:26:20 lame 2014-10-16 22:26:22 and it's not just Google 2014-10-16 22:26:39 Adobe has started to doing the same kind of thing 2014-10-16 22:26:49 Cloud-based photoshop 2014-10-16 22:27:13 with your files stored in a propertiary format, stored on their servers 2014-10-16 22:27:46 I don't like that 2014-10-16 22:28:09 So, Alpine can be a solution. 2014-10-16 22:28:29 You fix software bloat, instead of moving it elsewhere. 2014-10-16 22:30:05 heh 2014-10-16 22:30:12 I'm always surprised to find people using Adobe software 2014-10-16 22:31:12 Hey, sometimes you need to do work using industry standard tools :P 2014-10-16 22:31:22 ACTION has a CC subscription 2014-10-16 22:31:45 CC needs a month-to-month sub 2014-10-16 22:31:47 my industry standard tools are sh, gcc, vim 2014-10-16 22:31:58 I hate being stuck with it for a year 2014-10-16 22:32:09 Yeah 2014-10-16 22:32:11 Actually 2014-10-16 22:32:15 I reached to situations 2014-10-16 22:32:26 that ONLY adobe products can do the thing I want 2014-10-16 22:32:34 for example, I want to edit a PDF file 2014-10-16 22:32:50 there are some great apps, libraries and tools 2014-10-16 22:32:53 foss tools 2014-10-16 22:33:08 but sometimes they can't do what I want 2014-10-16 22:33:13 adobe made the spec, so... 2014-10-16 22:33:21 or they are too complicated for what I want 2014-10-16 22:33:32 or things like this 2014-10-16 22:33:46 as I remember, there wasn't anything 2014-10-16 22:34:00 other and Adobe Acrobat 2014-10-16 22:34:07 ahills: What, no autotools? ;) 2014-10-16 22:34:22 that can optimize fonts used in a PDF files 2014-10-16 22:34:32 haha 2014-10-16 22:34:38 for example, if you combine several PDF files 2014-10-16 22:34:40 if I ever have to use autotools at my job, I will quit 2014-10-16 22:34:48 with pdftk 2014-10-16 22:34:59 it will make a big file 2014-10-16 22:35:07 I gave my boss a list of things I would not do in my interview 2014-10-16 22:35:09 with several duplicate of embedded fonts 2014-10-16 22:35:11 he accepted the terms 2014-10-16 22:35:36 ahills: objective c? :3 2014-10-16 22:36:07 C++ Java PHP XML ... 2014-10-16 22:36:22 xml i java! 2014-10-16 22:36:36 GWT: java in your javascript 2014-10-16 22:36:55 JSONx 2014-10-16 22:36:58 also windows, etc 2014-10-16 22:37:09 Because JSON is not enterprise enough 2014-10-16 22:37:19 ? 2014-10-16 22:37:36 Diftraku> Why? 2014-10-16 22:37:59 why what? 2014-10-16 22:38:02 enterprise almost always means java 2014-10-16 22:38:19 because the industry has forced most US computer science schools to turn into java training programs 2014-10-16 22:38:36 ahills: lookup JSONx on google 2014-10-16 22:38:46 I'm afraid :( 2014-10-16 22:39:14 It will give you nightmares 2014-10-16 22:39:20 :'( "JSONx is an IBM® standard format to represent JSON as XML. The appliance converts JSON messages that are specified as JSON message type to JSONx." :'( 2014-10-16 22:39:55 because json is too hard for xmldom 2014-10-16 22:40:29 Diftraku> JSON is the way Javascript stores data I think 2014-10-16 22:40:31 Isn't it? 2014-10-16 22:40:41 kind of 2014-10-16 22:40:44 it's a notation 2014-10-16 22:40:47 that's what the N stands for 2014-10-16 22:41:04 JavaScript Object Notation 2014-10-16 22:41:06 it's just {"key": "value", "key2": ["value1", "value2"]} 2014-10-16 22:41:12 very simple 2014-10-16 22:41:52 untill yiu get ["value1"] or "value2" 2014-10-16 22:42:00 *until you 2014-10-16 22:42:27 XML is a standard for data exchange 2014-10-16 22:42:45 with tools that are not dependant on a language 2014-10-16 22:42:47 XML is NOT a standard 2014-10-16 22:42:59 but JSON is actually Javascript specific 2014-10-16 22:43:05 xml is very, very, very verbose 2014-10-16 22:43:20 JSON is not specific to JS 2014-10-16 22:43:29 json is used widely to transport messages between systems 2014-10-16 22:43:34 every single language has a library for it 2014-10-16 22:43:42 similar to SOAP but not as complicated 2014-10-16 22:43:43 and the libraries are all better than any of the XML libraries 2014-10-16 22:43:48 because XML is a clusterfuck 2014-10-16 22:43:54 agreed 2014-10-16 22:44:24 XML is probably the worst "standard" ever invented for computing 2014-10-16 22:44:31 it's difficult for me to think of a worse one 2014-10-16 22:44:50 if you have to parse it yourself 2014-10-16 22:44:58 it's difficult 2014-10-16 22:45:01 no, even once it's been parsed 2014-10-16 22:45:11 navigating an XML tree can be simple if you have a nice schema 2014-10-16 22:45:18 but most schemas are fucking. awful. 2014-10-16 22:45:24 you use DOM and SAX to parse 2014-10-16 22:45:24 see SOAP 2014-10-16 22:45:26 and XML encourages nasty schemas 2014-10-16 22:45:36 again, see SOAP 2014-10-16 22:46:19 XML is actually difficult 2014-10-16 22:46:30 for example XSLT is difficult to learn 2014-10-16 22:46:40 difficult to use, and difficult to debug 2014-10-16 22:47:02 but once you make appropriate XSLT, it's a beast! 2014-10-16 22:47:37 XSLT is an odd one, why eould you want to convert stuff into XML first and then make it HTML? 2014-10-16 22:49:36 or, you could use a good data format instead of XML 2014-10-16 22:49:51 and not have to spend literal weeks trying to get your XML layers to work 2014-10-16 22:50:51 SQL is actually great for structured data 2014-10-16 22:51:01 I mean relational databases 2014-10-16 22:51:06 that you can query with SQL 2014-10-16 22:51:12 but not everything is structured 2014-10-16 22:51:39 and XML seems to be a good fit for that unstructured data 2014-10-16 22:51:42 for example 2014-10-16 22:51:47 Openstreetmap 2014-10-16 22:51:54 uses XML for storing maps 2014-10-16 22:52:00 ugh 2014-10-16 22:52:07 and XSLT for generating output from them 2014-10-16 22:55:04 thanks for the warning; I will never use OSM 2014-10-16 22:55:37 :-)) 2014-10-16 22:55:48 You can use it, without XML 2014-10-16 22:55:58 There are several tools and APIs out there 2014-10-16 22:56:12 that can help using that 2014-10-16 22:56:27 When you work in IT environment 2014-10-16 22:57:08 there's not good or bad tools! There are tools that you HAVE TO use and you can't get rid of them. And there are tools that you can avoid. 2014-10-16 22:58:10 some tools ARE bad, and some tools ARE good 2014-10-16 22:58:25 whether or not I'm forced to use them does not affect whether they're good or bad 2014-10-16 22:58:41 and if I were forced to work with XML at my job, I would quit and get another one 2014-10-16 22:59:03 using someone's api once, sure, whatever... I will not develop an application with it 2014-10-16 23:02:01 Heh... maybe in the future I can afford to be picky :P 2014-10-16 23:02:25 Right now it's VB.NET and PHP for me, yayyyyy 2014-10-16 23:07:50 :'( 2014-10-16 23:11:00 Eh, whtever... I can moonlight with Haskell, Python and C 2014-10-17 00:58:33 anyone running ZNC on Alpine? 2014-10-17 00:58:51 I am going through ZNCs instructions on chrooting the server 2014-10-17 00:59:52 but I can’t find libnss_dns.so.2 or libresolv.so.2 even after installing bind-libs, dhcpcd, and dhcp-dev 2014-10-17 01:00:07 I figured I would need to install something like that because of http://git.alpinelinux.org/cgit/aports/tree/extra/dhcp/dhcpd.initd?h=2.1-stable&id=aad5f530b5a45159e9420d654173345006bde998 2014-10-17 01:34:39 systmkor: libnss_dns and libresolv are both glibc libraries, they're builtin to the musl libc. 2014-10-17 01:34:52 It looks like ZNC depends on libssl.so.1.0.0 libcrypto.so.1.0.0 libz.so.1 libstdc++.so.6 libgcc_s.so.1 libc.musl-x86_64.so.1 2014-10-17 01:34:57 jlyo: thanks 2014-10-17 01:35:04 yah 2014-10-17 01:35:21 their guide made it look like you had to use ldd to find dependencies 2014-10-17 01:35:27 and then regardless of that 2014-10-17 01:35:35 you had to have libnss_dns and libresolv 2014-10-17 01:37:23 You could probably install it with apk --root=$chroot add znc 2014-10-17 01:45:19 jlyo: that seems a lot cleaner 2014-10-17 01:45:44 and doesn’t rely on znc devs getting a chroot correct 2014-10-17 08:34:08 systmkor: we have similiar issues for building CEPH for alpine. libresolv is not present because of musl. not all internals are implemented. for instance see http://bugs.alpinelinux.org/issues/3241 2014-10-17 12:32:15 libresolv is implemented, at least paritially 2014-10-17 12:32:38 the problem is that applications depend on the internals of the glibc libresolv implementation 2014-10-17 12:32:46 which is wrong 2014-10-17 12:33:07 apps should not depend on implementation internals 2014-10-17 15:25:30 programmers are mostly bad 2014-10-17 15:32:20 <^7heo_> yeah, but what did you answer to? 2014-10-17 15:33:48 <@ncopa> apps should not depend on implementation internals 2014-10-17 15:33:57 re: glibc 2014-10-17 15:34:14 <^7heo_> ah yeah 2014-10-17 15:34:25 <^7heo_> I agree with what's being discussed. 2014-10-17 15:34:28 <^7heo_> All along the line. 2014-10-17 17:13:30 glibc sort of made these legacy _res internals a public interface... :( 2014-10-17 17:13:48 but that still doesn't make it a good idea to use them 2014-10-17 17:14:29 and if you want to use your own nameservers rather than the default ones, you can just do res_mkquery and sendto() yourself rather than res_send with a modified resolver state 2014-10-18 03:00:04 i've done some research on how to uninstall mysql, but i just get a list of options 2014-10-18 03:00:20 i'm using apk "remove mysql" 2014-10-18 04:47:05 ben12: it's: "apk del" 2014-10-18 04:47:32 Goodbox: thank you 2014-10-18 12:15:44 <^7heo> damn 2014-10-18 12:15:50 <^7heo> sudo docker -d 2014-10-18 12:16:12 <^7heo> Listening for HTTP on unix (/var/run/docker.sock) \n no such file or directory 2014-10-18 12:16:20 <^7heo> whyyyy? 2014-10-18 12:16:38 <^7heo> shouldn't docker bring up the socket itself? 2014-10-18 13:27:06 <[xming]> anyone here with an alpine desktop care to test mintest (client) for me, it's in testing 2014-10-18 16:50:36 <^7heo> [xming]: why should we test it for you? 2014-10-18 17:53:36 <[xming]> ^7heo: because I don't have an alpine desktop, I have only tested the server part, not the client 2014-10-18 19:10:16 has there been any more investigation into malfunctions in firefox on alpine? 2014-10-19 00:53:21 <^7heo> dalias: no, but firefox's shit, and alpine's more secure than the average. So at least you know why it's crashing. 2014-10-19 00:53:59 ^7heo, does this mean you've packaged chromium ? :) 2014-10-19 00:54:34 i know some of the problems are grsec/pax kernel but i'm using vanilla 2014-10-19 00:56:51 <^7heo> well, you're still using musl 2014-10-19 00:57:09 <^7heo> and no, I'm not packaging chromium. 2014-10-19 00:57:36 <^7heo> even if I should, it's google, I can't, it's against my nature. 2014-10-19 00:57:41 <^7heo> and it's 3 AM 2014-10-19 00:57:43 <^7heo> I shall sleep 2014-10-19 00:57:45 <^7heo> o/ 2014-10-19 00:59:21 ^7heo, i know, i wrote it after all 2014-10-19 01:00:08 thats why i'm interested in finding out if there's been any more progress figuring out what about firefox isn't 100% working, which probably relates to musl (but is likely firefox's fault :) 2014-10-19 01:00:43 checking out and paying for a shipping label on usps.com failed for me last week :( 2014-10-19 01:03:30 ACTION is stuck with chromium 2014-10-19 01:03:53 as for grsec, quick hack is to disable sandboxing, sure can be configured properly but meh 2014-10-19 01:05:00 the grsec kernel has a lot more usability problems for me 2014-10-19 01:05:26 like not being able to inspect process memory usage in /proc/$pid/smaps 2014-10-19 01:05:48 i like being able to actually observe how much less memory software is using without glibc :) 2014-10-19 01:06:20 it sure could use more fine-grained controls in kernel.grsecurity.* 2014-10-19 01:06:39 spender's concerns are touching, but it indeed makes system almost unusable for any kind of troubleshooting :( 2014-10-19 01:06:45 most of the X stuff i'm running is 100-150 MB vsz but only 1-3 MB in vszrw 2014-10-19 01:07:02 which is quite nice 2014-10-19 01:07:22 i've never seen numbers that low for vszrw on debian 2014-10-19 01:07:35 unfair comparison :) 2014-10-19 01:07:54 Psh, who needs to troubleshoot when you have SECURITY 2014-10-19 01:08:00 are you joking, or do you have a good reason it's unfair? 2014-10-19 01:08:01 indigo: :) 2014-10-19 01:08:15 dalias: in debian, firefox works ^_^ 2014-10-19 01:08:18 haha 2014-10-19 01:08:23 i wasnt comparing firefox 2014-10-19 01:08:37 it's 833 and 551 MB right now 2014-10-19 01:08:40 nah, the point is, with musl, ton of stuff still breaks 2014-10-19 01:08:42 which is not bad 2014-10-19 01:08:54 not the programs being compared here 2014-10-19 01:09:21 Thunar, xfce4-panel, ... 2014-10-19 01:09:37 i haven't experienced problems with any of them 2014-10-19 01:10:04 yeah, i run 1gb vm for unsafe browsing and its definitely way better than previous debian vm 2014-10-19 01:10:15 closest to it was only maybe netbsd 2014-10-19 01:41:41 Hmmm... it would be cool to package opensmtpd for alpine 2014-10-19 02:20:37 what are your thoughts on linux containers becoming a more standard means of deploying applications/services? 2014-10-19 02:21:49 who are you asking? just everyone in general? 2014-10-19 02:22:14 yah 2014-10-19 02:22:23 I thought it would be an intersting discussion 2014-10-19 02:29:04 dalias: nothing? 2014-10-19 02:39:32 i think it's a lot better than stupid 'shared hosting' setups that are usually full of problems where one customer's crap can make security (or just reliability/resource-usage) issues for others 2014-10-19 02:43:14 can you be a bit more specific on shared hosting 2014-10-19 09:50:14 <^7heo> dalias: what did you write? 2014-10-19 10:41:44 <^7heo> Jean-Scotch: Is the correct English translation of your name: John Duct Tape? 2014-10-19 11:06:32 ^7heo: lol 2014-10-19 11:07:22 it was my nickname when I was student ;) 2014-10-19 11:07:56 <^7heo> Why? Something to do with highlands liquor? 2014-10-19 11:08:30 something, yes. I was a collector of sort back then 2014-10-19 11:08:52 I had mor than 100 differents sorts of in my dorm 2014-10-19 11:09:32 not really sort of, but bottle of every kind qnd sizes 2014-10-19 11:10:01 That was in the late eigties... 2014-10-19 11:10:37 *eighties 2014-10-19 11:13:09 <^7heo> Jean-Scotch: since you were a student, I can expect those to have been very low end ones... 2014-10-19 11:13:12 <^7heo> right? 2014-10-19 11:13:55 no, no. some where of cause. but most of them where rare or unusual. 2014-10-19 11:14:12 <^7heo> how did you pay for that? Oo 2014-10-19 11:14:24 *were (sorry but just wake up and first coffe is on his way) 2014-10-19 11:14:40 <^7heo> I'm kind of an amateur of whiskey, and here, the prices aren't THAT high compared to where you live... And it's 30E minimum for an okay bottle. 2014-10-19 11:15:01 those were gifts most of the time. 2014-10-19 11:15:06 <^7heo> "of course" also I guess? 2014-10-19 11:15:20 <^7heo> ah okay ;) 2014-10-19 11:15:31 <^7heo> so, crowdfunded scotch collection? :P 2014-10-19 11:15:49 lol. yes 2014-10-19 11:16:06 <^7heo> Sharing is caring right? :P 2014-10-19 11:16:17 <^7heo> you were sorta the scotch keeper I guess. 2014-10-19 11:16:27 my friends usually enjoyed my place ;) 2014-10-19 11:16:57 (they still do) 2014-10-19 11:23:02 <^7heo> I can easily imagine why 2014-10-19 11:24:13 change your nick to Jean-beer and ill happily join the fun :) 2014-10-19 11:24:20 <^7heo> okay, I still have to figure out why docker -d crashes 2014-10-19 11:24:30 <^7heo> clandmeter: where are you situated? 2014-10-19 11:24:39 nl 2014-10-19 11:24:45 <^7heo> oh 2014-10-19 11:24:51 <^7heo> Amsterdam? 2014-10-19 11:24:52 <^7heo> or? 2014-10-19 11:24:55 close to the belgium border 2014-10-19 11:25:00 <^7heo> oh so no 2014-10-19 11:25:06 <^7heo> I've not been there. 2014-10-19 11:25:18 <^7heo> I've been only in Amsterdam and Alkmaar 2014-10-19 11:25:24 eindhoven is not the place to be. 2014-10-19 11:25:27 <^7heo> both of which are lovely places. 2014-10-19 11:25:35 <^7heo> why not 2014-10-19 11:25:55 industrial modern area 2014-10-19 11:26:00 <^7heo> oh 2014-10-19 11:26:08 <^7heo> yeah, seems like depressing 2014-10-19 11:26:16 well, if you like IT stuff, eindhoven is the place to be. :) 2014-10-19 11:26:19 <^7heo> old industrial can be lots of fun, but modern... 2014-10-19 11:26:25 clandmeter: :p I come from a beer country but moved to a wine one ;) 2014-10-19 11:26:30 <^7heo> IT is my job, my passion, but not my way of life. 2014-10-19 11:26:42 ^7heo: correct 2014-10-19 11:26:46 <^7heo> Jean-Scotch: bist du Deutsch? 2014-10-19 11:27:00 nein (or nee) 2014-10-19 11:27:02 Jean-Scotch: wine is also fine :) 2014-10-19 11:27:26 <^7heo> wine is for pussies and old people 2014-10-19 11:27:53 im not a fan of whisky 2014-10-19 11:28:03 ^7heo: I guess I'm becoming old... 2014-10-19 11:28:05 <^7heo> 'cause you didn't try the right ones :) 2014-10-19 11:28:21 <^7heo> Jean-Scotch: if you were in uni during the 80', I can imagine so. 2014-10-19 11:28:54 ^7heo: is that an educated guess? 2014-10-19 11:30:06 <^7heo> about the right one? 2014-10-19 11:30:13 :) 2014-10-19 11:30:35 <^7heo> it's just that the richness of different whiskies and bourbons isn't that obvious at first. 2014-10-19 11:30:57 <^7heo> mainly because most people only know a few names 2014-10-19 11:31:07 <^7heo> mac allan, laphroig, etc. 2014-10-19 11:31:10 I totally agree with that statement 2014-10-19 11:31:13 <^7heo> glenfiddish ofc. 2014-10-19 11:31:30 <^7heo> so you try those 2014-10-19 11:31:38 <^7heo> as a "standard" 2014-10-19 11:31:43 <^7heo> and you declare "I don't like it" 2014-10-19 11:31:54 <^7heo> but you can find so, SO many more different types... 2014-10-19 11:32:30 <^7heo> for example, the first time I tried the Oban, I was completely shocked. Not that I fell in love with it, but that was completely beyond charted areas for whiskey for me :P 2014-10-19 11:33:12 <^7heo> then you learn about blends, single malts, the different types of casks and woods, etc etc. 2014-10-19 11:33:23 <^7heo> knowledge is power, and that also applies to whiskey :D 2014-10-19 11:34:22 <^7heo> but ofc if you go for a supermarket or hypermarket, and get the "standard" looking bottle, you'll have the standard taste, and it's even possible that it's made in india or anything. 2014-10-19 11:34:35 <^7heo> so it has chances to be a low quality ripoff 2014-10-19 11:35:36 <^7heo> rule of thumb, if you go for a supermarket, go for something else than whiskey; even baileys is better from a supermarket, because you know what to expect, and get it 100% of the time. 2014-10-19 11:37:55 <^7heo> anyway 2014-10-19 11:38:00 <^7heo> I've straced docker 2014-10-19 11:38:08 <^7heo> mount("/var/lib/docker", "/var/lib/docker", 0xc2100d7810, MS_BIND, NULL) = -1 ENOENT (No such file or directory) 2014-10-19 11:38:20 <^7heo> that's what's causing docker to not work. 2014-10-19 11:38:25 <^7heo> or to crash, even. 2014-10-19 11:39:51 <^7heo> the docker package was broken, I fixed it by "sudo mkdir /var/lib/docker" 2014-10-19 11:40:01 <^7heo> I guess it's maybe a good thing to fix it? 2014-10-19 11:40:07 <^7heo> I shall submit a bug report. 2014-10-19 11:41:33 or a patch and ill commit it 2014-10-19 11:42:06 <^7heo> yeah okay. 2014-10-19 11:42:10 ;-) 2014-10-19 11:42:11 <^7heo> gimme a moment to replicate it. 2014-10-19 11:42:26 <^7heo> btw, can I safely dd the alpine iso to a usb drive? 2014-10-19 11:42:29 <^7heo> 'cause it's not booting... 2014-10-19 11:42:34 sure you can 2014-10-19 11:42:38 but it wont boot 2014-10-19 11:42:55 we dont ship hybrid iso's 2014-10-19 11:43:47 we have a script that will do it for you 2014-10-19 11:45:59 <^7heo> okay I did expect it to be an hybrid iso 2014-10-19 11:46:01 <^7heo> thanks for the info 2014-10-19 11:46:25 <^7heo> what's the script, 'cause I don't find "hybridiso"? 2014-10-19 11:47:07 setup-bootable --help 2014-10-19 11:47:23 setup-bootable -h even :) 2014-10-19 11:47:26 <^7heo> danke. 2014-10-19 11:47:27 i didnt write it 2014-10-19 11:47:42 <^7heo> so 2014-10-19 11:47:53 <^7heo> setup-bootable alpine.iso /dev/sdb ? 2014-10-19 11:48:10 i think so, i never use it 2014-10-19 11:48:24 i just copy it myself to usb and install bootloader 2014-10-19 11:48:47 im pre setup-bootable :) 2014-10-19 11:48:50 <^7heo> yeah 2014-10-19 11:49:09 <^7heo> I just want a no-brainer for now, as I don't have the resources to learn it atm; but still need to boot it. 2014-10-19 11:50:11 I always have a copy of syslinux.exe on my windows box. then its just a matter of format the stick fat32 and copy the contents to / 2014-10-19 11:50:39 after just syslinux with correct opts 2014-10-19 11:50:51 <^7heo> clandmeter: does the disk need a filesystem prior to setting up bootable? 2014-10-19 11:51:02 i dont think so 2014-10-19 11:51:04 <^7heo> my guess is yes since you don't use the iso fs, right? 2014-10-19 11:51:06 <^7heo> ah? 2014-10-19 11:51:07 <^7heo> okay 2014-10-19 11:51:18 i guess the script takes care of it 2014-10-19 11:51:35 <^7heo> failed to remount sdb 2014-10-19 11:51:37 <^7heo> so yeah 2014-10-19 11:51:42 <^7heo> at least it needs a partition table. 2014-10-19 11:52:00 <^7heo> well, you need to point out the correct partition at least. 2014-10-19 11:52:03 <^7heo> so... 2014-10-19 11:52:32 could be, i never used it myself. 2014-10-19 11:52:48 <^7heo> mount: wrong fs type, bad option, bad superblock on /dev/sdb1, 2014-10-19 11:52:56 <^7heo> so yeah, partition required. 2014-10-19 11:53:42 correct 2014-10-19 11:53:47 by the looks of the src 2014-10-19 11:54:39 <^7heo> makes sense because it lets you chose what fs type you want. 2014-10-19 11:55:00 <^7heo> however 2014-10-19 11:55:05 <^7heo> the docker shit is not portable. 2014-10-19 11:55:09 <^7heo> \o/ 2014-10-19 11:55:15 <^7heo> try "docker pull ubuntu" 2014-10-19 11:55:20 <^7heo> and witness the crash(es) 2014-10-19 11:55:41 <^7heo> I expect something like the wrong shebang or something. 2014-10-19 11:55:56 <^7heo> I get a LOT of errors 2014-10-19 11:56:31 i use lxc and im happy. didnt find a reason (yet) to use docker. 2014-10-19 11:57:10 <^7heo> many more things are automated. 2014-10-19 12:01:34 <^7heo> damn 2014-10-19 12:01:43 <^7heo> I'll have to do it by hand I guess. 2014-10-19 12:01:47 <^7heo> "cannot find operating system" 2014-10-19 12:01:51 <^7heo> upon boot... 2014-10-19 12:02:26 <^7heo> wait no 2014-10-19 12:02:29 <^7heo> "Missing operating system" 2014-10-19 12:02:39 <^7heo> same meaning, but different operating system search 2014-10-19 12:02:53 <^7heo> s/operating system/search engine/ 2014-10-19 12:03:03 <^7heo> have to do less things at the same time. 2014-10-19 12:05:04 <^7heo> damnit 2014-10-19 12:05:17 <^7heo> the setup-bootable doesn't enforce the bootable flag on the partition 2014-10-19 12:05:20 <^7heo> suck 2014-10-19 12:06:50 <^7heo> now it boots, but it's stuck at "loading hardware drivers" 2014-10-19 12:09:22 <^7heo> okay, I bbl 2014-10-19 17:33:49 ^7heo, musl 2014-10-19 22:22:30 <^7heo> someone shoud package docker in a current version... 2014-10-19 22:22:36 <^7heo> Client version: 0.11.1 2014-10-19 22:22:42 <^7heo> Server version: 0.11.1 2014-10-19 22:22:49 <^7heo> Last stable version: 1.3.0, please update docker 2014-10-19 22:22:59 <^7heo> 0.11 / 1.3 ... 2014-10-19 22:23:16 <^7heo> I know docker evolves fast but seriously? 2014-10-19 22:24:17 version in edge is 1.2 2014-10-19 22:24:29 1.3 is from 2 days ago 2014-10-19 22:24:31 <^7heo> okay, I'll try it 2014-10-19 22:24:44 if you can do it better, go for it ;-) 2014-10-19 22:25:43 <^7heo> (1/1) Upgrading docker@edge (0.11.1-r1 -> 1.2.0-r0) 2014-10-19 22:25:51 <^7heo> clandmeter: for now I have to solve my own problems. 2014-10-19 22:25:55 <^7heo> but as soon as I got time 2014-10-19 22:26:01 <^7heo> I gotta contribute to alpine. 2014-10-19 22:26:11 <^7heo> because it's the best software I found out there 2014-10-19 22:26:22 <^7heo> and I'd even like to have Alpine UNIX rather than Alpine Linux. 2014-10-19 22:26:29 <^7heo> because it can work with other kernels. 2014-10-19 22:26:36 <^7heo> would be awesome. 2014-10-19 22:27:34 <^7heo> 1.2 works FAR better than 0.11 2014-10-19 22:27:35 <^7heo> :D 2014-10-19 22:30:01 <^7heo> but my connection is far to slow for that... 2014-10-19 22:30:20 <^7heo> anyway now itworks 2014-10-19 22:30:29 <^7heo> have a nice evening everyone 2014-10-19 22:30:34 itworks™ 2014-10-19 22:30:35 <^7heo> o/ 2014-10-19 22:30:48 godd night 2014-10-19 22:30:51 <^7heo> danke 2014-10-20 02:56:50 i need to learn how to setup a glibc-based system inside a container on alpine so i have a way to compare size/performance and test bugs on glibc without pulling out my old debian laptop 2014-10-20 03:40:50 there's a wiki article about arch in a chroot 2014-10-20 03:41:10 dalias: http://wiki.alpinelinux.org/wiki/Running_glibc_programs 2014-10-20 03:43:09 *nod* 2014-10-20 04:34:16 btw has anyone built chromium on alpine? 2014-10-20 04:35:27 stitch, that article looks really outdated, and using chroot rather than a container 2014-10-20 04:35:53 and parts don't make sense 2014-10-20 04:36:09 i don't see why you'd want/need a glibc-linked busybox if doing it that way 2014-10-20 04:36:24 using a static-linked busybox from the host libc would work just as well 2014-10-20 04:36:37 anyway, zzzz for now 2014-10-20 06:39:57 Hi 2014-10-20 06:40:15 Does anybody knows how to force to apk to use packages from 'testing' repo? 2014-10-20 06:40:35 Example: firefox is located in both repositories... 31.1.1 in 'main' and 33.0 in 'testing 2014-10-20 06:40:57 I want use 33.0 version but apk install 31.1.1 from 'main' each time 2014-10-20 06:43:35 (I have activated both repos in '/etc/apk/repositories', of course...) 2014-10-20 06:59:09 pinning 2014-10-20 09:33:28 good day, god dag, bonjour to all 2014-10-20 09:34:14 <^7heo_> moin 2014-10-20 13:23:42 quiet on here today! 2014-10-20 14:37:26 ouch - if alpine already has an ip during startup and /etc/init.d/networking wants to set one, this fails and therefore sshd won't be started m( 2014-10-20 14:57:24 How do you set your IP? /etc/network/interfaces? 2014-10-20 14:58:53 because starting in some VM or container? in this cas just deactivate the startup script of the network. 2014-10-20 16:06:01 yeah... to clarify: the ip is defined in /etc/network/interfaces with "iface eth0 inet static\n..." which will be read by /etc/init.d/networking which fails if the interface is already pre-configured outside the container 2014-10-20 16:07:04 just setting "iface eth0 inet manual" helps and i can still configure the other interfaces :-) 2014-10-20 16:07:53 is that so that sshd would start if the networking script was removed from startup? 2014-10-20 16:08:44 Jean-Scotch: ^ 2014-10-20 16:10:56 jomat: that's it ^^ you are better teacher that I'm ;) 2014-10-20 16:11:29 if the IP is set, ssh should start 2014-10-20 16:11:46 that's how it works on some container I manage 2014-10-20 16:12:34 and it's like that when I run alpine in a chroot on my desktop 2014-10-20 16:12:36 ssh refuses to start as networking can't start 2014-10-20 16:13:35 btw: do you have an idea how to start a lxc container in a pre-defined netns? 2014-10-20 16:24:09 jomat: http://wiki.alpinelinux.org/wiki/LXC 2014-10-20 16:24:52 No bridges 2014-10-20 16:25:58 with dummy module is it what you need ? 2014-10-20 16:27:00 No :-/ 2014-10-20 16:29:18 :/ 2014-10-20 16:29:23 I want to have network connectivity between containers by veth-pairs... my idea for that is: 2014-10-20 16:29:37 - create net namespaces, i. e. netns0 and netns1 2014-10-20 16:29:56 - create a pair of veth-interfaces 2014-10-20 16:30:29 - move these interfaces to netns0 and netns1 2014-10-20 16:30:39 no problems until here 2014-10-20 16:30:52 last step: start vm0 in netns0 and vm1 in netns1 2014-10-20 16:31:54 that should ensure that i can reboot vm0 and vm1 without loosing connectivity 2014-10-20 16:33:53 for example if i create vm0 with a normal veth device and if I specify this veth device as phys device in vm1 I'll also have a member of the veth pair in each vm 2014-10-20 16:34:53 but if vm1 is restarted I can't find the veth member of vm0 anywhere anymore 2014-10-20 16:34:56 and so on... 2014-10-20 16:38:48 <^7heo_> redshift is REALLY needed in alpine. 2014-10-20 16:38:52 <^7heo_> I'll package it. 2014-10-20 16:43:19 That's the spirit 2014-10-20 16:43:27 <^7heo_> yeah 2014-10-20 16:43:30 <^7heo_> I want to contribute 2014-10-20 16:43:31 <^7heo_> but 2014-10-20 16:44:32 <^7heo_> I have like, a flat to repaint, a flat to clean, a 2.5x2.5x2 meters structure to design, buy and assemble, a car to bring to the garage, many appointments to do during the day while I should be working, etc. 2014-10-20 16:44:40 <^7heo_> so, atm, I'm a bit short, time-wise. 2014-10-20 16:45:08 <^7heo_> and, that moment continuing for 2 months now. 2014-10-20 16:45:58 Heh 2014-10-20 16:46:22 I have a vacation in two weeks. First one in two years 2014-10-20 16:46:31 <^7heo_> yeah, moving, splitting with my ex, and getting a new job at the same time. 2014-10-20 16:46:32 Highly recommend vacations btw. 2014-10-20 16:46:46 <^7heo_> new job, I'm not rich, vacation-wise. 2014-10-20 16:47:06 <^7heo_> but once this is done 2014-10-20 16:47:13 Ha! Chill out in my flat writing scripts — that's my vacations 2014-10-20 16:47:14 <^7heo_> it's going to be champagen. 2014-10-20 16:47:38 <^7heo_> (reference to futurama, understands who can) 2014-10-21 07:23:24 Greetings all! 2014-10-21 07:33:40 morning ginjachris 2014-10-21 07:40:56 <^7heo_> moin 2014-10-21 07:51:28 hi 2014-10-21 08:29:49 <^7heo_> ncopa: 3.0.5 has no problem with the keyboards anymore. 2014-10-21 08:30:00 bug number? 2014-10-21 08:30:16 good to hear 2014-10-21 08:30:52 <^7heo_> moment 2014-10-21 08:31:22 <^7heo_> 3374 2014-10-21 08:31:47 <^7heo_> I've used 3.0.4 in kvm 2014-10-21 08:31:54 <^7heo_> was hanging 2014-10-21 08:31:58 <^7heo_> then 3.0.5 2014-10-21 08:31:59 <^7heo_> wasn't. 2014-10-21 08:58:23 #3374 2014-10-21 08:58:32 algitbot: hi 2014-10-21 08:58:37 algitbot: hi 2014-10-21 08:58:40 #3374 2014-10-21 08:59:20 ^7heo: ok. thanks for letting me know 2014-10-21 08:59:33 <^7heo_> no problem :) 2014-10-21 12:08:20 gotta go get food...in the meantime if anyone has any ideas about "rc-service knot start *rc-service: Permission denied" please dump them in here and I will look later :) 2014-10-21 12:11:39 Hello 2014-10-21 12:12:20 since a while now i've been looking into an issue, i'm trying to have AL install on a VPS in a VM 2014-10-21 12:12:24 so vm-in-vm 2014-10-21 12:12:39 hmm, christ! 2014-10-21 12:12:40 it works 2014-10-21 12:13:04 boot option nomodeset text avoids a problem in the ssh session 2014-10-21 12:17:56 i have a couple issues -- should i file bugs for these? 2014-10-21 12:19:37 1. the "networking" service fails to 'start' due to failure to configure lo (ip: RTNETLINK answers: File exists) and with lo removed, it fails to start because there are no interfaces at all 2014-10-21 12:20:34 2. the "ntp" service wrongly depends on networking (there's no reason it can't be started while there's no network, e.g. on a machine with only intermittent wifi, and merely produce no results until a network link is up) 2014-10-21 13:12:53 hmm, not sure 1 is a bug....you should always have loopback, if nothing else 2014-10-21 13:24:43 Is there a reason I cannot login to acf as root as the wiki page notes? Do I have to enable anything else? 2014-10-21 13:27:33 iudicare: you see the login page or a page to create an ADMIN user? 2014-10-21 13:28:00 tdtrask: i saw just the login page 2014-10-21 13:28:21 here is the acf passwd "root:x:Admin account:ADMIN" 2014-10-21 13:28:24 can you try 'admin' 'admin' 2014-10-21 13:28:35 ACTION thinks ncopa might have changed the setup-acf script to reuse the system root password 2014-10-21 13:28:41 have you tried that yet? 2014-10-21 13:29:32 i think with setup-scf you can then login with the root account (at least last time i tried) 2014-10-21 13:29:33 I try that and read over the setup-acf 2014-10-21 13:29:50 might be different if you don't use setup-acf 2014-10-21 13:30:11 if that doesn't work, just delete /etc/acf/passwd and it should prompt you to create an ADMIN user 2014-10-21 13:30:14 i think acf2 (in testing repo) has admin/admin by default 2014-10-21 13:30:58 acf2 is completely different 2014-10-21 13:31:07 ACTION is referring to the released ACF 2014-10-21 13:31:22 admin admin fails 2014-10-21 13:31:35 yeah, thought it might, but you gotta try 2014-10-21 13:32:03 iudicare: might be a good idea to also delete any lock files from /tmp/ 2014-10-21 13:32:15 and you can't login with 'root' user and your configured root password? 2014-10-21 13:32:18 removing the passwd file gave me the user creation script thank!!! 2014-10-21 13:32:22 you 2014-10-21 13:32:26 once you fail 10x, you are locked out for 30 min 2014-10-21 13:32:38 iudicare: welcome 2014-10-21 13:34:49 clear 2014-10-21 13:34:55 exit 2014-10-21 13:35:04 I like my pf... three connections in three minutes, blacklisted for 12 hours. 2014-10-21 13:35:39 course that's on bsd 2014-10-21 13:46:22 NameBrand: the "no outgoing for non-root users" pf rule i have there is also great 2014-10-21 13:46:31 for dmz webservers etc 2014-10-21 13:46:45 right I've retested a couple times on different machines.......rc-service refuses to start 'knot' (new package from bug 3466) with "Permission Denied" 2014-10-21 13:46:54 not ever seen this with rc-service before 2014-10-21 13:47:24 my "amazing (ahem) troubleshooting skills tell me this may be a permissions issue :) 2014-10-21 13:51:29 alright think I mostly got it 2014-10-21 13:52:10 file /etc/init.d/knot needs to be made executable 2014-10-21 13:52:15 darkfader: *nice* I've got per user restrictions and logging myself. 2014-10-21 13:52:52 then /etc/knot/knot.conf needs to be created (not a big deal, can move/copy the example config) 2014-10-21 13:53:06 then, tries to run as knot:knot 2014-10-21 13:53:12 these don't exist 2014-10-21 13:53:40 I'll update the feature request 2014-10-21 14:02:38 http://bugs.alpinelinux.org/issues/3466 updated :) 2014-10-21 14:04:09 hmm, not sure 1 is a bug....you should always have loopback, if nothing else 2014-10-21 14:04:33 loopback fails to configure via "networking" because it's already configured (probably by default by the kernel) 2014-10-21 14:04:38 that's the bug 2014-10-21 14:38:17 apologies, I misunderstood 2014-10-21 14:41:35 katlogic: hi 2014-10-21 14:41:54 and if you omit lo from interfaces file (seems like the obvious workaround) 2014-10-21 14:42:06 then 'networking' fails to start because there are no interfaces 2014-10-21 14:42:27 so it's either fail-because-lo-is-already-configured or fail-because-there-are-no-interfaces 2014-10-21 14:42:42 no way to make it succeed without stupid things like a dummy interface 2014-10-21 14:46:06 Hi, am i mistaking or does alpine not provide an upgrade all updated packages using apk ? 2014-10-21 14:46:18 Jarth: apk upgrade -a -U 2014-10-21 14:47:31 why -a and -U ? 2014-10-21 14:49:04 it seems -a is a mistake of mine, but strange at that 2014-10-21 14:49:11 -U fetches the updated index 2014-10-21 14:49:15 -a seemed odd to me 2014-10-21 14:49:19 apk -U upgrade 2014-10-21 14:49:23 most likely to do what he wants 2014-10-21 14:49:24 -U just does "apk update" along with the upgrade? 2014-10-21 14:49:28 yes 2014-10-21 14:49:34 nice 2014-10-21 14:49:56 i thought there's -a flag, was it removed by chance? 2014-10-21 14:50:25 -a flag exists but its behavior sounds weird 2014-10-21 14:51:24 darkfader: thanks 2014-10-21 14:52:01 tried before, it but just reports 200MB updates, nothing done 2014-10-21 14:52:54 if it looks like this: 2014-10-21 14:52:54 OK: 424 MiB in 166 packages 2014-10-21 14:53:00 then there's no update 2014-10-21 14:53:12 and like this is if it found one 2014-10-21 14:53:13 xen03:~# apk upgrade -a U 2014-10-21 14:53:13 (1/4) Upgrading libcrypto1.0 (1.0.1i-r2 -> 1.0.1j-r0) 2014-10-21 14:53:19 that was a typo 2014-10-21 14:53:29 but it worked on like it doesn't care :) 2014-10-21 14:54:38 ahwr! 2014-10-21 15:02:10 dalias: the -a flag is for 'use whatever package is available in repo' 2014-10-21 15:02:18 you can build a package locally for example 2014-10-21 15:02:33 so it has the exact same version number than the one existing in repository 2014-10-21 15:02:51 but apk will notice that checksum differs 2014-10-21 15:03:43 so when you upgrade from v2.7 (uclibc) to musl, even if package has same version number, it was built with different libc 2014-10-21 15:03:54 and you surely want the version with the correct libc 2014-10-21 15:03:56 thus -a 2014-10-21 15:05:54 ah 2014-10-21 15:12:42 hi 2014-10-21 15:13:44 <^7heo_> yo 2014-10-21 15:14:02 yesterday I asked a question, but I do not get an answer... 2014-10-21 15:14:11 ncopa, re: the issues i mentioned, should i file bugs? 2014-10-21 15:14:21 ...so I try again: 2014-10-21 15:14:58 Does anybody knows how to force to apk to use packages from 'testing' repo? 2014-10-21 15:15:16 Example: firefox is located in both repositories... 31.1.1 in 'main' and 33.0 in 'testing' 2014-10-21 15:15:41 add testing repo above main 2014-10-21 15:17:10 <^7heo_> AmatCoder: in /etc/apk/repositories, tag your repositories with @tag 2014-10-21 15:17:17 Hi again, when i'm using LXC with the Debian template, this would mean i'd be running an Alpine Linux container ( using the grsec alpine kernel ) with a Debian system on top ? 2014-10-21 15:17:23 <^7heo_> and then 2014-10-21 15:17:30 <^7heo_> apk add package@tag 2014-10-21 15:20:13 xen_roger: That did not work... 2014-10-21 15:20:50 <^7heo_>: how can I tag my repositories? 2014-10-21 15:20:53 check the way ^7heo_ mentions 2014-10-21 15:21:04 <^7heo_> @tag http://address/... 2014-10-21 15:21:27 <^7heo_> in /etc/apk/repositories 2014-10-21 15:22:01 something like: @something http://nl.alpinelinux.org/alpine/edge/testing 2014-10-21 15:22:10 I will try right now... 2014-10-21 15:25:19 hmm, it did not work neither... 2014-10-21 15:27:27 <^7heo_> AmatCoder: what do you do, and what do you expect? 2014-10-21 15:28:00 <^7heo_>: Firefox is located in both repositories... 31.1.1 in 'main' and 33.0 in 'testing' 2014-10-21 15:28:09 <^7heo_> yeah and? 2014-10-21 15:28:18 <^7heo_>: I want use 33.0 2014-10-21 15:28:19 <^7heo_> apk add firefox@testing 2014-10-21 15:28:32 <^7heo_> and don't tell me that it doesn't work, that's what I use. 2014-10-21 15:28:58 <^7heo_> (firefox doesn't work, alright, but you can install the version 33 nevertheless) 2014-10-21 15:29:14 doesn't work? 2014-10-21 15:29:17 how? 2014-10-21 15:29:23 a few things seem broken but overall it mostly works 2014-10-21 15:29:37 <^7heo_> dalias: I got a "symbol not found" error 2014-10-21 15:29:42 <^7heo_> dalias: but the system's at home. 2014-10-21 15:29:47 <^7heo_> so I can't say here and now. 2014-10-21 15:30:02 that shouldn't happen 2014-10-21 15:30:11 <^7heo_> well, it does. 2014-10-21 15:30:18 maybe you've got plugins it's trying to load and the error is in a plugin? 2014-10-21 15:30:23 <^7heo_> I can give you more info later 2014-10-21 15:30:25 <^7heo_> but now I'm @work 2014-10-21 15:30:36 ok 2014-10-21 15:30:42 <^7heo_> nah, no plugin afaik 2014-10-21 15:30:44 please do 2014-10-21 15:31:48 <^7heo_> yeah I have nothing better to do than to contribute, during the week. 2014-10-21 15:31:51 <^7heo_> Well, to sleep, too. 2014-10-21 15:31:54 ^7heo_: can you force to install the version from 'main'? 2014-10-21 15:32:39 <^7heo_> apk del firefox && apk add firefox 2014-10-21 15:34:14 Have you firefox into '/etc/apk/world' ? 2014-10-21 15:34:26 bye all, have a good evening! 2014-10-21 15:35:03 because here it's not working... :( 2014-10-21 15:41:47 Oh, I got it... 2014-10-21 15:42:21 I needed 'apk del firefox-pdfjs' 2014-10-21 15:42:42 (1/5) Purging firefox-pdfjs (31.1.1-r0) 2014-10-21 15:42:49 (2/5) Installing icu-libs (53.1-r0) 2014-10-21 15:42:57 (3/5) Upgrading firefox@testing (31.1.1-r0 -> 33.0-r1) 2014-10-21 15:43:04 (4/5) Purging libbz2 (1.0.6-r3) 2014-10-21 15:43:06 ^7heo_: Error relocating /usr/lib/firefox-33.0/libmozalloc.so: malloc_usable_size: symbol not found 2014-10-21 15:43:09 This one? 2014-10-21 15:43:11 (5/5) Purging xulrunner (31.1.1-r0) 2014-10-21 15:44:22 <^7heo_> jomat: yes 2014-10-21 15:44:28 <^7heo_> this very one 2014-10-21 15:46:04 Here it is working fine (one minute test) 2014-10-21 15:46:17 Start Page works at least! 2014-10-21 15:48:04 ^7heo_: http://git.musl-libc.org/cgit/musl/tree/WHATSNEW 2014-10-21 15:48:13 ^7heo_: 2014-10-21 15:48:28 1.1.5 release notes: new features: … malloc_usable_size function (nonstandard) 2014-10-21 15:48:29 <^7heo_> I suspected a change in musl yes. 2014-10-21 15:48:39 <^7heo_> but I didn't check more than that 2014-10-21 15:48:59 <^7heo_> so what you mean is 2014-10-21 15:49:16 <^7heo_> if I want to install the newest firefox, I also gotta use a newer musl? 2014-10-21 15:53:01 i think so... but i don't know how to install it... 2014-10-21 15:53:09 % sudo apk add musl@edge 2014-10-21 15:53:09 OK: 1236 MiB in 521 packages 2014-10-21 15:53:17 % apk info -L musl 2014-10-21 15:53:17 musl-1.1.4-r3 contains: 2014-10-21 15:53:18 ... 2014-10-21 15:55:11 <^7heo_> I 2014-10-21 15:55:14 ah 2014-10-21 15:55:16 <^7heo_> will try' 2014-10-21 15:56:16 if edge repositories are activated into '/etc/apk/repositories'... 2014-10-21 15:56:24 ...just do 'apk update && apk upgrade' 2014-10-21 15:58:42 AmatCoder: @edge http://dl-3.alpinelinux.org/alpine/edge/main 2014-10-21 15:58:59 i have it tagged in my apk/repositories 2014-10-21 16:00:10 AmatCoder: http://dl-3.alpinelinux.org/alpine/edge/main/x86_64/APKINDEX.tar.gz also has musl-1.1.5 2014-10-21 16:00:35 but apk version musl only knows about 1.1.4-r3 2014-10-21 16:01:05 hmm, strange... 2014-10-21 16:03:36 apk policy musl: http://sprunge.us/jTKd 2014-10-21 16:05:34 Has you tried '--update-cache' option ? 2014-10-21 16:09:57 Yeah 2014-10-21 16:10:32 apk add musl=1.1.5 tells me about a lot of unsatisfiable constraints 2014-10-21 16:11:12 ironically incl. firefox-33.0-r1[so:libc.musl-x86_64.so.1] 2014-10-21 16:11:33 I guess that you can not upgrade only musl... You must upgrade all packages to edge 2014-10-21 16:12:06 But I dont know what "musl=1.1.5" actually means 2014-10-21 16:12:21 apk add musl=1.1.4 does the same :-D 2014-10-21 16:13:07 apk does not understand "musl=xxxx" as far as I know 2014-10-21 16:14:06 At least not as we understand it 2014-10-21 16:15:29 sudo apk -X http://dl-3.alpinelinux.org/alpine/edge/main add musl <- that did something with 7 packets... but not with musl 2014-10-21 16:17:20 none package depends on those 7 packages...that's why 2014-10-21 16:18:13 (1/7) Updating pinning libxscrnsaver (1.2.2-r0) 2014-10-21 16:18:15 ... 2014-10-21 16:19:47 downloading the apk and running 'sudo apk add musl-1.1.5-r0.apk musl-dev@edge' did it 2014-10-21 16:20:16 <^7heo_> ah yeah you need to update apk, I guess. 2014-10-21 16:20:34 I tried 'apk add musl@edge musl-utils@edge musl-dbg@edge', that did nothing 2014-10-21 16:21:19 hmm, I hope that it does not break anything 2014-10-21 16:21:42 and it brought back my firefox 29 >:( 2014-10-21 16:23:05 Not really... firefox starts, the about dialog says 29.0.1 2014-10-21 16:23:22 and apk version firefox: firefox-33.0-r1 2014-10-21 16:23:53 /usr/lib/firefox-33.0/firefox is the binary... wtf 2014-10-21 16:24:33 ah... there was another firefox window still open 2014-10-21 16:24:49 and now it tells me my sqlite is too old 2014-10-21 16:26:01 Sure, you are mixing packages from two repositories... 2014-10-21 16:26:07 you need: sqlite-libs@edge 2014-10-21 16:26:39 <^7heo_> damn, redshift becomes intense at night 2014-10-21 16:28:08 and the Pentadactyl Nightly 2014-10-11 also works with it, yeah :-D 2014-10-21 16:31:38 <^7heo_> ncopa: I have an apology to make. I did give you some wrong information. The keyboard isn't working better with 3.0.5; at least not with a real machine. 2014-10-21 16:35:06 <^7heo_> sorry I have tried the 3.0.5 before trying your fix... 2014-10-21 16:35:07 <^7heo_> but actually 2014-10-21 16:35:18 <^7heo_> modprobe usbkbd was what made the keyboard work. 2014-10-21 16:35:38 <^7heo_> usbhid just prevented usbkbd from being inserted in the first place. 2014-10-21 16:35:50 <^7heo_> =/ 2014-10-21 16:36:01 <^7heo_> so now it's broken, and not fixable; while before, it was fixable. 2014-10-21 16:36:12 <^7heo_> https://bugs.alpinelinux.org/issues/3374 2014-10-21 16:36:36 <^7heo_> (I have a working keyboard, which is not a cherry, and I have tried modprobe usbkbd, it doesn't find it) 2014-10-21 16:36:46 <^7heo_> (rmmod usbhid breaks the working keyboard on the other hand) 2014-10-21 16:37:54 <^7heo_> I have to flash the key with 3.0.4 again. 2014-10-21 16:37:59 <^7heo_> and then I can make it work... 2014-10-21 16:38:01 <^7heo_> but that's crap. 2014-10-21 16:39:10 <^7heo_> how can I re-open the bug? 2014-10-22 01:03:13 I'd like to use openvpn with polarssl instead of openssl... other than compiling from source, what are my options? 2014-10-22 06:43:22 hay! any quick tutorial on how to set up Alpine as a client?.. like Thunderbird ?? thanks! 2014-10-22 07:21:20 Good Morning all 2014-10-22 07:25:35 morning 2014-10-22 07:26:38 shodan45: maybe edit openvpn APKBUILD to support a USE_POLARSSL conditional or so? 2014-10-22 07:39:17 <^7heo> moin ncopa 2014-10-22 07:39:24 <^7heo> did you see my messages from yesterday? 2014-10-22 07:40:06 oh 2014-10-22 07:40:14 so we introduced a new problem? 2014-10-22 07:40:49 <^7heo> yessir 2014-10-22 07:41:14 <^7heo> i.e. before, I had to plug another keyboard, run a command as root, and the first keyboard was working 2014-10-22 07:41:35 <^7heo> now, I can't do anything to solve it with the non-working keyboard: usbkbd isn't found at all (ofc) 2014-10-22 07:44:15 hm 2014-10-22 07:44:36 is hid-generic loaded? 2014-10-22 07:44:44 <^7heo> I can check. 2014-10-22 07:44:49 <^7heo> moment, I have a ticket to solve first. 2014-10-22 07:45:05 fabled: you had an issue with usbkbd? 2014-10-22 07:45:16 which was the proper module that should be used instead of usbkbd? 2014-10-22 07:45:35 bah 2014-10-22 07:45:38 you know what 2014-10-22 07:45:43 # CONFIG_HID_CHERRY is not set 2014-10-22 07:45:53 i think usbkbd is wrong module 2014-10-22 07:46:59 ^7heo: do you think you could file a new issue? 2014-10-22 07:47:08 i think removing usbkbd was the correct thing to do 2014-10-22 07:47:16 but we need enable various CONFIG_HID_* modules 2014-10-22 07:47:27 including CONFIG_HID_CHERRY 2014-10-22 07:47:28 <^7heo> ncopa: I know, I made a patch for that one 2014-10-22 07:47:31 <^7heo> didn't send it tho 2014-10-22 07:47:44 <^7heo> but that HID variables would be for the media keys anyway, or? 2014-10-22 07:47:56 <^7heo> I might have the patch with me, even 2014-10-22 07:48:05 <^7heo> since I exceptionally have my personal laptop @work 2014-10-22 07:48:10 <^7heo> (needed to configure access to my vps) 2014-10-22 07:48:14 Help text 2014-10-22 07:48:14 Support for Cherry Cymotion keyboard. 2014-10-22 07:53:20 ncopa, my issue was with usbmouse, if it's loaded it took over the hid driver and made extended keys not work 2014-10-22 07:57:56 <^7heo> ncopa: so 2014-10-22 07:58:14 <^7heo> I need to clone the repo containing grsec 2014-10-22 07:58:24 ^7heo: i think the proper thing to do is file a bug for cherry keyb support 2014-10-22 07:58:25 <^7heo> then modify the file, commit 2014-10-22 07:58:29 <^7heo> but then, where do I send the patch? 2014-10-22 07:58:31 <^7heo> irc/query? 2014-10-22 07:58:34 <^7heo> or ml? 2014-10-22 07:58:40 i think i can just do it 2014-10-22 07:58:46 <^7heo> okay 2014-10-22 07:58:51 <^7heo> but why file a bug? 2014-10-22 07:58:57 for history 2014-10-22 07:59:15 so when i look in future, why did we enable this? 2014-10-22 07:59:32 or if someone else gets a keyb problem 2014-10-22 07:59:39 they can search the bugs database and find it 2014-10-22 08:01:50 ^7heo: also, i might forget the whole thing when go looking for coffe... :) 2014-10-22 08:01:58 which i'll do now 2014-10-22 08:05:13 it could be problem is that hid-generic is not loaded 2014-10-22 08:05:46 <^7heo> yeah 2014-10-22 08:05:49 <^7heo> have to test that. 2014-10-22 08:06:30 <^7heo> okay, now is the right time. 2014-10-22 08:11:19 <^7heo> okay tested. 2014-10-22 08:11:20 <^7heo> so 2014-10-22 08:11:34 <^7heo> hid-generic appaears in lsmod (in the right part) 2014-10-22 08:11:56 <^7heo> and it's not working, even after doing "rmmod hid-generic; modprobe hid-generic" 2014-10-22 08:12:03 <^7heo> (as root ofc) 2014-10-22 08:14:58 super, thanks 2014-10-22 08:15:07 then i think we only need enable the proper hid module 2014-10-22 08:17:16 yes 2014-10-22 08:17:37 cypress has malformed HID descriptor - the HID/cypress driver fixes it up and allows the keyboard to work 2014-10-22 08:19:11 seems that we need add it to initramfs too 2014-10-22 08:25:07 hum 2014-10-22 08:25:10 and new grsecurity patch... 2014-10-22 08:28:08 no, no new grsec patch. i was looking at wrong versin 2014-10-22 08:33:31 ^7heo: coudl you please file an issue sayin that while fixing #3374, the cypress keyb support was removed 2014-10-22 08:34:48 <^7heo> ncopa: I can. 2014-10-22 08:34:54 thanks 2014-10-22 08:35:00 i want this fixed for v3.0.6 2014-10-22 08:35:14 <^7heo> my problem right now is that if I use alpine @work, if I do ANY apk upgrade 2014-10-22 08:35:19 <^7heo> I can't use the keyboard again. 2014-10-22 08:35:20 <^7heo> =/ 2014-10-22 08:35:29 <^7heo> and I would GLADLY work with alpine. 2014-10-22 08:35:43 thats why i want this fixed asap 2014-10-22 08:35:53 <^7heo> yeah 2014-10-22 08:35:56 <^7heo> that'd be awesome. 2014-10-22 08:36:14 <^7heo> even if I need a kernel from edge or testing, at least it would work. 2014-10-22 08:36:23 <^7heo> and I can pin the kernel from edge and use the rest from stable. 2014-10-22 08:36:25 <^7heo> which is perfect. 2014-10-22 08:36:34 i just want an issue number that i can refer to in the git commit 2014-10-22 08:37:04 <^7heo> no problem. 2014-10-22 08:42:39 <^7heo> crap I forgot my password. 2014-10-22 08:43:09 <^7heo> good thing I have my laptop @work today 2014-10-22 08:50:59 <^7heo> ncopa: 3469 2014-10-22 08:53:21 thanks1 2014-10-22 08:53:45 <^7heo> anytime, that's normal to help. 2014-10-22 08:53:47 #3469 2014-10-22 08:53:51 #3469 2014-10-22 08:53:54 algitbot: hi 2014-10-22 08:53:57 algitbot: hi 2014-10-22 08:53:58 <^7heo> archlinux rocks so much :P 2014-10-22 08:54:01 #3469 2014-10-22 08:54:11 <^7heo> alpinelinux I mean 2014-10-22 08:54:12 <^7heo> damn 2014-10-22 08:54:18 <^7heo> why this name seriously 2014-10-22 08:54:24 <^7heo> r and l are pronounced almost the same 2014-10-22 08:54:29 <^7heo> so al-ar same sound 2014-10-22 08:54:40 <^7heo> p and ch aren't so far away neither 2014-10-22 08:54:48 <^7heo> so alp and arch are still close enough 2014-10-22 08:54:58 <^7heo> and that's enough for me to confuse them ALL the time. 2014-10-22 08:55:13 <^7heo> I wish I could destroy archlinux with mind bullets 2014-10-22 08:55:20 <^7heo> and that nobody would remember it ever. 2014-10-22 08:55:32 <^7heo> so when I say archlinux, people would answer: you mean alpinelinux right? 2014-10-22 08:56:31 you mead adpidelidux, dight? 2014-10-22 08:56:46 <^7heo> haha :D 2014-10-22 08:58:18 <^7heo> on the other hand, archlinux rock too, differently: when you install archlinux on a computer, you get a rock. 2014-10-22 08:59:01 lol 2014-10-22 08:59:02 a brick 2014-10-22 08:59:24 suggestion for new distro name: brickedlinux 2014-10-22 09:00:18 <^7heo> I did think about "stablinux" 2014-10-22 09:00:46 <^7heo> STatic Basic Linux 2014-10-22 09:00:51 <^7heo> sorry 2014-10-22 09:00:53 <^7heo> STAtic Basic Linux 2014-10-22 09:01:12 <^7heo> the goal of that Linux would be to automatize the package management to the extreme... 2014-10-22 09:01:22 <^7heo> to allow people to publish buildfiles 2014-10-22 09:01:34 <^7heo> which would be feteched by distributed build systems 2014-10-22 09:01:38 <^7heo> to build static binaries 2014-10-22 09:01:55 <^7heo> so you can drag/drop any binary anywhere provided that you can execute that binary format and have the correct arch. 2014-10-22 09:02:05 <^7heo> but I'd better contribute to alpine first. 2014-10-22 09:23:09 drats 2014-10-22 09:23:14 i enabled wrong thing 2014-10-22 09:55:47 <^7heo> ncopa: in #3469, by "cypress" you mean "cherry", right? 2014-10-22 11:05:03 hello again all 2014-10-22 11:05:17 <^7heo> moin moin 2014-10-22 11:05:25 morning 2014-10-22 12:05:44 Hi, community :) 2014-10-22 12:07:16 Got a quick question regarding user-built packages 2014-10-22 12:07:36 How can I know which packages to upgrade from git ports ? 2014-10-22 12:08:29 I mean, let's say I build "minidlna" version 1.1.4 2014-10-22 12:08:53 and the ports gets upgraded to 1.1.7 2014-10-22 12:09:31 after a "git pull", how can I check if minidlna was upgraded ? 2014-10-22 12:10:07 (A general command would be nice, as I build a few packages by myself) 2014-10-22 12:14:10 z3bra: not sure i catch your point 2014-10-22 12:14:34 you want to know if there is a new version upstream? 2014-10-22 12:14:54 yeah 2014-10-22 12:15:10 if so, ncopa write a script that will compare our aports tree with others (like archlinux) 2014-10-22 12:15:42 Something that will tell me "Hey, you use package X in version Y. But the version upstream is Z. Just sayin'.." 2014-10-22 12:15:52 there is no *real* solution to this problem 2014-10-22 12:16:00 Yeah I'm aware of that 2014-10-22 12:16:12 I just wanted to know how other users do ? 2014-10-22 12:16:19 there was a/is a website that keeps track of pkg versions 2014-10-22 12:16:43 would be nice :) 2014-10-22 12:22:50 Can I get all the packages I built myself ? 2014-10-22 12:24:11 ? 2014-10-22 12:24:48 the packages in my /usr/abuild/{main,testing,...} 2014-10-22 12:25:22 can apk give me a list of packages that I did not installed from remote repo ? 2014-10-22 12:25:47 you want apk list the repo the pkg was installed from? 2014-10-22 12:26:11 i think so 2014-10-22 12:26:21 no i dont think its possible 2014-10-22 12:26:30 it was something i wanted to ask fabled 2014-10-22 12:26:47 fabled: is it possible to list which repo the pkg came from? 2014-10-22 12:27:32 i think apk threats all repo's as one, except pinned repo's 2014-10-22 12:27:49 pinned ? 2014-10-22 12:28:07 lets say you build them local 2014-10-22 12:28:27 yep 2014-10-22 12:28:28 they would be in ~/packages/main/arch/whatever.apk 2014-10-22 12:28:32 clandmeter, "apk policy " 2014-10-22 12:28:52 ^ 2014-10-22 12:29:03 ah ok 2014-10-22 12:29:08 thx fabled 2014-10-22 12:29:31 fabled: i guess there is no way to have that listed in apk info? 2014-10-22 12:29:43 damn, i scarred him 2014-10-22 12:30:38 z3bra: i use this for upstream monitoring: http://git.alpinelinux.org/cgit/ncopa/upstream-monitor/tree/ 2014-10-22 12:30:43 i have some local changes i think 2014-10-22 12:30:46 for PKG in $(apk info); do REPO="$(apk policy $PKG | tail -n1)"; echo $PKG - $REPO; done 2014-10-22 12:30:48 that checks github 2014-10-22 12:32:16 http://i.imgur.com/bOMnIF6.png 2014-10-22 12:32:26 clandmeter, is that what you want ? 2014-10-22 12:32:59 kind of 2014-10-22 12:36:06 z3bra: repo pinning, you can prefix a repo with for instance @test 2014-10-22 12:36:16 then you can add it like myapk@test 2014-10-22 12:37:49 this way you can mix edge with stable. (still need to be carefull with abi changes though) 2014-10-22 12:37:52 ncopa, I just want to check alpine repos 2014-10-22 12:38:18 SOme packages need to be built by hand (eg. minidlna, from main) 2014-10-22 12:38:34 what is wrong with minidlna? 2014-10-22 12:38:40 And I just want to know wether my version is up to date with upstream 2014-10-22 12:38:47 nothing is wrong 2014-10-22 12:38:55 it's an example like any other 2014-10-22 12:39:20 sure, but i guess its a real example 2014-10-22 12:39:24 thats why i ask. 2014-10-22 12:39:50 It could be polarss, umurmur, protobuf, ... 2014-10-22 12:39:56 there are many of ther 2014-10-22 12:39:59 *them 2014-10-22 12:40:21 sure, but im interested in minidnla, not the rest. 2014-10-22 12:40:33 what do you want to know ? 2014-10-22 12:40:43 why you build it local 2014-10-22 12:40:52 apk search minidlna 2014-10-22 12:40:57 -> nothing 2014-10-22 12:41:32 ah, now it's avail 2014-10-22 12:42:20 :) 2014-10-22 12:42:44 has been since few weeks 2014-10-22 12:43:04 I think I installed it right after it was pushed to main 2014-10-22 12:43:22 and thus, was not already built on remote repos 2014-10-22 12:43:40 I then assumed I'd have to build it by hand forever.. 2014-10-22 12:43:43 synced probably 2014-10-22 12:43:59 use nl.a.o if you want ti quick 2014-10-22 12:44:03 Now I realise.. That's was dumb to think so xD 2014-10-22 12:44:10 I use it 2014-10-22 12:44:13 :p 2014-10-22 12:44:22 then when its pushed its there after build 2014-10-22 12:44:44 Mmmh... 2014-10-22 12:44:46 and with minidlna, that takes about 2 minutes :) 2014-10-22 12:44:54 if not faster. 2014-10-22 12:45:12 so basically, I should NEVER build package from main by hand 2014-10-22 12:45:20 (only if I want to modify them) 2014-10-22 12:45:25 correct 2014-10-22 12:45:32 and just wait for them to appear in repo 2014-10-22 12:46:02 Mmmh 2014-10-22 12:46:03 if you find them in aports, they should be on our mirror 2014-10-22 12:46:04 okay 2014-10-22 12:46:12 thanks for the enlightenment 2014-10-22 12:46:17 or you are so fast, its not ready with building them. 2014-10-22 12:46:49 ^^ 2014-10-22 12:47:03 and stuff you miss, we can probably add it. 2014-10-22 12:47:22 of course we prefer spoon fed with patches 2014-10-22 12:47:24 Yeah I need to submit pull request... 2014-10-22 12:47:43 or tarballs or whatever 2014-10-22 12:47:49 are you on github? 2014-10-22 12:47:56 I am now 2014-10-22 12:48:09 but I don't really us it as I have my own repo online 2014-10-22 12:48:23 it doesnt matter for us what you use. 2014-10-22 12:48:30 aslong we can read it. 2014-10-22 12:48:35 and git of course 2014-10-22 12:48:39 ;) 2014-10-22 12:48:51 btw, what do you prefer ? 2014-10-22 12:49:17 if you send us one patch a week, then sprunge them is easiest 2014-10-22 12:49:24 z3bra 2014-10-22 12:49:24 `- testing 2014-10-22 12:49:24 `- pkg1 2014-10-22 12:49:24 `- pkg2 2014-10-22 12:49:33 or directly under "z3bra" ? 2014-10-22 12:49:51 no 2014-10-22 12:49:59 you need to clone our aports 2014-10-22 12:50:11 add stuff to aports/testing 2014-10-22 12:50:29 Ah, I see.. 2014-10-22 12:50:53 in github its easy to fork 2014-10-22 12:51:00 For me to 2014-10-22 12:51:04 git clone 2014-10-22 12:51:05 :P 2014-10-22 12:51:10 else just git clone git.a.o 2014-10-22 12:51:24 I'll just not be able to push 2014-10-22 12:51:42 correct. 2014-10-22 12:52:06 http://git.z3bra.org/cgit.cgi/alpine/tree/ 2014-10-22 12:52:11 ^ here it is actually 2014-10-22 12:52:25 ones we get quality patches regular, we can give ppl push access. 2014-10-22 12:52:31 It just bother me that on MY repo, will be a full mirror of git.a.o 2014-10-22 12:52:49 Oh, okay 2014-10-22 12:53:17 then use github 2014-10-22 12:53:51 the fun thing about github is you can create PR's 2014-10-22 12:54:04 then we can discus about the actual commits. 2014-10-22 12:54:21 similar like if you want send a patch to our ml 2014-10-22 12:54:31 Ah yeah 2014-10-22 12:54:38 I need to subscribe to it also 2014-10-22 12:55:44 we didnt receive a PR yet :). we still need to test it. 2014-10-22 12:55:55 ^^ 2014-10-22 12:56:09 I'll fix my packages first 2014-10-22 12:56:26 then set up everything correctly and sync it to github 2014-10-22 12:56:38 And we'll try eventually 2014-10-22 12:56:59 i would suggest to create a branch, and create the pr from that branch. 2014-10-22 12:57:18 but i think thats discribed in githubs docs. 2014-10-22 12:57:53 I'm not used to github yet 2014-10-22 13:06:51 fatal: No names found, cannot describe anything. 2014-10-22 13:06:57 what does that mean ? 2014-10-22 13:07:50 which cmd did you run? 2014-10-22 13:08:24 abuild -r 2014-10-22 13:08:58 that sounds more like a git msg 2014-10-22 13:09:17 ah, yeah 2014-10-22 13:14:59 I patched mutt to include the "sidebar" patch 2014-10-22 13:15:19 should I create a mutt-sidebar package, or update the main package ? 2014-10-22 13:18:52 can it be turned off/on? 2014-10-22 13:19:13 yeah 2014-10-22 13:19:24 throw in the notmuch patches while you're at it :P 2014-10-22 13:19:38 if you have verified its working, then its fine. 2014-10-22 13:19:38 what's this ? 2014-10-22 13:19:43 it is ! 2014-10-22 13:19:54 (I've been using it for 3 years now) 2014-10-22 13:22:58 what's the status of alpine/arm? 2014-10-22 13:23:18 we have an rpi image which works 2014-10-22 13:23:44 i'm getting some arm hardware for testing/development of musl and was thinking of putting alpine on it 2014-10-22 13:23:56 it should also work on wandboard, which fabled uses as builder. 2014-10-22 13:31:38 <^7heo> ncopa: any way I can get the usbkbd module back? 2014-10-22 13:31:44 <^7heo> ncopa: should I build it? 2014-10-22 13:34:34 z3bra: I was half-joking there https://github.com/karelzak/mutt-kz has a bunch of other patches merged in. perhaps it'd make nice separate package. 2014-10-22 13:35:29 using standalone notmuch with a bunch of shell scripts personally, so no vested interest in mutt anymore 2014-10-22 13:37:20 I find mutt pretty convenient to sort my mails by hand 2014-10-22 13:47:36 cool, keepassx compiles fine... 2014-10-22 13:47:53 I really need this :) 2014-10-22 13:51:18 ^7heo: do you think you could test the http://nl.alpinelinux.org/alpine/edge/releases/x86_64/alpine-edge-141022-x86_64.iso 2014-10-22 13:51:25 with your keyboard 2014-10-22 13:51:59 just plug the keyboard into anything that boots that is 2014-10-22 13:52:07 <^7heo> ncopa: yep 2014-10-22 13:52:20 <^7heo> ncopa: but please allow me some time to do so, many tickets to solve. 2014-10-22 13:54:08 <^7heo> I made a postit for it. 2014-10-22 13:54:15 <^7heo> so I'll do that after the other postits. :P 2014-10-22 13:54:43 <^7heo> (yes I use post-its as a kanban board) 2014-10-22 13:54:54 real postits, or the digital kinds? 2014-10-22 13:55:02 <^7heo> real ones. 2014-10-22 13:55:06 <^7heo> on the wall on my right 2014-10-22 13:55:12 damn, you are hardcore 2014-10-22 13:55:18 <^7heo> when i do a task, I place the post-it on my desk in front of me. 2014-10-22 13:55:22 i used a whiteboard before 2014-10-22 13:55:27 <^7heo> so if I'm ever interrupted, I know what to resume. 2014-10-22 13:55:30 in my previous office 2014-10-22 13:55:35 <^7heo> yeah that works too 2014-10-22 13:55:38 <^7heo> but I like post its 2014-10-22 13:55:40 <^7heo> it works super well 2014-10-22 13:56:28 i actually have a few post-its on my desk as well 2014-10-22 13:56:39 and i can "close" one of them. thanks for the reminder :) 2014-10-22 13:56:54 actually, i can close both 2014-10-22 13:57:03 <^7heo> anytime :D 2014-10-22 14:01:14 <^7heo> and 2014-10-22 14:01:18 <^7heo> the AWESOME part of it 2014-10-22 14:01:27 <^7heo> it's that blockers can mask the rest 2014-10-22 14:01:33 <^7heo> like, if a ticket is blocked by another one 2014-10-22 14:01:41 <^7heo> I place the blocker in front of the blocked. 2014-10-22 14:01:48 <^7heo> so I always see ONLY the tickets I can work on. 2014-10-22 14:01:59 <^7heo> but when they are resolved, I directly see the newly unblocked ones :P 2014-10-22 14:02:02 <^7heo> great, right? 2014-10-22 14:02:12 indeed 2014-10-22 14:03:42 <^7heo> the only problem I have is to have more than one blocker on a single ticket 2014-10-22 14:03:53 <^7heo> 'cause the order is relevant, and if you have no particular order, it's harder to place :P 2014-10-22 14:07:47 <^7heo> and the person able to do that postit wall in a terminal would be my god. 2014-10-22 14:08:03 <^7heo> best would be over git, with remote repo and web integration. 2014-10-22 14:08:14 <^7heo> then we would be able to replace jira with a much better product :P 2014-10-22 14:08:23 <^7heo> (w/out java !! :D) 2014-10-22 14:10:23 heard of taskwarrior? 2014-10-22 14:11:19 I've done some scripting around that can dump groff documents for schedules, postit wall shouldn't be significantly harder 2014-10-22 14:11:32 <^7heo> taskwarrior is shit. 2014-10-22 14:11:37 <^7heo> it's written in C++ for starters 2014-10-22 14:11:47 <^7heo> and then, I've used it for some time, but it simply sucks. 2014-10-22 14:12:19 your username is pretty annoying to type 2014-10-22 14:12:20 <^7heo> maybe they did improve it significantly since I tried it (2010 if my memory's right) 2014-10-22 14:12:32 <^7heo> z3bra: type 7 and use tab... 2014-10-22 14:12:38 well, then there's the whole orgmode crowd 2014-10-22 14:12:40 woah 2014-10-22 14:12:45 <^7heo> you're welcome. 2014-10-22 14:12:50 ^^ 2014-10-22 14:12:55 <^7heo> CcxCZ: that's even worse. 2014-10-22 14:13:10 yeah, couldn't get into that at all 2014-10-22 14:13:27 <^7heo> sorry, but, taskwarrior being the less bad CLI alternative, it's not enough. 2014-10-22 14:13:39 <^7heo> problem with my constant ranting at what sucks 2014-10-22 14:13:42 <^7heo> I don't implement nearly enough. 2014-10-22 14:13:54 <^7heo> so since I don't want to shut up, I have to implement. 2014-10-22 14:16:48 I love cli when it's good. that usually involves me writing bunch of convenience functions in zsh and perhaps zaw or some other nifty widget 2014-10-22 14:17:07 <^7heo> please define "good" 2014-10-22 14:17:23 <^7heo> for me, "good" is "stripped down to the bare minium" 2014-10-22 14:17:37 <^7heo> which taskwarrior isn't. 2014-10-22 14:20:36 good :- convenient(CommonCase), flexible, concise. 2014-10-22 14:20:50 <^7heo> that's very vague. 2014-10-22 14:21:05 <^7heo> anything can be described like that. Even systemd. 2014-10-22 14:21:52 should thrown in "one thing well" I guess 2014-10-22 14:22:23 still, tw looked better than any competing thing I've tried 2014-10-22 14:22:39 could be modularized better, I agree 2014-10-22 14:33:53 <^7heo> what do you mean by "modularized better"? 2014-10-22 14:36:45 make core that just emits parseable output for queries and have the UI(s) as a separate thing, like notmuch 2014-10-22 14:37:51 hmm basically it boils down to it being RDBMS :D 2014-10-22 14:53:28 <^7heo> CcxCZ: you mean that it should be done with different filters rather than modularized components? 2014-10-22 15:16:48 Ok guys, I forked aports repo, and added my packages in a new branch 2014-10-22 15:16:50 now what ? 2014-10-22 15:17:32 <^7heo> now git diff? 2014-10-22 15:17:47 git diff or pull request ? 2014-10-22 15:18:00 <^7heo> not enough information to tell 2014-10-22 15:18:26 It adds 11 packages to testing 2014-10-22 15:18:52 so git diff might be consequent :O 2014-10-22 15:18:59 <^7heo> $ git remote -v 2014-10-22 15:21:49 z3bra: for each package make a commit 2014-10-22 15:22:02 and use the same commit msg we normally use 2014-10-22 15:22:24 ^7heo, http://sprunge.us/MTgB 2014-10-22 15:22:53 <^7heo> z3bra: pull request. 2014-10-22 15:23:29 testing/pkgname: new aport 2014-10-22 15:23:55 okay 2014-10-22 15:24:08 are you going to use github? 2014-10-22 15:24:11 idk 2014-10-22 15:24:14 as you wnat 2014-10-22 15:24:16 *want 2014-10-22 15:24:19 what you want 2014-10-22 15:24:20 I don't really care 2014-10-22 15:24:24 ^^ 2014-10-22 15:24:31 me neither 2014-10-22 15:24:38 you can also just sprunge them 2014-10-22 15:25:03 do you have a clean aports local? 2014-10-22 15:25:52 yeah 2014-10-22 15:26:32 i would advise to create a sep branch 2014-10-22 15:26:39 for each package ? 2014-10-22 15:26:48 no 2014-10-22 15:27:04 just call it dev 2014-10-22 15:27:05 <^7heo> clandmeter: he uses github. Look at the paste. 2014-10-22 15:27:13 doesnt matter 2014-10-22 15:27:19 That's what I did then 2014-10-22 15:27:25 <^7heo> yes it does matter, he needs to do a PR then. 2014-10-22 15:27:31 And I added all my packages to that branch 2014-10-22 15:27:32 <^7heo> not to submit a patch or whatever. 2014-10-22 15:27:45 <^7heo> well, he can submit a patch, but it's easier to do a PR then. 2014-10-22 15:28:05 like i said, it depends how you want to do it. 2014-10-22 15:28:07 PR is fine 2014-10-22 15:28:37 if you want pr its almoist the same 2014-10-22 15:29:09 create a branch 2014-10-22 15:29:14 as I said, I don't mind. github act as a mirror for my personnal repo 2014-10-22 15:29:28 check it out and add each pkg in its own commit 2014-10-22 15:29:47 then push the new branch to github 2014-10-22 15:30:07 go to that branch and create a PR 2014-10-22 15:30:08 How do you delete a branch on github ? 2014-10-22 15:31:09 git push origin --delete 2014-10-22 15:31:29 thanks 2014-10-22 15:31:54 a branch is just so you can keep your master clean 2014-10-22 15:32:31 branches are cheap. I like cheap, im dutch. 2014-10-22 15:34:32 <^7heo> :D 2014-10-22 15:34:39 256mb for a VIA usb3 driver... 2014-10-22 15:34:45 got to love windows 2014-10-22 15:34:57 <^7heo> wtf?! 2014-10-22 15:35:07 <^7heo> 0xFF MB for a usb3 driver!? 2014-10-22 15:35:14 <^7heo> or wait no, worse 2014-10-22 15:35:17 <^7heo> 0x100 2014-10-22 15:36:51 lol 2014-10-22 15:36:54 and its not enough 2014-10-22 15:37:05 this board also has intel usb3 2014-10-22 15:37:21 and windows doesnt have it on its magical update server. 2014-10-22 15:38:18 ah intel is only 4.5mb 2014-10-22 15:38:45 <^7heo> intel knows how to do drivers. 2014-10-22 15:38:47 and thats of course including a ghaphical installer 2014-10-22 15:38:49 <^7heo> and when you come to think of it 2014-10-22 15:38:56 <^7heo> 5MB for a driver is already freaking HUGE. 2014-10-22 15:39:20 a simple inf + sys would be enough. 2014-10-22 15:39:46 <^7heo> if only =/ 2014-10-22 15:39:49 <^7heo> I'd be SO HAPPY. 2014-10-22 15:39:52 <^7heo> I could put that in git 2014-10-22 15:40:00 <^7heo> and organize ALL the drivers like I want. 2014-10-22 15:40:22 <^7heo> but no, manufacturers want to force you to install hundreds of bloated binaries each time you plug yet another device. 2014-10-22 15:41:00 except gigabyte, they are the best of course. 2014-10-22 15:46:30 Why is a HDD installed Alpine needed for building? 2014-10-22 15:47:19 ? 2014-10-22 15:47:28 From wiki... 2014-10-22 15:47:43 http://wiki.alpinelinux.org/wiki/Setting_up_the_build_environment_on_HDD 2014-10-22 15:47:51 well 2014-10-22 15:48:11 you have enough memory to build on tmpfs? 2014-10-22 15:48:26 I have 8 gb 2014-10-22 15:48:37 its enough for small stuff 2014-10-22 15:48:44 in other words, no 2014-10-22 15:48:52 hehe 2014-10-22 15:48:52 i have 16 and run in trouble with hd even 2014-10-22 15:49:06 try to build webkit 2014-10-22 15:49:08 or java 2014-10-22 15:49:26 it will make your memory cry for mama 2014-10-22 15:50:31 well, i am thinking about desktop apps... 2014-10-22 15:51:10 building or installing? 2014-10-22 15:51:20 building 2014-10-22 15:51:46 yeah... glwt 2014-10-22 15:52:47 you could run tmpfs, but you should put your aports tree on a hd with only 8gb. 2014-10-22 15:53:30 ccache could be on tmpfs, maybe that will be a bit faster. 2014-10-22 15:54:33 great, I will try 2014-10-22 15:55:07 or you could upgrade to 128gb memory 2014-10-22 15:55:14 that should be enough. 2014-10-22 15:55:30 oh, men... 2014-10-22 15:55:32 but memory prices are not that low latelty 2014-10-22 15:57:36 time go call it a day. nice evening. z3bra good luck with your commits! 2014-10-22 15:57:43 They're done ! 2014-10-22 15:57:56 what's the difference between unified/split pull requests ? 2014-10-22 15:58:17 ah, no, neverming 2014-10-22 15:58:25 its on github already? 2014-10-22 15:58:42 yeah 2014-10-22 15:58:54 https://github.com/z3bratabs/aports 2014-10-22 15:58:56 im looking at the wrong username? 2014-10-22 15:58:57 "dev" branch 2014-10-22 15:58:59 yeah 2014-10-22 15:59:06 this cocksucker... 2014-10-22 15:59:14 "z3bra" is idling there 2014-10-22 15:59:19 doing nothing 2014-10-22 15:59:35 and support team don't want to release the username atm 2014-10-22 15:59:44 https://help.github.com/articles/using-pull-requests 2014-10-22 16:00:08 Yeah I read that 2014-10-22 16:00:19 I'm commenting it to give explaination about packages 2014-10-22 16:00:32 ok then just create the PR and ill review it later tonight 2014-10-22 16:01:32 ok thanks ;) 2014-10-22 16:04:20 <^7heo> http://www.bbc.com/news/uk-england-hampshire-29568551 2014-10-22 16:04:23 <^7heo> rofl. 2014-10-22 16:06:58 <^7heo> okay, going to test the keyboard with edge. 2014-10-22 16:09:39 <^7heo> ncopa: works like a charm. 2014-10-22 16:09:43 <^7heo> awesome work. 2014-10-22 16:09:49 <^7heo> what did you do? 2014-10-22 16:10:08 woah, my first PR 2014-10-22 16:10:23 <^7heo> :P 2014-10-22 16:11:35 <^7heo> crap! 2014-10-22 16:11:45 <^7heo> ncopa: I could have fixed that one or two months ago... 2014-10-22 16:11:52 see ya ! 2014-10-22 16:11:57 <^7heo> it's exactly what I did in my patch... 2014-10-22 16:12:14 <^7heo> now I feel bad =/ 2014-10-22 16:12:25 <^7heo> I was so convinced that my patch wouldn't solve anything... 2014-10-22 16:14:09 <^7heo> (especially since the module usbkbd did solve the problem) 2014-10-22 17:08:42 <^7heo> ncopa: so basically, to fix that issue with the keyboard, I install alpine stable, and I apk add linux-grsec@edge and done?? 2014-10-22 17:45:01 ^7heo: ncopa is not here after 5 cet 2014-10-22 17:45:15 mostly... 2014-10-22 17:48:49 <^7heo> damnit 2014-10-22 17:48:56 <^7heo> thanks for the info clandmeter 2014-10-22 17:49:06 <^7heo> I'll try what I said, it should have fixed it. 2014-10-22 17:49:22 <^7heo> or, are the kernel modules in another package? 2014-10-22 17:50:01 no 2014-10-22 17:50:10 same packager 2014-10-22 17:50:13 -r 2014-10-22 17:50:33 but maybe he change something in initramfs? 2014-10-22 17:52:43 <^7heo> yeah 2014-10-22 17:52:56 <^7heo> but if you install a new linux-grsec, it runs mkinitramfs right? 2014-10-22 17:54:11 yes 2014-10-22 17:56:35 disk based only ofc 2014-10-22 18:12:52 is there any plans of moving from 3.14 to 3.16 kernel? 2014-10-22 18:30:53 Well, my first attempt to make an aport: 2014-10-22 18:31:01 http://sprunge.us/PWNO 2014-10-22 18:31:38 I am not sure if it is correct... 2014-10-22 18:33:28 ...probably not ;) 2014-10-22 18:34:14 AmatCoder: sweet 2014-10-22 18:34:22 i'll check it out in a few minutes 2014-10-22 18:34:39 systmkor: thanks! 2014-10-22 18:43:22 AmatCoder: Looks good, the only thing I can see after reading it is the source could use $pkgname and $pkgver to make version bumps easier. 2014-10-22 18:43:26 source="http://www.keepassx.org/releases/${pkgname}-${pkgver}.tar.gz 2014-10-22 18:43:57 oops, you are right 2014-10-22 18:48:45 Fixed: http://sprunge.us/ENTT 2014-10-22 18:53:07 Well, I have to go 2014-10-22 18:53:13 No hurry on this... I am just learning. 2014-10-22 18:53:33 thanks to jlyo and systmkor 2014-10-22 18:53:37 bye! 2014-10-22 19:59:10 z3bra: merged :) 2014-10-22 20:05:47 I noticed 2014-10-22 20:05:54 I saw your fiexxes too 2014-10-22 20:05:57 wow 2014-10-22 20:05:59 fixes 2014-10-22 20:06:16 sorry for not being consistent enough yet 2014-10-22 20:06:42 I'll review the APKBUILDs tomorrow to fix a few things 2014-10-22 20:07:07 Anyway thanks ;) 2014-10-22 20:11:57 im just nitpicking... 2014-10-22 20:41:27 <^7heo> okay 2014-10-22 20:41:30 <^7heo> back home 2014-10-22 20:43:23 nice, so you can finaly do some real work. 2014-10-22 20:45:07 <^7heo> like watching series? 2014-10-22 20:45:09 <^7heo> : 2014-10-22 20:45:11 <^7heo> :D* 2014-10-22 20:45:19 <^7heo> nah 'cause seriously 2014-10-22 20:45:21 <^7heo> I have 1h 2014-10-22 20:45:24 <^7heo> then I have to sleep 2014-10-22 20:45:28 <^7heo> and then I go back to work 2014-10-22 20:45:52 where are you from? 2014-10-22 20:46:01 <^7heo> I think you know :P 2014-10-22 20:46:07 <^7heo> if you ask, you know. 2014-10-22 20:46:23 i asked before? 2014-10-22 20:46:40 <^7heo> nah, but you're asking because I'm refusing to work :P 2014-10-22 20:47:13 i dont get it :) 2014-10-22 20:47:41 i was just wondering because of the timezone 2014-10-22 20:47:47 <^7heo> oh 2014-10-22 20:47:50 <^7heo> GMT+2 2014-10-22 20:47:56 <^7heo> I'm French. Unfortunately. And the French people aren't exactly hard workers if you get my meaning. 2014-10-22 20:48:01 <^7heo> Also 2014-10-22 20:48:03 <^7heo> I live in Germany 2014-10-22 20:48:05 ah haha 2014-10-22 20:48:17 <^7heo> Which is a better country than France, by far, on many aspects (not only beer) 2014-10-22 20:48:33 <^7heo> they also happen to think 2014-10-22 20:49:28 i like both countries 2014-10-22 20:49:34 <^7heo> Well, me too. 2014-10-22 20:49:46 <^7heo> Germany and Netherlands are both very nice countries. 2014-10-22 20:50:14 Im happy to be close to 2 borders 2014-10-22 20:50:27 <^7heo> meh 2014-10-22 20:50:34 <^7heo> The north of France is actually okay. 2014-10-22 20:50:37 im going to luxembourg next week 2014-10-22 20:50:48 so maybe ill hop into france. 2014-10-22 20:51:03 they seem to have public holidays at that day 2014-10-22 20:51:43 in luxembourg i mean. 2014-10-22 20:51:48 <^7heo> for fucksake 2014-10-22 20:51:55 <^7heo> redshift build tools are shit. 2014-10-22 20:52:05 <^7heo> $ ./bootstrap 2014-10-22 20:52:06 <^7heo> ./bootstrap: line 7: autopoint: not found 2014-10-22 20:52:09 <^7heo> v_v 2014-10-22 20:52:24 <^7heo> I just want to type "make" and be done with it, that's not too hard, is it? 2014-10-22 20:53:44 well, if you dont have to patch it together, youll be fine. 2014-10-22 20:54:00 <^7heo> whatcha mean? 2014-10-22 20:54:16 i mean in general. 2014-10-22 20:54:31 like im trying to package pulseaudio 2014-10-22 20:54:40 because its a hard dep... 2014-10-22 20:54:45 and it doesnt like musl 2014-10-22 20:55:03 or lets say, musl doesnt like pulseaudio 2014-10-22 20:55:17 and lennard doesnt like musl... 2014-10-22 20:55:38 <^7heo> wait 2014-10-22 20:55:42 <^7heo> pulseaudio? 2014-10-22 20:55:44 <^7heo> why?! 2014-10-22 20:55:49 <^7heo> nah seriously why?! 2014-10-22 20:55:53 i just told you 2014-10-22 20:55:55 <^7heo> alpine is such a nice distro... 2014-10-22 20:55:59 <^7heo> why ruin everything? 2014-10-22 20:56:19 <^7heo> Nah seriously 2014-10-22 20:56:23 <^7heo> hard dep or what 2014-10-22 20:56:26 i dont like it either 2014-10-22 20:56:26 <^7heo> I don't flying care. 2014-10-22 20:56:32 <^7heo> if people require it 2014-10-22 20:56:35 <^7heo> they shall burn in hell 2014-10-22 20:56:41 <^7heo> with their shitware 2014-10-22 20:56:56 <^7heo> now, if people want to use software that relies on it, they should either re-implement the software 2014-10-22 20:57:04 <^7heo> or build it themselves. 2014-10-22 20:57:05 most desktop env need parts of gnome 2014-10-22 20:57:12 except a few basic ones. 2014-10-22 20:57:12 <^7heo> not dwm 2014-10-22 20:57:16 <^7heo> and dwm rocks. 2014-10-22 20:57:24 <^7heo> nah but really 2014-10-22 20:57:32 <^7heo> gtk3 is already a big step backwards 2014-10-22 20:57:41 <^7heo> you're not going the package that shit, are you? 2014-10-22 20:57:48 gtk3? 2014-10-22 20:57:53 <^7heo> first, it's time that would be MUCH better invested elsewhere 2014-10-22 20:57:58 <^7heo> nah pulseaudio 2014-10-22 20:58:08 <^7heo> gtk3 is done already, you can't change what's done. 2014-10-22 20:58:08 if i get the chance i will 2014-10-22 20:58:16 <^7heo> but why?! 2014-10-22 20:58:21 <^7heo> you're so going to ruin everything! 2014-10-22 20:58:27 <^7heo> you know how I solve sound problems?! 2014-10-22 20:58:30 <^7heo> ALWAYS works? 2014-10-22 20:58:41 <^7heo> I FUCKING REMOVE FUCKING PULSEFUCKINGAUDIO SHIT! 2014-10-22 20:58:52 <^7heo> so nah seriously. 2014-10-22 20:58:58 <^7heo> Please mate. 2014-10-22 20:59:05 <^7heo> You have a brain, use it on things that matter. 2014-10-22 20:59:18 <^7heo> What's so important that pulse audio is required in that beautiful alpine? 2014-10-22 20:59:26 :) 2014-10-22 20:59:38 <^7heo> I mean, who uses Gnome anyway? 2014-10-22 20:59:46 <^7heo> it looks like unity, it's just worse. 2014-10-22 20:59:59 <^7heo> you're not going to package systemd are you? 2014-10-22 21:00:50 i just like a desktop env with a little more eyecandy. 2014-10-22 21:00:57 <^7heo> yeah, well 2014-10-22 21:01:09 <^7heo> eye candy has nothing to do with bloat. 2014-10-22 21:01:21 <^7heo> so please re-implement. 2014-10-22 21:01:24 <^7heo> don't package shit,. 2014-10-22 21:01:37 <^7heo> you know 2014-10-22 21:01:40 <^7heo> right now 2014-10-22 21:01:48 <^7heo> I'm trying to package that redshift stuff 2014-10-22 21:01:51 well, it can run without libpulse 2014-10-22 21:01:53 <^7heo> it's mostly shit, by design. 2014-10-22 21:01:59 <^7heo> but it's useful. 2014-10-22 21:02:00 <^7heo> so 2014-10-22 21:02:02 just not build without... 2014-10-22 21:02:05 <^7heo> if I can't package it 2014-10-22 21:02:09 <^7heo> I'm just going to re-implement it. 2014-10-22 21:02:18 <^7heo> I mean the ONLY hard stuff is to talk to X 2014-10-22 21:02:26 <^7heo> to change the temperature of the display. 2014-10-22 21:02:28 <^7heo> That's AKLL. 2014-10-22 21:02:29 <^7heo> ALL* 2014-10-22 21:02:33 <^7heo> the rest is uber easy. 2014-10-22 21:02:43 <^7heo> and so I'm just going to copy paste that. 2014-10-22 21:02:58 <^7heo> and who cares if it doesn't work on Mac|Windows? I don't. 2014-10-22 21:03:11 <^7heo> I just want it to run on BSD and Linux. 2014-10-22 21:03:26 <^7heo> and obviously, if it can run on BSD, it has chances to run on Mac anyway. 2014-10-22 21:03:50 <^7heo> but I am SO not going to build any gtk or Qt interface with it... 2014-10-22 21:03:56 <^7heo> just 'cause it sucks. 2014-10-22 21:04:05 <^7heo> I prefer to use libcaca 2014-10-22 21:04:48 <^7heo> wtf?! What do they smoke!!? 2014-10-22 21:05:15 <^7heo> https://github.com/jonls/redshift/blob/master/src/solar.c#L50-54 2014-10-22 21:05:19 <^7heo> nah but REALLY?! 2014-10-22 21:05:27 <^7heo> a one line macro expanded in a fucking function?! 2014-10-22 21:05:34 <^7heo> YEAH LET'S JUUMP! 2014-10-22 21:05:43 <^7heo> YEAAAAH FUCK YEAH! 2014-10-22 21:05:59 <^7heo> and they do it for EVERYFUCKINGTHING in the top of that =/ 2014-10-22 21:06:09 <^7heo> it's like the file is full of that, 2014-10-22 21:06:38 <^7heo> omg, I am so going to rewrite that. 2014-10-22 21:06:46 <^7heo> anyway, bed now. 2014-10-22 21:06:49 <^7heo> o/ 2014-10-22 21:06:52 <^7heo> have a nice one people. 2014-10-22 21:07:02 gnite, hope you can sleep. 2014-10-22 21:07:09 <^7heo> I have good beer. 2014-10-22 21:07:13 <^7heo> nothing can beat good beer. 2014-10-22 21:07:22 whisky? 2014-10-22 21:07:22 <^7heo> not even code THAT awful. 2014-10-22 21:07:28 <^7heo> finished it gestern 2014-10-22 21:07:34 ach so 2014-10-22 21:07:36 <^7heo> ja 2014-10-22 21:07:43 <^7heo> I have to get more. 2014-10-22 21:07:46 <^7heo> MEAR!!! 2014-10-22 21:08:02 <^7heo> anyway 2014-10-22 21:08:09 <^7heo> guten abend mate. 2014-10-22 21:08:11 <^7heo> o/ 2014-10-22 21:08:36 <^7heo> oder gute nacht even 2014-10-22 21:08:47 <^7heo> tchuss :) 2014-10-23 07:51:30 <^7heo> moin 2014-10-23 07:51:48 <^7heo> ncopa: just to get a confirmation; to make the keyboard work with stable: 2014-10-23 07:52:38 <^7heo> $ sed 's/^\(.*edge.*\)$/@edge \1/' /etc/apk/repositories > /etc/apk/repositories.new 2014-10-23 07:52:47 <^7heo> $ mv /etc/apk/repositories.new /etc/apk/repositories 2014-10-23 07:52:51 <^7heo> $ apk update 2014-10-23 07:52:58 <^7heo> $ apk add linux-grsec@edge 2014-10-23 07:53:00 <^7heo> right? 2014-10-23 07:53:28 <^7heo> oh wait, I made a mistake 2014-10-23 07:53:34 <^7heo> $ sed 's/^#\(.*edge.*\)$/@edge \1/' /etc/apk/repositories > /etc/apk/repositories.new 2014-10-23 07:53:37 <^7heo> that's the right line 2014-10-23 08:24:16 good morning everyone 2014-10-23 08:25:01 <^7heo> moin ginjachris 2014-10-23 12:26:17 <^7heo> ncopa: ping 2014-10-23 12:38:19 ^7heo: pong 2014-10-23 12:38:56 <^7heo> awesome 2014-10-23 12:39:08 <^7heo> do you have the log from 4:30 ago? 2014-10-23 12:40:20 i have the log, but 4:30 what timezone? 2014-10-23 12:40:23 utc? 2014-10-23 12:40:41 <^7heo> ago 2014-10-23 12:41:06 <^7heo> around 10 AM CEST 2014-10-23 12:41:10 <^7heo> (aka UTC+2) 2014-10-23 12:41:29 <^7heo> you should be able to read it anyway, normally. 2014-10-23 12:41:35 <^7heo> it should be just above 2014-10-23 12:42:00 about getting keyboard work with stable? 2014-10-23 12:42:03 <^7heo> yep 2014-10-23 12:42:12 i'd say wait for 3.0.6 2014-10-23 12:42:17 i am trying to get it out today 2014-10-23 12:42:24 <^7heo> point is 2014-10-23 12:42:28 <^7heo> I have some PCs to setup 2014-10-23 12:42:33 <^7heo> oh 2014-10-23 12:42:34 <^7heo> today 2014-10-23 12:42:42 <^7heo> can I help somehow? 2014-10-23 12:42:57 <^7heo> aside leaving you work in peace? 2014-10-23 12:43:21 http://bugs.alpinelinux.org/projects/alpine/issues?fixed_version_id=89&set_filter=1&status_id=o 2014-10-23 12:43:41 i think i have all important done 2014-10-23 12:44:10 ah 2014-10-23 12:44:14 there is a xen sec issue 2014-10-23 12:44:33 rnalrd: could you please bump linux-virt-grsec to 3.14.22? 2014-10-23 12:44:45 sure 2014-10-23 12:45:36 ^7heo: you can help by answering ppl in this channel so i can work in peace ;) 2014-10-23 12:45:46 been pretty quiet today though 2014-10-23 12:47:02 all postits gone? 2014-10-23 12:49:12 <^7heo> clandmeter: no. 2014-10-23 12:49:18 <^7heo> and I had to create new ones. 2014-10-23 12:49:33 in german or french? 2014-10-23 12:49:47 <^7heo> English. 2014-10-23 12:50:01 <^7heo> basically I try to write everything in English 2014-10-23 12:50:03 <^7heo> so everyone can read. 2014-10-23 12:50:11 except the germans? 2014-10-23 12:50:36 <^7heo> they can read english too :P 2014-10-23 12:50:47 \o/ 2014-10-23 12:51:07 <^7heo> \o/ 2014-10-23 13:30:32 3.0.6 is in the oven... 2014-10-23 13:30:39 <^7heo> \o/ 2014-10-23 13:30:48 <^7heo> so maybe I can DL the iso before EOB? 2014-10-23 13:31:37 you can see in #alpine-devel when its ready 2014-10-23 13:31:49 should be any second now 2014-10-23 13:45:11 ^7heo: build-3-0-x86_64: release v3.0.6 uploaded 2014-10-23 13:52:12 <^7heo> ncopa: awesome 2014-10-23 13:56:43 <^7heo> ncopa: http://nl.alpinelinux.org/alpine/v3.0/releases/x86_64/alpine-3.0.6-x86_64.iso right? 2014-10-23 13:58:52 yep, works for me! 2014-10-23 14:25:43 Oh wow, 3.0.6 \o/ ... I just updated my build-test-vm to 3.0.5 a few hours ago %) 2014-10-23 14:27:44 \o/ 2014-10-23 14:29:37 :) 2014-10-23 14:30:10 thanks all for testing/reporting and help fixing the issues 2014-10-23 14:33:10 <^7heo> ncopa: one remark: it'd be nice to have sshd NOT need net on the install disk 2014-10-23 14:35:23 <^7heo> http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually first thing I'm going to do. 2014-10-23 14:37:21 <^7heo> ncopa: how can I be sure that I have enough entropy when I do "dd if=/dev/urandom of=/dev/sda bs=4096" > 2014-10-23 14:37:24 <^7heo> ?* 2014-10-23 14:37:50 dunno 2014-10-23 14:37:57 isnt /dev/random for that? 2014-10-23 14:38:07 <^7heo> dunno 2014-10-23 14:38:10 <^7heo> basically 2014-10-23 14:38:16 iirc /dev/random will block when its out of entropy 2014-10-23 14:38:19 <^7heo> I'm going to fill a 240GB ssd with random stuff 2014-10-23 14:38:30 <^7heo> how can I feed more entropy? 2014-10-23 14:38:53 i think move mouse and use keyb 2014-10-23 14:39:01 but do that for 240GB... 2014-10-23 14:39:02 <^7heo> even in term? 2014-10-23 14:39:09 good luck... :) 2014-10-23 14:39:11 <^7heo> yeah that's the point. 2014-10-23 14:39:47 http://en.wikipedia.org/?title=/dev/random 2014-10-23 14:40:07 the /dev/urandom is supposed to be good enough for crypto 2014-10-23 14:40:27 <^7heo> yeah that's why I took urandom by default 2014-10-23 14:41:10 i think you'll have to feed entropy for /dev/random for years 2014-10-23 14:41:12 literally 2014-10-23 14:41:27 <^7heo> =/ 2014-10-23 14:41:35 <^7heo> how do the other system do that? 2014-10-23 14:41:38 <^7heo> like debian/ubuntu 2014-10-23 14:41:45 i think they just use /dev/urandom 2014-10-23 14:42:53 i saw some tool that grabbed noise from mic input and feeded that to the entropy pool 2014-10-23 14:42:59 i dont remember the name of it 2014-10-23 14:43:22 <^7heo> might be worth it 2014-10-23 14:43:29 <^7heo> would transform years on hours 2014-10-23 14:44:08 i am trying to get some binary to run on edge. ldd gives me Error loading shared library libnsl.so.1 2014-10-23 14:44:21 might be weeks, i dont know 2014-10-23 14:44:22 can't find it in packages 2014-10-23 14:44:32 mfehlhaber: precompiled binary? 2014-10-23 14:44:40 yes 2014-10-23 14:44:48 readelf -d 2014-10-23 14:45:02 make sure all the [NEEDED] are available 2014-10-23 14:45:05 and found 2014-10-23 14:45:41 musl should provide at least partial ABI compat with libc.so.6 2014-10-23 14:46:09 so if you have NEEDED libc.so.6 in there, then it might work to symlink it 2014-10-23 14:46:37 otherwise, you can copy libc.so.6 from debian/ubuntu/fedora something 2014-10-23 14:47:15 ^7heo: /dev/urandom is fine for everything except before it gets seeded at boot time 2014-10-23 14:47:34 <^7heo> indigo: like 2014-10-23 14:47:40 sorry i'm not used to alpine... apk search -v readelf finds nothing 2014-10-23 14:47:42 <^7heo> I can use urandom to feed a 240GB drive?! 2014-10-23 14:47:55 <^7heo> wait, how can that happen? Where does the entropy comes from? 2014-10-23 14:47:59 checking musl.. 2014-10-23 14:48:00 mfehlhaber: apk add binutils 2014-10-23 14:48:21 you can also use scanelf from pax-utils to do the same 2014-10-23 14:48:27 ^7heo: It's a CSPRNG, it only needs however many bits of entropy that its cipher requires in the beginning 2014-10-23 14:48:51 <^7heo> indigo: hmm, thanks for the info. I'm going to look that up right now. 2014-10-23 14:50:03 <^7heo> https://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator#NSA_backdoor_in_the_Dual_EC_DRBG_PRNG 2014-10-23 14:50:07 <^7heo> interesting :) 2014-10-23 14:50:34 seems i'm missing [libm.so.6 libnsl.so.1 libdl.so.2 libpthread.so.0 libc.so.6] 2014-10-23 14:51:25 no, i'm not... just libnsl.so.1 2014-10-23 14:52:20 I like the way the BSD's /dev/random device works 2014-10-23 14:52:38 It blocks until it gets enough entropy to seed the CSPRNG, and then does not block anymore 2014-10-23 14:53:55 mfehlhaber: i think those are from gnu libc 2014-10-23 14:54:19 i think so too 2014-10-23 14:54:31 <^7heo> indigo: do you know how the Linux implementation behaves when it doesn't have the required entropy? 2014-10-23 14:54:33 try create symlinks to musl libc 2014-10-23 14:55:01 ^7heo: http://en.wikipedia.org/?title=/dev/random 2014-10-23 14:55:15 <^7heo> ?? 2014-10-23 14:55:18 ^7heo: Yes, if you search online there are quite a few articles about /dev/urandom behaving poorly on routers and VMs if it doesn't get seeded property on startup. 2014-10-23 14:55:37 "When the entropy pool is empty, reads from /dev/random will block until additional environmental noise is gathered.[4]" 2014-10-23 14:55:47 <^7heo> okay 2014-10-23 14:55:57 <^7heo> but what's the diff between random and urandom then? 2014-10-23 14:56:16 "A counterpart to /dev/random is /dev/urandom ("unlimited"[5]/non-blocking random source[6]) which reuses the internal pool to produce more pseudo-random bits. This means that the call will not block, but the output may contain less entropy than the corresponding read from /dev/random. While it is still intended as a pseudorandom number generator suitable for most cryptographic purposes, it is not recommended for the generation of long-term 2014-10-23 14:56:16 cryptographic keys." 2014-10-23 14:56:16 ^7heo: /dev/random uses magic voodoo to estimate how much entropy you have 2014-10-23 14:56:33 <^7heo> did ttso implement it? 2014-10-23 14:57:01 ^7heo: http://www.2uo.de/myths-about-urandom/ should have more info 2014-10-23 14:57:15 <^7heo> vielen dank 2014-10-23 15:00:21 ln -s /lib/libc.musl-x86_64.so.1 libnsl.so.1 got me one step closer 2014-10-23 15:01:04 getting Error relocating ... backtrace: symbol not found 2014-10-23 15:02:17 mfehlhaber: backtrace is not implemented in musl 2014-10-23 15:02:20 i know that for sure 2014-10-23 15:02:49 i'd copy all the above gnu libc files from debian or so 2014-10-23 15:03:55 ncopa: ok i'll try that 2014-10-23 15:04:07 thanks 2014-10-23 15:11:02 hi, i get an error: 2014-10-23 15:11:08 alpine:~ sudo apk add linux-grsec-dev 2014-10-23 15:11:17 (1/1) Installing linux-grsec-dev (3.14.22-r1) 2014-10-23 15:11:33 ERROR: Failed to extract lib/modules/3.14.22-1-grsec/build: Read-only file system 2014-10-23 15:11:43 wget: short write 2014-10-23 15:11:49 ERROR: linux-grsec-dev-3.14.22-r1: Read-only file system 2014-10-23 15:11:55 1 errors; 269 MiB in 172 packages 2014-10-23 15:11:55 AmatCoder: you cannot install kernel on tmpfs setup 2014-10-23 15:12:17 well, you can if you unmount modloop: /etc/init.d/modloop stop 2014-10-23 15:12:38 but things will break on reboot if you install kernel (and lbu commit it) 2014-10-23 15:13:01 build files either? 2014-10-23 15:14:03 I need to compile a module 2014-10-23 15:15:51 i'd recommend using disk when compiling stuff 2014-10-23 15:16:04 development tend to use lots of space 2014-10-23 15:16:31 you could do it on you desktop linux (in an alpine linux lxcontainer) 2014-10-23 15:16:58 if you have enough ram on the box and dont mind loing your work on reboot, then you can /etc/init.d/modloop 2014-10-23 15:17:10 and apk add linux-grsec-dev shoudl work 2014-10-23 15:17:25 <^7heo> ncopa: did you stock up on snacks and drinks? 2014-10-23 15:17:35 <^7heo> 'cause you're going to do support quite some time it seems :D 2014-10-23 15:17:43 gotta go 2014-10-23 15:17:45 its late 2014-10-23 15:17:46 <^7heo> :P 2014-10-23 15:17:47 see u 2014-10-23 15:17:49 <^7heo> o/ 2014-10-23 15:17:51 <^7heo> have a nice one 2014-10-23 15:18:14 yes, I don't mind loosing my work, just I am testing... 2014-10-23 15:18:44 see you, thanks 2014-10-23 17:12:11 Nice: http://rationallyparanoid.com/articles/diskless-ssh-honeypot-alpine-linux.html 2014-10-23 17:17:48 :) 2014-10-23 17:18:07 i RT'd that; nice to get alpine some more exposure 2014-10-23 20:19:47 Hi. I have booted alpine from a live usb system, installed some packages, and I want to install it tomorrow to a hard disk. today, I want to backup the whole system to the USB stick that I used to boot the system. how do I do that? 2014-10-23 20:29:13 I see there is already the 'cache' directory on my usb stick. what exactly will persist when I reboot the system? 2014-10-23 20:30:35 I remember that I have selected 'usb' as the cache location 2014-10-23 20:30:35 quit 2014-10-23 20:36:29 currently I am running in diskless mode I think, and I want all changes I made (including packages and files in /home) to persist 2014-10-23 20:40:49 going to sleep, sorry. see you. 2014-10-23 21:00:07 <^7heo> I just found a problem with 3.0.6 2014-10-23 21:00:26 <^7heo> cryptsetup -y -v luksFormat /dev/sdx segfaults 2014-10-23 21:01:01 <^7heo> basically 2014-10-23 21:01:14 <^7heo> you yave to type the answers to the output without the output. 2014-10-23 21:01:16 <^7heo> if not 2014-10-23 21:01:18 <^7heo> it segfaults. 2014-10-23 21:01:47 <^7heo> or only the first YES actually 2014-10-23 21:02:50 <^7heo> funny, the "are you sure (type uppercase yes)" is delayed until the end of the exection 2014-10-23 21:03:50 buffering? 2014-10-23 21:04:04 <^7heo> maybe 2014-10-23 21:04:17 <^7heo> nevertheless, the behavior isn't consistent 2014-10-23 21:04:30 <^7heo> you type the command, then it "seems to hang" 2014-10-23 21:04:37 <^7heo> if you press enter it segfaults 2014-10-23 21:04:45 <^7heo> if you type anything and enter it segfaults 2014-10-23 21:04:57 <^7heo> the ONLY way to make it happen is to type YES (so to know what is expected) 2014-10-23 21:05:32 ^7heo: that's been happening to me since 3.0.1 when I started using alpine 2014-10-23 21:05:47 but it also happened to me on one debian xen machine so I dunno 2014-10-23 21:07:28 <^7heo> I straced the thing, the memory in which you type the answer is 1024 byte long, and memset to 0 2014-10-23 21:07:31 <^7heo> so it's not that 2014-10-23 21:08:01 <^7heo> must be some double free or something 2014-10-23 21:08:29 <^7heo> I expect it to be from the cleanup-exit code 2014-10-23 21:09:15 the prompt problem is probably some line-buffering mistake 2014-10-23 21:10:03 ^7heo: it uses gnu readline to read the response 2014-10-23 21:10:17 <^7heo> hmm 2014-10-23 21:10:20 <^7heo> prolly yes 2014-10-23 21:10:35 <^7heo> GNU is so incrusted into every and anything 2014-10-23 21:10:39 <^7heo> that's depressing 2014-10-23 21:10:48 is there a better readline library? 2014-10-23 21:10:53 readline makes using the command line so much easier... 2014-10-23 21:11:07 I've included it in tons of my code at work 2014-10-23 21:11:08 <^7heo> meh I dunno 2014-10-23 21:11:12 <^7heo> scanf is easy enough 2014-10-23 21:12:51 ahills: libtcl 2014-10-23 21:13:11 no, I mean for the user 2014-10-23 21:13:22 having ^A, ^E, ^W, ^U, command line history, etc 2014-10-23 21:13:26 letting them use their own .inputrc 2014-10-23 21:13:29 all that great stuff 2014-10-23 21:14:34 rlwrap will take care of all that 2014-10-23 21:15:08 ugh tcl 2014-10-23 21:15:18 if you find that you need more customizations than it allows, maybe it's about time to think of providing a proper interface for your program. you know, there were some advancements since teletypes. 2014-10-23 21:15:38 no, the idea is that the user can use their existing .inputrc which they use with literally every other readline program 2014-10-23 21:15:50 so that the command line in my software was an easy interface to adopt 2014-10-23 21:15:55 rlwrap will use it since it uses readline? 2014-10-23 21:16:12 right but that still requires readline to be on the system... 2014-10-23 21:16:20 <^7heo> damnit 2014-10-23 21:16:30 <^7heo> I did luksFormat 2014-10-23 21:16:33 <^7heo> on /dev/sda 2014-10-23 21:16:35 but now that's user's choice 2014-10-23 21:16:40 <^7heo> and I forgot the /boot... 2014-10-23 21:16:41 <^7heo> =/ 2014-10-23 21:16:45 <^7heo> ACTION stupid 2014-10-23 21:17:46 <^7heo> wait, how to make an ext2 fs in the alpine install disk? 2014-10-23 21:18:23 <^7heo> please ignore that, I've installed the e2fsprogs package 2014-10-23 21:21:52 ahills: Linenoise? 2014-10-23 21:22:26 https://github.com/antirez/linenoise ? 2014-10-23 21:22:37 looks col, thanks 2014-10-23 21:22:39 cool, even 2014-10-23 21:23:58 ahills: Are you anti-GNU? ;P 2014-10-23 21:24:36 <^7heo> how to tell the system to reload the parition table? 2014-10-23 21:24:40 <^7heo> I did fdisk my disk 2014-10-23 21:24:50 <^7heo> and now I need to reload it, because there's no sda1 or sda2 2014-10-23 21:24:59 <^7heo> and I don't know how to do it. 2014-10-23 21:25:18 <^7heo> indigo: not hardware anti-gnu, but I think that it has much more traction than it deserves. 2014-10-23 21:25:23 <^7heo> hardcore 2014-10-23 21:25:25 <^7heo> I need to sleep 2014-10-23 21:25:29 <^7heo> see you tomorrow. 2014-10-23 21:25:34 ^7heo: 'Night! 2014-10-23 21:31:14 well, fsf/gnu cloned unix and managed to keep things together for a long time, you have to give them that. 2014-10-23 21:33:32 it's easy to hate gnu, but they maintain a production quality compiler and you don't. 2014-10-23 21:36:05 Haha 2014-10-23 22:20:47 ovf: Yes, alpine/musl is a bit of counterculture against gnu. In the GNU utopia, all code is equal. Thou shall not kill code, no matter how ugly. And things get stacked one atop of another. 2014-10-23 22:20:58 we're here to break this vicious cycle of code decay 2014-10-23 22:23:49 see, i hate gnu just like the next guy. (and my ibm t23 runs plan 9). but i feel like someone's got to remind what compiler you ship by default. :-) 2014-10-23 22:53:34 indigo: no, but I don't like some of the software... just someone was complaining about readline, and I was wondering what the alternatives are 2014-10-23 22:53:42 since I use readline in lots of software, and generally like it 2014-10-23 22:54:09 Ah gotcha 2014-10-23 22:54:21 the API isn't too difficult to learn 2014-10-23 22:54:41 ovf: 9front? :o 2014-10-23 22:55:47 as an aside, this whole discussion started because of my typo. i meant to say gnu (as in glibc) getline, not readline 2014-10-23 22:56:12 indigo: that install predates 9front by a few years. 2014-10-23 23:03:06 Oh cool 2014-10-23 23:03:15 Haha you're the first person I've met that actually runs Plan 9 :P 2014-10-23 23:37:00 something broke in my X on alpine... 2014-10-23 23:37:22 modesetting_drv.so fails to load with missing symbol 2014-10-23 23:37:38 ms_shadowUpdate32to24 2014-10-23 23:38:04 this is after upgrade 2014-10-23 23:38:11 any idea what could be wrong? 2014-10-23 23:51:39 ncopa, the patch you added for modesetting on 9/11 seems to have broken it 2014-10-24 00:00:30 ovf, getline is in posix 2008 2014-10-24 00:00:53 *sigh* going to try rebuilding this package from source 2014-10-24 00:02:23 nobody else has hit this problem? 2014-10-24 00:02:48 modesetting_drv.so has an undefined symbol that's not defined anywhere as far as i can tell 2014-10-24 00:06:25 hmm, first going to try a binary patch... 2014-10-24 00:07:18 dalias: weird, it's right there in the master->sh3224 patch: http://git.alpinelinux.org/cgit/aports/commit/?id=c51482993c8410af07f96da6d9dc1dc01248c755 2014-10-24 00:08:41 yes, i wonder if somehow that .c file is not getting linked tho 2014-10-24 00:08:50 because it's definitely U in the nm -D output 2014-10-24 00:09:15 anyway i binary-patched the symbol table to refer to abort() instead of that symbol (hoping it won't be used) and now it loads but my X is still broken 2014-10-24 00:10:32 so i dunno wtf is wrong 2014-10-24 00:12:50 well, i think i see why it doesn't get linked in (no alpine system nearby, so static analysis, sorry). the patch (correctly) modifies Makefile.am to add sh3224.c, but nobody bothers to re-run autofools (APKBUILD just runs configure). 2014-10-24 00:13:00 ... 2014-10-24 00:14:27 hmm i'm getting "failed to set drm interface version 1.4: permission denied" in the Xorg log file 2014-10-24 00:34:04 how can i pull these packages from 3.0 rather than edge? 2014-10-24 00:34:13 i just want to try downgrading my whole X to alpine 3.0 2014-10-24 00:38:59 figured out how, and that fixed it 2014-10-24 00:41:21 back up and running :) 2014-10-24 00:46:24 the version in edge must be doing something really messed up because it leaves the fb in a corrupted state (and never even initializes the fb right to begin with as far as i can tell) 2014-10-24 00:47:00 all this with what video driver? 2014-10-24 00:47:55 intel 2014-10-24 06:17:21 mornign 2014-10-24 06:18:20 dalias: you mean xf86-video-modesetting? 2014-10-24 06:18:49 i added 2 different patches 9/11 2014-10-24 06:22:23 dalias: is it xf86-video-intel or xf86-video-modesetting? 2014-10-24 06:22:25 and what kernel 2014-10-24 06:22:41 i know grsec kernel causes issues with mesa due to sysfs protrection 2014-10-24 08:37:54 <^7heo> ncopa: moin 2014-10-24 08:38:29 hi 2014-10-24 08:38:58 dmesg 2014-10-24 08:39:17 <^7heo> ncopa: since you probably don't have the time to read the complete log 2014-10-24 08:39:49 <^7heo> ncopa: here is some info from me: GNU readline is used in cryptsetup and it makes it segfault if not used as expected. 2014-10-24 08:40:26 <^7heo> also, there is unexpected buffering happening, which prevents any normal user (who would rely on stdout) to operate the software as expected. 2014-10-24 08:40:47 sounds like bugs in readline and/or cryptsetup 2014-10-24 08:40:53 <^7heo> tl;dr: GNU readline (and maybe another thing) is making cryptsetup crash. 2014-10-24 08:41:09 <^7heo> well, I checked with strace, the segfault isn't happening during the scanf 2014-10-24 08:41:20 <^7heo> the buffer is memset to 0, 1024 byte long. 2014-10-24 08:41:34 <^7heo> and it's filled with the data you type in. 2014-10-24 08:41:41 <^7heo> so no worries on that side. 2014-10-24 08:41:48 <^7heo> should I submit a bug? 2014-10-24 08:41:54 yes 2014-10-24 08:42:09 and possibly also file a bug for upstream 2014-10-24 08:42:20 please post a stable way to reproduce it 2014-10-24 08:42:31 eg if you can do it with a temp file or loopback device 2014-10-24 08:44:18 <^7heo> ncopa: of course I'll post a way to reproduce, it's the main reason why I ask if I should submit a br 2014-10-24 08:44:32 <^7heo> I have worked into testing companies, so I think I know how to file a bug :P 2014-10-24 08:44:46 :) 2014-10-24 08:45:05 <^7heo> but if in anycase I'm not fitting your standards, please feel free to inform me :) 2014-10-24 08:45:23 <^7heo> basically, it's uber easy to reproduce 2014-10-24 08:45:28 <^7heo> 1. boot alpine iso, log as root 2014-10-24 08:45:36 <^7heo> 2. # apk add cryptsetup 2014-10-24 08:45:49 <^7heo> 3. # cryptsetup -y -v luksSetup /dev/sda 2014-10-24 08:45:58 <^7heo> 4. # press enter 2014-10-24 08:46:01 does it happen with tmpfile? 2014-10-24 08:46:08 instead of /dev/Sda 2014-10-24 08:46:19 <^7heo> I expect it to. It's input related afaik 2014-10-24 08:46:21 <^7heo> but I can test. 2014-10-24 08:46:24 <^7heo> right now. 2014-10-24 08:46:36 dd if=/dev/null of=tmpfile bs=1M count=10 2014-10-24 08:46:38 or so 2014-10-24 08:47:15 <^7heo> basically, I did dd if=/dev/urandom of=/dev/sda bs=4096 for the whole evening 2014-10-24 08:47:36 <^7heo> so I expect the data there to be semantically equivalent to null 2014-10-24 08:48:09 <^7heo> ncopa: when you say tmpfile, you mean "any file"? 2014-10-24 08:48:19 <^7heo> or is it a specific type of file? 2014-10-24 08:50:10 any file 2014-10-24 08:50:15 <^7heo> okay :) 2014-10-24 08:50:40 <^7heo> exact same behavior. 2014-10-24 08:50:43 <^7heo> As I said. 2014-10-24 08:50:52 <^7heo> stdin dependant, the rest of the program works as expected. 2014-10-24 08:52:09 <^7heo> raah naaah 2014-10-24 08:52:14 <^7heo> nah naaah naaaaaah! 2014-10-24 08:52:21 <^7heo> godamnit! 2014-10-24 08:52:38 <^7heo> I didn't note my b.a.o password... 2014-10-24 08:52:49 <^7heo> and I can't sign it now v_v 2014-10-24 08:52:55 <^7heo> that f***ing sucks! 2014-10-24 08:53:59 <^7heo> okay, I'll submit the br tonight 2014-10-24 08:54:02 <^7heo> sorry for that =/ 2014-10-24 08:54:27 can you reset the passwd? 2014-10-24 08:54:38 luksSetup 2014-10-24 08:54:46 cryptsetup: Unknown action. 2014-10-24 08:54:57 did you mean luksFormat? 2014-10-24 08:55:25 <^7heo> 1. I can reset the passwd but not access my email from work. So same problem. 2014-10-24 08:55:41 so you'll have to pay me toreset it for you ;) 2014-10-24 08:55:47 <^7heo> 2. complete command is: # cryptsetup -y -v luksFormat /irrelevant/file 2014-10-24 08:55:56 <^7heo> ncopa: how can I pay? 2014-10-24 08:55:59 <^7heo> can I send you beer? :D 2014-10-24 09:34:28 <^7heo> new bugs: #3470 & #3471 2014-10-24 11:12:33 ^7heo: you can pay by reporting it upstream to cryptsetup devs 2014-10-24 11:13:08 you can send them this patch: http://git.alpinelinux.org/cgit/aports/tree/main/cryptsetup/flush-stdout.patch 2014-10-24 11:13:21 <^7heo> ncopa: thanks 2014-10-24 11:14:30 <^7heo> ncopa: thanks a bunch for editing the br too :) 2014-10-24 11:16:07 <^7heo> damnit, there's no date on code.google.com/p//issues/list 2014-10-24 11:16:17 <^7heo> so I can't check if they still use that bug tracker or not =/ 2014-10-24 11:16:52 <^7heo> last comment there is from may 18, 2014... 2014-10-24 11:17:38 <^7heo> okay, it seems in use. Not very active but in use. Posting br. 2014-10-24 11:18:13 <^7heo> haha, flush stdout :D 2014-10-24 11:18:15 <^7heo> noobs. 2014-10-24 11:18:17 <^7heo> ACTION hides. 2014-10-24 11:18:37 i am amused that it does not happen with other libc 2014-10-24 11:18:39 <^7heo> or, "how a bad libc can lead you to have very poor coding habits" 2014-10-24 11:18:51 yes 2014-10-24 11:19:35 <^7heo> "the libc does that for me. I don't care wether it's good or bad, or even if in the POSIX specs and portable at all. I want GNU to dominate the world because nobody good can turn evil, right?!" 2014-10-24 11:20:07 <^7heo> anyway, I'm going to report it; I'll also copy a lot of data from #3471 2014-10-24 11:20:31 you can link to alpine bug tracker 2014-10-24 11:21:31 <^7heo> yeah maybe. 2014-10-24 11:21:39 <^7heo> ticket title: "stdout doesn't get flushed after fputs with non-GNU libc" 2014-10-24 11:21:50 <^7heo> do you have any improvement to suggest? 2014-10-24 11:22:35 or "yesDialog" does not display question prompt 2014-10-24 11:23:01 <^7heo> that's more specific about the bug, but less about the problem. 2014-10-24 11:23:04 "stdout doesn't get flushed after fputs with non-GNU libc" makes it sound that the problem is in the libc 2014-10-24 11:23:12 <^7heo> right 2014-10-24 11:23:21 <^7heo> I'd better put that in the description with more details. 2014-10-24 11:23:26 yesDialog does not flush stdout 2014-10-24 11:23:28 or similar 2014-10-24 11:23:46 <^7heo> yeah that last one is very good. 2014-10-24 11:24:33 problem is that the outputted string does not have '\n' 2014-10-24 11:24:38 so its not flushed automatically 2014-10-24 11:24:47 http://stackoverflow.com/questions/1716296/why-does-printf-not-flush-after-the-call-unless-a-newline-is-in-the-format-strin 2014-10-24 11:24:54 <^7heo> you're right 2014-10-24 11:24:59 i think its strange that it does not happen on glibc 2014-10-24 11:25:07 <^7heo> I forgot about that newline thing 2014-10-24 11:25:36 <^7heo> I think that the GNU libc has some wrapper around any puts that checks for a newline and flushes if it does not. 2014-10-24 11:25:42 <^7heo> or maybe flushes all the time anyway. 2014-10-24 11:30:05 <^7heo> oh damn 2014-10-24 11:30:20 <^7heo> I'm going to post the issue with my work profile... 2014-10-24 11:30:26 <^7heo> good thing I check that out =/ 2014-10-24 11:30:42 <^7heo> ncopa: do you have a google profile? 2014-10-24 11:31:26 <^7heo> if you do, lemme provide you with all the needed data, it only takes two clicks and two copy/paste. 2014-10-24 11:36:25 <^7heo> why the fuck are people using this google code stuff... 2014-10-24 11:36:41 <^7heo> I hate having to get a google account in order to post bugs 2014-10-24 11:37:46 <^7heo> I can't submit a bug without a google profile, seriously. 2014-10-24 11:37:57 yes that's bad 2014-10-24 11:39:32 yes, and a google profile is associated with phone numbers and other kraaken suction. die google, die. 2014-10-24 11:39:51 <^7heo> yep 2014-10-24 11:40:03 (when i found security critical bugs in the go crypto package (openpgp) i couldnt submit them (back then i didnt have mobile phone and google account creation required one here) so i had to spend a lot of effort to find round about ways to submit the patch) 2014-10-24 11:40:08 <^7heo> and all the profiling analysis on your communications 2014-10-24 11:40:18 <^7heo> what's the content of everything you have sent or received 2014-10-24 11:40:21 <^7heo> semantic content I mean 2014-10-24 11:40:41 <^7heo> and maybe even profiling on the way you communicate, not only on the words 2014-10-24 11:41:10 <^7heo> but sentence structure, vocabulary cardinality, frequency and distribution, 2014-10-24 11:41:13 <^7heo> etc. 2014-10-24 11:41:23 <^7heo> that can tell a lot already. 2014-10-24 11:45:20 <^7heo> so now I need to find a freaking google profile 2014-10-24 11:46:14 can't you just send a mail to one of the devs? 2014-10-24 11:46:22 or maybe there's a mailinglist? 2014-10-24 11:46:27 <^7heo> yeah, same problem. 2014-10-24 11:46:41 <^7heo> ffs, google asks for SO much data from you to create an account, now... 2014-10-24 11:46:52 you can send mails without having a google profile, no? 2014-10-24 11:46:58 <^7heo> yes 2014-10-24 11:47:30 <^7heo> I can send a mail to cle...@gmail.com 2014-10-24 11:47:32 so? problem solved? 2014-10-24 11:47:50 aah? really? they should state their email address in the (c) license 2014-10-24 11:47:52 <^7heo> but I guess I'll get a Notification back from the mail daemon 2014-10-24 11:47:57 <^7heo> hmm okay. 2014-10-24 11:48:06 <^7heo> I love how simple that is. 2014-10-24 11:48:28 <^7heo> yeah! :) 2014-10-24 11:48:34 <^7heo> they're going to love this. 2014-10-24 11:48:37 also i believe in the git/svn commit logs there might be reall addresses 2014-10-24 11:48:53 <^7heo> I'm going to send the mail to 3 different people 2014-10-24 11:51:22 <^7heo> ncopa: sent. 2014-10-24 11:53:03 thanks 2014-10-24 11:53:37 <^7heo> well, I merely sent an email to the people in the AUTHORS file. 2014-10-24 11:53:49 <^7heo> not sure how much that is going to help 2014-10-24 11:53:59 <^7heo> but it's the best I can do without a google profile. 2014-10-24 12:07:52 note that gmail can easily mark your mail as spam if you are not writing from a well known webmail service (which i learnt recently when i submitted patches to the mujs project) 2014-10-24 12:08:05 so yes this is painful 2014-10-24 12:09:58 actuall the bozos at google put me on their blacklist because one db update changed every record, and send a mail for each change to some people (also with gmail addresses) after a few 10 thousand they labeled me as a bulk mailer. and all their forms to get of the blacklist are redirected to /dev/null so i had to change ip addresses. motherfuckers. 2014-10-24 13:00:58 <^7heo> YX9KwoR6rk2l: thats Google for you. 2014-10-24 13:24:28 <^7heo> which musl is in 3.0.6? 2014-10-24 13:24:39 <^7heo> okay 1.1..4 2014-10-24 13:24:51 <^7heo> sorry for asking, I forgot I had a running system I could connect to. 2014-10-24 14:04:04 <^7heo> there's one thing I don't get in fdisk 2014-10-24 14:04:10 <^7heo> Units = cylinders of 16065 * 512 = 8225280 bytes 2014-10-24 14:04:24 <^7heo> Start: 1 2014-10-24 14:04:28 <^7heo> Units = sectors of 1 * 512 = 512 bytes 2014-10-24 14:04:34 <^7heo> Start: 63 2014-10-24 14:04:55 <^7heo> 1 * 8225280 bytes != 63 * 512 bytes; so what?! 2014-10-24 14:05:26 <^7heo> it's actually 255 times less... 2014-10-24 14:05:40 <^7heo> wtf happens here? 2014-10-24 14:21:01 Good afternoon all! 2014-10-24 14:22:16 <^7heo> yo 2014-10-24 14:36:03 hi 2014-10-24 14:36:06 where can I find libutil? 2014-10-24 14:36:16 empty.c:544:3: warning: implicit declaration of function 'login_tty' [-Wimplicit-function-declaration] 2014-10-24 14:36:16 login_tty(slave); 2014-10-24 14:36:16 ^ 2014-10-24 14:43:17 Zap-w: I think that login_tty is a function from glibc 2014-10-24 14:44:17 Zap-w: I don't think that it is implemented in musl 2014-10-24 14:44:45 Zap-w: but I am not an expert 2014-10-24 15:27:57 Have a great weekend everyone! (try not to miss me too much!) 2014-10-24 15:46:37 Hi ! 2014-10-24 15:53:18 I have an error with abuild: 2014-10-24 15:53:28 >>> ERROR: mednafen*: Found /usr/lib/charset.alias 2014-10-24 15:53:59 in postcheck 2014-10-24 16:04:07 yeah, a couple autogen scripts seem to create that, haven't figured out where it comes from 2014-10-24 16:04:20 just add a "rm" line after your make install 2014-10-24 16:06:21 yes, it is created by package 2014-10-24 16:06:26 thanks! 2014-10-24 16:07:41 ah damn, anyone has a vagrant alpine image? 2014-10-24 16:20:46 <^7heo> # lvm lvcreate -l100%FREE -n sys /dev/mapper/lvm 6e1d5e673000-6e1d5e8ad000 r-xp 00000000 00:0f 6306 /lib/libdevmapper.so.1.02: mlock failed: Out of memory 6e1d5e8b2000-6e1d5eab7000 r-xp 00000000 00:0f 687 /lib/libdevmapper-event.so.1.02: mlock failed: Out of memory Logical volume "sys" created 2014-10-24 16:20:51 <^7heo> wtf? 2014-10-24 16:21:03 fbettag: i thought i saw one on vagrantcloud a while back. seems to have gone 2014-10-24 16:21:47 ^7heo: looks messy in irc. can you sprunge it? 2014-10-24 16:22:04 <^7heo> ScrumpyJack: yes 2014-10-24 16:22:42 <^7heo> http://sprunge.us/XURB 2014-10-24 16:25:27 not a clue :) 2014-10-24 16:27:35 i take it you beleive it to be wrong, and you're not out of memory? :) 2014-10-24 16:28:59 #3107 2014-10-24 16:29:38 algitbot: #3107 2014-10-24 16:29:44 arg 2014-10-24 16:29:55 http://bugs.alpinelinux.org/issues/3107 2014-10-24 16:29:56 there 2014-10-24 16:30:31 ScrumpyJack: yeah i scoured google for it but didnt find it :/ 2014-10-24 16:36:15 ScrumpyJack: the error message is bullshit 2014-10-24 16:36:47 i get it even if i have 63GB ram available 2014-10-24 16:36:55 yeah, looks like #3107 agrees with you 2014-10-24 16:37:15 probably you need to put linux=high,umb on the kernel commandline 2014-10-24 16:37:17 *now* he wakes up? 2014-10-24 16:37:30 eyo darkfader 0) 2014-10-24 16:37:34 moin 2014-10-24 16:37:53 blimey, moin? what TZ? 2014-10-24 16:39:35 fbettag: wanna create a feature request for it on b.a.o? 2014-10-24 16:40:13 ScrumpyJack: i'd rather build a packer.io json description for how to auto-build them and then launch a feature request ;) 2014-10-24 16:40:27 just looked for how auto-builds are made 2014-10-24 16:40:30 https://bitbucket.org/ariya/packer-vagrant-linux/src/d0359d7e64e25a4ca6c2cfe53da80fb79d0dc8c4/ubuntu-12.04-amd64.json?at=master 2014-10-24 16:40:53 looks straight enough 2014-10-24 16:40:58 "boot_command": [ 2014-10-24 16:40:58 "", 2014-10-24 16:40:59 "/install/vmlinuz noapic preseed/url=http://{{ .HTTPIP }}:{{ .HTTPPort }}/ubuntu-12.04-amd64/preseed.cfg ", 2014-10-24 16:41:02 hihi 2014-10-24 16:41:03 esc esc enter wait 2014-10-24 17:05:19 <^7heo> the alpine setup script behaves poorly with lvm 2014-10-24 20:18:22 ^7heo http://bugs.alpinelinux.org/issues/3107#change-10868 2014-10-24 20:18:58 sorry didnt saw that ScrumpyJack wrote it already 2014-10-24 20:57:18 ScrumpyJack: https://github.com/wasted/alpine-packer 2014-10-24 20:57:41 ScrumpyJack: produces a vagrant box out of the alpine iso ;) 2014-10-24 20:57:59 girlfriend doesn't like that i've spent my evening on it :/ 2014-10-24 20:58:08 lol 2014-10-24 22:06:36 ScrumpyJack: the only problem with vagrant is that it requires virtualbox guest utils, which are not available on alpine :/ 2014-10-24 22:10:25 Has anyone successfully built libreSSL on alpine? 2014-10-24 22:13:20 systmkor3: havent tried it on alpine yet 2014-10-24 22:13:35 I was running into some issues with it yesterday 2014-10-24 22:13:37 ScrumpyJack: if http://git.alpinelinux.org/cgit/aports/plain/testing/virtualbox-additions-grsec/APKBUILD gets into main, vagrant will work ;) 2014-10-24 22:13:57 ncopa: how far is that along? :s 2014-10-24 22:20:26 what are the implentation or design differences with the apk package manager 2014-10-24 22:20:47 compared to other package managers 2014-10-24 22:28:19 ncopa: what are the general design principles of the apk package manager 2014-10-24 22:28:47 and why go with yet another package manager? 2014-10-24 22:50:21 lol, it's described in full on the wiki 2014-10-24 22:51:43 and it works almost exactly like pkg_add on openbsd, but with more features 2014-10-24 23:08:22 are there any 'contrib' repositories for alpine? 2014-10-24 23:27:34 khm: the google repo ;) 2014-10-25 00:01:21 ahills: the wiki discusses the command line interface and that is about it 2014-10-25 00:01:35 you can also use the help file, etc 2014-10-25 00:01:35 ahills: whichs is not what I am talking about 2014-10-25 00:01:53 the design is not very complex, you can infer the entirety from the usage 2014-10-25 00:02:02 ahills: if you mean apk --help 2014-10-25 00:02:06 again that is just the UI 2014-10-25 00:02:16 or the man page 2014-10-25 00:02:21 anyways apk and abuild make up most of it 2014-10-25 00:02:22 not how it's designed or varied in implementation 2014-10-25 00:03:55 ahills: yup as far as I gathered apk, abuild, and alpine-sdk comprise just about everything for the build environment 2014-10-25 00:04:42 ahills: reason I ask is I am looking at pkgin and nix package managers and see how they deal with undefined state, rollbacks, deterministic builds etc. 2014-10-25 00:04:50 ahills: and what apk does to do that 2014-10-25 00:05:54 ahills: e.g. nix has a pure functional language for their package build language 2014-10-25 00:06:35 ahills: the other reason I am asking since ncopa is one of the developers is that I am curious what the goals of the package manager ar 2014-10-25 00:06:54 just to be different/role with your own 2014-10-25 00:07:12 re-implementation of another package manager (which I wonder why not just use that) 2014-10-25 00:07:36 or to be one of the most robust, secure, and well designed package managers 2014-10-25 00:08:03 then I am curious what the road map is to there, what design referenence is being used, and why not just use the nix package manager 2014-10-25 00:17:02 systmkor3: I assume that they don't use nix's package manager because of the complexity involved 2014-10-25 00:17:23 indigo: in what way 2014-10-25 00:17:40 indigo: while I would agree there is definitely some upfront complexity costs 2014-10-25 00:18:00 overall it appears to dramatically simplify and stabalize package management 2014-10-25 00:18:53 Right, in exchange for an increased implementation complexity 2014-10-25 00:20:03 yes agreed 2014-10-25 00:20:46 indigo: but if the marginal increase in implementation complexity produces a net sum of less system complexity 2014-10-25 00:20:52 why wouldn't that be a good trade-off? 2014-10-25 00:21:11 indigo: in addition of not having to maintain a custom package manager 2014-10-25 00:21:12 systmkor3: Because then you end up with things like systemd. 2014-10-25 00:21:28 indigo: that is a weird leap in logic 2014-10-25 00:21:52 indigo: I am not suggesting everything become some weird monolith 2014-10-25 00:21:59 Not really... systemd init files are very simple and concise 2014-10-25 00:22:08 A lot of the logic is handled via the implementation 2014-10-25 00:22:30 systemd is the cancer that has befallen the GNU-lands 2014-10-25 00:22:46 fbettag: ehhhhhh, not sure about that 2014-10-25 00:23:01 fbettag: I would argue that people need to put pressure on systemd not being a monolith implementation 2014-10-25 00:23:02 Breaking POSIX, from the guy who brought us the pulse-audio root vuln 2014-10-25 00:23:03 Given that Alpine includes musl instead of glibc, busybox instead of coreutils, they seem to be looking for things that have simpler implementations 2014-10-25 00:23:21 So it's just a different focus :) 2014-10-25 00:23:23 systmkor3: it can'T get any worse 2014-10-25 00:23:36 systmkor3: oh sorry, it can 2014-10-25 00:23:43 yah... 2014-10-25 00:23:47 fbettag: you could have Windows 2014-10-25 00:23:48 systmkor3: by having your init system allocate a fixed pid during the whole uptime 2014-10-25 00:23:57 killing pid 1 meaning your system is rebooting 2014-10-25 00:24:14 fbettag: again not saying that the current implementation is correct 2014-10-25 00:24:28 systmkor3: it's totally incorrect by any standard 2014-10-25 00:24:29 just saying their more nebulus goals are on the right path 2014-10-25 00:24:44 systmkor3: the "do one thing, and do it right" 2014-10-25 00:24:56 this simple philosophy that brought us great pieces of code 2014-10-25 00:25:10 indigo: I would say the musl from the comparison chart I was given in #musl shows that it is mainly just as feature complete but minus a significant amount of code 2014-10-25 00:25:16 just thrown out for some weird fuck who didnt understand POSIX and said "hey, lets disregard it" 2014-10-25 00:25:38 systmkor3: sorry, touchy topic 2014-10-25 00:25:43 no problem 2014-10-25 00:25:51 whenever i read systemd my heartrate is up 2014-10-25 00:25:53 fbettag: but one thing that is neglected is that re-implementation of the same configuration 2014-10-25 00:25:55 fbettag: It's not that bad once you get used to it >:) 2014-10-25 00:25:57 via init scripts 2014-10-25 00:26:05 also produces poor implementation and configuration 2014-10-25 00:26:22 indigo: It appears to be the same wiht coreutils v. busybox 2014-10-25 00:26:47 systmkor3: busybox is missing quite a few features compared to coreutils 2014-10-25 00:26:55 indigo: i was just about to say that 2014-10-25 00:27:06 but back to the tradeoffs 2014-10-25 00:27:17 it would seem busybox works better for their goals 2014-10-25 00:27:21 Yep :) 2014-10-25 00:27:23 or what I assume the goals of alpine are 2014-10-25 00:27:42 and as far as I can tell it looks like nix package manager looks inline with those goals 2014-10-25 00:27:53 while an initial increase in implementation complexity 2014-10-25 00:28:00 indigo: i'd rather penetrated my testicles with hot needles 2014-10-25 00:28:08 and put it on youtube 2014-10-25 00:28:13 fbettag: do it 2014-10-25 00:28:14 :P 2014-10-25 00:28:26 systmkor3: not as long als there is alpine ;) 2014-10-25 00:28:34 and upstart on ubuntu :P 2014-10-25 00:28:40 and openrc on gentoo 2014-10-25 00:28:47 fbettag: I heard Ubuntu was ditching upstart for systemd 2014-10-25 00:28:47 and just shell scripts on bsd 2014-10-25 00:28:58 fbettag: see I would say shell scripts would be worse 2014-10-25 00:29:00 systmkor3: not in 14.04 LTS ;) 2014-10-25 00:29:05 ahhh 2014-10-25 00:29:16 i'm stuck for another 4 years 2014-10-25 00:29:17 :D 2014-10-25 00:29:22 maybe they ditch it by then 2014-10-25 00:29:29 eehhh unlikely 2014-10-25 00:29:38 wait for the first 2-3 root exploits 2014-10-25 00:29:39 fbettag: but if you code a fork of systemd removing the crap 2014-10-25 00:29:45 or get people behind moving the crap out of it 2014-10-25 00:29:48 systmkor3: that's what openbsd logind is doing 2014-10-25 00:29:51 probably more likely then Ubuntu wil ditch it 2014-10-25 00:30:06 also openrc is still superior in quality to systemd 2014-10-25 00:30:14 shellscript even more so 2014-10-25 00:30:21 since they've been around for decades and still work ;) 2014-10-25 00:30:24 (for me at least) 2014-10-25 00:30:25 i can understand openrc 2014-10-25 00:30:29 but shell script.. no 2014-10-25 00:30:35 and even if you want something fundamentally better 2014-10-25 00:30:43 solaris smf has been around eons 2014-10-25 00:30:53 and it still beats the shit out of systemd 2014-10-25 00:31:11 =) 2014-10-25 00:31:18 darkfader will love me for mentioning smf 2014-10-25 00:31:18 yes but solaris is Oracle and Oracle is probably one of the worst forces in computer science 2014-10-25 00:31:26 nah, they killed mysql 2014-10-25 00:31:29 best thing ever 2014-10-25 00:31:41 ehh to increase Oracle DB 2014-10-25 00:31:51 so ehhh not really best thing ever 2014-10-25 00:32:01 i dont see mysql as a database 2014-10-25 00:32:06 it's more of a dump 2014-10-25 00:32:09 like /tmp 2014-10-25 00:32:13 =) 2014-10-25 00:32:28 ehhh you may see it that way but technically it is a db 2014-10-25 00:32:39 but I am not trying to stand up for mysql 2014-10-25 00:32:54 good ting 2014-10-25 00:32:55 :D 2014-10-25 00:32:58 indigo: so what would be a good way to see the interopability of nix or something similar with alpine 2014-10-25 00:32:59 nobody should 2014-10-25 00:33:01 anyway 2014-10-25 00:33:02 off to bed 2014-10-25 00:33:06 n8 everyone 2014-10-25 00:33:11 systmkor3: Port it to alpine and check it out ;) 2014-10-25 00:33:21 well running on alpine at the moment 2014-10-25 00:33:25 haven't ported it 2014-10-25 00:33:45 indigo: i sent an email to the mailing list to ask about apk design 2014-10-25 00:33:47 but no response yet 2014-10-25 00:35:30 well I could probably get the program installed 2014-10-25 00:35:55 but I would have to port their existing build scripts to have PaX flags 2014-10-25 00:35:57 etc. 2014-10-25 00:36:42 indigo: also curious about integration of lxc, cgroups, and grsec policy with package management 2014-10-25 00:36:58 19:24 < fbettag> khm: the google repo ;) 2014-10-25 00:37:00 what does this mean 2014-10-25 00:37:05 maybe a flag or something for a package 2014-10-25 00:37:26 systmkor3: Not sure, haven't really delved into that 2014-10-25 00:37:29 like how it seems that all alpine packages like to keep the program seperate from the documentation which makes sense 2014-10-25 00:37:51 indigo: so like gcc, gcc-doc, gcc-lxc, gcc-cgroup, gcc-policy 2014-10-25 00:38:11 wer the gcc-lxc enables gcc to be ran within a container 2014-10-25 00:38:27 gcc may not be a good example but 2014-10-25 00:39:29 i want reproducible builds, thrustworthy sigs on all sources, quality https for downloading. does nix do that? does apk do that? 2014-10-25 00:39:29 Hmmm why would there be a separate package to enable a program to run in a container 2014-10-25 00:39:58 indigo: basicall a modifier on top of the main package, I am not certain 2014-10-25 00:40:02 just spit balling 2014-10-25 00:40:15 I think it something that should be discussed / examined more closely 2014-10-25 00:40:37 YX9KwoR6rk2l: I can't verify that 2014-10-25 00:40:53 YX9KwoR6rk2l: however nix is shooting for atomic package upgrades 2014-10-25 00:41:01 and reproducibility 2014-10-25 00:41:29 YX9KwoR6rk2l: I don't know what their common build flags for programs are and I don't know how locked down sigs on sources are 2014-10-25 00:41:56 YX9KwoR6rk2l: their design philosophy assumes building from source but opts for pulling pre built binaries when possible 2014-10-25 00:42:36 YX9KwoR6rk2l: quality https, that I assume would be more related to the mirrors that they currently run 2014-10-25 00:42:39 YX9KwoR6rk2l: I think nix is aiming for deterministic builds in the future 2014-10-25 00:42:44 https://nixos.org/wiki/Long-term_open_issues:build_determinism 2014-10-25 00:43:14 indigo: just seems like their is similar goals with nix, so why not assimilate :P 2014-10-25 00:44:41 systmkor3: Heh, I think it's completely different 2014-10-25 00:45:07 indigo: i don't think so 2014-10-25 00:45:16 I know :) 2014-10-25 00:45:25 ah :P 2014-10-25 00:47:46 indigo: okay well I'll mess around with making a port of it 2014-10-25 00:47:55 Cool :D 2014-10-25 00:49:05 yah I have a few things i am hoping ported 2014-10-25 00:49:38 libressl, dtrace, 3.16 kernel 2014-10-25 00:49:52 and also get ARM more mainline with alpine 2014-10-25 00:50:25 oh and more software defined programs e.g. openvswitch 2014-10-25 00:51:53 will be back on in a couple of hours 2014-10-25 13:14:36 <^7heo> moin @ 2014-10-25 13:15:18 moin 2014-10-25 14:59:22 moin 2014-10-25 14:59:39 anyone got an idea if vboxsf is available as module somewhere? 2014-10-25 16:12:50 <^7heo> fbettag: I think not. 2014-10-26 08:42:42 Hi all. I want to have 5 similar Alpine setup running from USB. Which best practive is recommended to generate identical setups in a row? I would like to have openvswitch, openssh-server, qemu, rsyslog, ssmtp, nbd-client, ntpdate et collectd installed and configured on every usb key with same configs.. 2014-10-26 08:43:48 the rsa pub key for root should be included in /root/.ssh/authorized_keys2 also 2014-10-26 08:43:57 and that's it 2014-10-26 08:45:09 they will be used to launch qemu instance from NBD images 2014-10-26 10:25:10 <^7heo> Jean-Scotch: write a config file, then 2014-10-26 10:25:24 <^7heo> cat config-file | nc -lp 4810 2014-10-26 10:25:30 <^7heo> and on the client 2014-10-26 10:25:37 <^7heo> $ ip link set eth0 up 2014-10-26 10:25:48 <^7heo> $ udchpc eth0 2014-10-26 10:26:02 <^7heo> $ nc 4810 | sh 2014-10-26 10:26:34 My goal is not even to log into the machine. just fire an usb key already configured. 5 times the same one. 2014-10-26 10:26:37 <^7heo> that way you can deploy the same config file to 5 different hosts 2014-10-26 10:26:44 <^7heo> ah 2014-10-26 10:27:14 <^7heo> it's easier to use the nc way, because you can correct the config VERY easily as you progress. 2014-10-26 10:27:15 I will send the keys by post to some young guy and he will go plug in into the DC ;) 2014-10-26 10:27:44 I will see the new machine in the dhcp log 2014-10-26 10:27:54 <^7heo> but yeah if you're not close to the clients, it's a bad idea to use anything else than a totally automatic setup. 2014-10-26 10:28:05 the login by ssh (the key wust already be present) and continue from there 2014-10-26 10:28:06 <^7heo> dhcp leases but yeah same idea :P 2014-10-26 10:28:18 <^7heo> wait 2014-10-26 10:28:25 I'm 1000 km away from the machines 2014-10-26 10:28:46 <^7heo> you indend to install 5 machines via ssh by hand typing "setup-alpine" and all?! 2014-10-26 10:30:14 I will then start a qemu from the real HDD and migrate theyr contents to the new storage cluster (ceph). the restart the VMs from the storage cluster trough a custom proxy between RBD and NBD as Alpnine still is not a CEPH client yet. 2014-10-26 10:30:31 notI ant the key to be already prepared. 2014-10-26 10:30:36 <^7heo> wait, are you installing the virt host or guests? 2014-10-26 10:30:48 the only différence will be the IP address received by the dhcp server 2014-10-26 10:32:57 rsyslog preconfigured to sennd all his log to some central servr, ntp configured to use local time server; ssmtp configured to use local relay; collectd configured to push collected data to local server; ssh pub key setup for root to connect passwordless; modules for nbd and kvm loaded; openvswitch configuref to offer 3 differents OVS. 2014-10-26 10:34:20 I can always prepare the keys by hand. no big deal. but I was wondering if a best practise for such use case was already published somewhere. 2014-10-26 10:35:00 I not, I will publish my notes on this topic. I yes, I may contribute to it if I change some procedural parts 2014-10-26 10:35:33 I'm in the process of preparing the bare metal install 2014-10-26 10:35:52 This mean the virt host I guess 2014-10-26 10:37:13 when up & running I will launch other kvm on top of it from with images pulled from some NBD server on the LAN 2014-10-26 10:39:33 I suppose I may prepare one by hand the copy it's overlay file accross several newly prepared keys... 2014-10-26 10:39:43 *then 2014-10-26 14:36:55 <^7heo> http://dilbert.com/strips/comic/2002-03-10/ 2014-10-27 07:08:18 morning 2014-10-27 07:08:24 ncopa: what are the general design principles of the apk package manager 2014-10-27 07:08:53 and why go with yet another package manager? 2014-10-27 07:09:00 those are reasonable questions 2014-10-27 07:09:05 specially the last 2014-10-27 07:09:15 we should add that to the FAQ or similar 2014-10-27 08:20:38 Hi all! 2014-10-27 09:02:51 <^7heo> moin 2014-10-27 14:00:04 clandmeter, uselessd is -5 already :) 2014-10-27 14:00:21 You might want to update it. Or wait for -6 that is in few days 2014-10-27 14:08:39 fr0stb1te: ill update it. 2014-10-27 14:08:45 are you using it? 2014-10-27 14:15:38 Nope, I'm mantaining uselessd for Arch 2014-10-27 14:15:53 The actual developer asked me to poke you :) 2014-10-27 14:24:46 fr0stb1te: http://tinyurl.com/o48bulo 2014-10-27 14:26:55 weee, thx :) 2014-10-27 14:27:51 You've removed the checksums and didn't add new ones 2014-10-27 14:27:55 Is that in purpose? 2014-10-27 14:29:06 what do you mean? 2014-10-27 14:29:09 <^7heo> we have arsch maintainers here? 2014-10-27 14:29:12 <^7heo> =/ 2014-10-27 14:29:23 <^7heo> this place has much lower standards than I thought. 2014-10-27 14:29:37 clandmeter, the APKBUILD has no checksums. 2014-10-27 14:29:46 hmm you are right 2014-10-27 14:29:51 dunno what happend 2014-10-27 14:30:02 Heh, stuff happens 2014-10-27 14:30:45 I keep pushing pkgbuilds to my personal build server and have my builds fail because of the weirdest errors on my part 2014-10-27 14:31:18 It builds btw. 2014-10-27 14:31:30 fr0stb1te: thx for reminding. updated now. 2014-10-27 14:32:18 Yep. Everything is fine now. 2014-10-27 14:32:29 I'll boot my alpine instance with it when I get home 2014-10-27 14:32:42 Might forget tho 2014-10-27 14:33:25 ^7heo, you know, excluding users of other systems is how the arch community became the toxic waste it is today. 2014-10-27 14:33:35 Well. One of the reasons. 2014-10-27 14:34:26 fr0stb1te: not sure the pkg will be sync yet 2014-10-27 14:34:32 we are cleaning up testing repo 2014-10-27 14:34:42 too much junk in there 2014-10-27 14:35:54 Eh, I can build it locally 2014-10-27 14:35:59 Not afraid of that :) 2014-10-27 14:41:51 <^7heo> fr0stb1te: imho it's because the developers are arrogant pricks. 2014-10-27 14:42:38 <^7heo> which *causes* what you're saying, but unlike what you just said, being arrogant pricks directly (as opposed to indirectly) caused arch to be a pile of scheiße 2014-10-27 14:43:05 <^7heo> anyhow, back to work. 2014-10-27 15:17:46 what's the problem with the libressl license? 2014-10-27 15:54:48 I really need a filesystem that supports a swap file... 2014-10-27 15:59:18 clandmeter: xfs doesn't work? :o 2014-10-27 16:00:28 im using Btrfs 2014-10-27 16:02:55 Ah 2014-10-27 16:02:59 Didn't know that was stable yet 2014-10-27 16:23:52 <^7heo> isn't it btrfs still unstable?! 2014-10-27 16:41:15 <^7heo> I don't know for you, but when I use lvm lvcreate, I get a Out Of Memory error. 2014-10-27 17:11:15 hahahah 2014-10-27 17:11:27 i just made some connection between btrfs and rng 2014-10-27 17:39:14 ^7heo: Same here 2014-10-27 17:42:15 clandmeter, btrfs now works ok ? 2014-10-27 17:52:27 kernel v3.18 has lots of fs related changes + overlayfs 2014-10-27 17:56:12 gtg 2014-10-27 18:41:15 <^7heo> jomat: I did create straces of both lvm lvcreate and lvm lvremove 2014-10-27 18:41:23 <^7heo> I'll read them tomorrow. 2014-10-27 18:41:30 <^7heo> I had way too many tickets to manage today 2014-10-27 18:41:38 same here 2014-10-27 18:41:44 <^7heo> :P 2014-10-27 18:42:04 <^7heo> please tell me those tickets weren't Jira tickets, please. 2014-10-27 18:42:11 But besides of the oom-errors, lvm works 2014-10-27 18:42:25 <^7heo> did you install alpine with it? 2014-10-27 18:42:43 ^7heo: We have OTRS at work... a bunch of crap in perl.... but still better than java /o\ 2014-10-27 18:42:55 <^7heo> genau 2014-10-27 18:43:03 <^7heo> I'd take everything else than Jira. 2014-10-27 18:43:07 ^7heo: Yes, my alpine server runs with this ooming lvm 2014-10-27 18:43:21 <^7heo> I even have a ticket in Jira to solve a major bug about Jira... 2014-10-27 18:43:22 creating a raid1-volume even thorows more oom 2014-10-27 18:43:48 <^7heo> possibly, I don't plan to do that. 2014-10-27 18:43:59 <^7heo> how did you install alpine with it? 2014-10-27 18:44:30 at first i booted the server into the rescue system my isp offers 2014-10-27 18:44:40 <^7heo> here, the install script doesn't detect the lvm volumes. 2014-10-27 18:44:48 <^7heo> ah, chroot install? 2014-10-27 18:44:51 created a linear lvm volume as they're missing the raid1-kernel-module 2014-10-27 18:45:18 mounted all that and ... lemme think... ah, there's somewhere a static apk package 2014-10-27 18:45:33 and with that i installed alpine-base and such stuff into the mounted lv 2014-10-27 18:45:56 fixed the bootloader, added lvm-kernelmodules to the ramdisk, and some other stuff 2014-10-27 18:46:10 booted alpine and converted the linear volume to a raid1 volume 2014-10-27 18:46:20 more details later, i have to catch my train now 2014-10-27 18:47:06 <^7heo> that should be done automatically by the install scripts 2014-10-27 18:47:38 <^7heo> at least with flags 2014-10-27 18:47:47 <^7heo> I don't mind having to read some doc' 2014-10-27 18:47:52 <^7heo> or having to type additional parameters 2014-10-27 18:48:21 i don't like scripts 2014-10-27 18:48:31 they tend to do more stuff then i need 2014-10-27 18:49:18 <^7heo> no problem with that, but a step by step doc would be better then. 2014-10-27 18:49:43 <^7heo> problem is, scripts save you an AWFUL lot of typing. 2014-10-27 18:50:22 <^7heo> type once, reuse forever. 2014-10-27 18:50:31 <^7heo> and then share, 'cause why re-invent the wheel, right? 2014-10-27 18:53:42 <^7heo> wait a minute... 2014-10-27 18:53:59 <^7heo> when I "apk search qemu-system" I get a LOT of results... 2014-10-27 18:54:16 <^7heo> are those ALL for the x86_64 platform!!!? 2014-10-27 18:56:00 <^7heo> okay, installing qemu conflicts with xen, but nothing is prevented, so it just fails during package install. 2014-10-27 18:56:35 <^7heo> btw, I wonder why xen has a "qemu-bridge-helper" file... seems like a very dirty hack (ripoff) to me :) 2014-10-27 19:05:17 <^7heo> downloading Alpine from http://mirror.yandex.ru/mirrors/alpine//v3.0/releases/x86/alpine-3.0.6-x86.iso 2014-10-27 19:05:22 <^7heo> it says 3h 2014-10-27 19:05:24 <^7heo> =/ 2014-10-27 19:06:35 <^7heo> funny enough 2014-10-27 19:06:40 <^7heo> http://nl.alpinelinux.org/alpine/v3.0/releases/x86_64/alpine-3.0.6-x86_64.iso takes only 6 minutes 2014-10-27 19:07:17 <^7heo> I advise to use a better geodetection if you don't want people do all take from nl 2014-10-27 19:13:35 <^7heo> okay, battery ran out 2014-10-27 19:15:01 <^7heo> laters 2014-10-27 19:30:17 anyone here using LXC containers on alpine? I can't seem to use the "lxc-attach" command 2014-10-27 19:31:45 re 2014-10-27 19:33:05 buckley310: yes 2014-10-27 19:33:13 buckley310: see dmesg 2014-10-27 19:34:58 jomat: ahh, thanks. i will poke at it some more. 2014-10-27 19:48:34 buckley310: I use lxc-console or ssh as long as I am not sure what I'm doing by disabling lxc-attach :-) 2014-10-27 19:50:03 <^7heo> back 2014-10-27 19:50:11 <^7heo> the monologue guy is coming back 2014-10-27 19:51:16 jomat: yeah, im googling it but it looks like i cant fix it without changing some compile-time kernel setting. im going to stick with ssh... 2014-10-27 19:58:29 ^7heo: ever used cat in interactive mode? ;-) 2014-10-27 19:58:50 buckley310: You can change it at runtime... i think the info is in this link, but it doesn't load in this train: dhcpcd[23694]: wwan0: soliciting an IPv6 router 2014-10-27 19:58:53 dhcpcd[23694]: wwan0: carrier lost 2014-10-27 19:58:57 wrong link 2014-10-27 19:58:59 http://en.wikibooks.org/wiki/Grsecurity/The_RBAC_System#Capability_Restrictions 2014-10-27 20:02:38 buckley310: echo 0 > /proc/sys/kernel/grsecurity/chroot_caps 2014-10-27 20:20:21 jomat: thanks :) 2014-10-27 20:21:04 glad i could help somehow 2014-10-27 20:21:05 npnp 2014-10-27 20:27:15 <^7heo> jomat: yes, ofc. And the emacs psychologue too 2014-10-27 20:27:36 :-D 2014-10-27 20:28:05 <^7heo> someone should try to package despotify in alpine 2014-10-27 20:28:17 what is it? 2014-10-27 20:28:38 does it have dependencies? 2014-10-27 20:28:44 which are missing on alpine 2014-10-27 20:28:48 <^7heo> it's a spotify open source client 2014-10-27 20:29:06 <^7heo> dependencies: pulseaudio backend 2014-10-27 20:29:09 I'll try openscad next 2014-10-27 20:29:09 <^7heo> libssl-dev zlib1g-dev libvorbis-dev libtool 2014-10-27 20:29:20 <^7heo> libgstreamer-plugins-base0.10-0 libgstreamer0.10-dev # For GStreamer backend 2014-10-27 20:29:25 <^7heo> libao-dev # For libao backend 2014-10-27 20:29:26 I hate poettering software 2014-10-27 20:29:40 <^7heo> yeah but you're not forced to use the pulseaudio backend, fortunately. 2014-10-27 20:29:50 <^7heo> I'd say libao or gstreamer 2014-10-27 20:29:54 I just realized it 2014-10-27 20:30:12 <^7heo> and byw 2014-10-27 20:30:17 <^7heo> btw* 2014-10-27 20:30:17 ^7heo: can you package mono for me? :-D 2014-10-27 20:30:22 I want to play openra! 2014-10-27 20:30:23 <^7heo> I mixed the dependencies 2014-10-27 20:30:31 <^7heo> jomat: die in flames. 2014-10-27 20:30:35 <^7heo> not mono. 2014-10-27 20:30:44 yeah :'-/ 2014-10-27 20:30:53 <^7heo> not that shitty bloatware. You can't complain about java and praise mono 2014-10-27 20:31:15 I don't praise it, I just need it for OpenRA /o\ 2014-10-27 20:31:21 <^7heo> it's like complaining about islamic integrists and being part of a christian church 2014-10-27 20:32:19 ACTION isn't part of neither of them 2014-10-27 20:32:41 good bless 2014-10-27 20:32:50 *god 2014-10-27 20:33:54 <^7heo> dog bless. 2014-10-27 21:05:24 <^7heo> damnit 2014-10-27 21:05:47 <^7heo> I can't get the setup-disk -L to accept a luks volume... 2014-10-27 21:08:47 <^7heo> I might be doing something wrong here. 2014-10-27 21:11:26 <^7heo> yes indeed, the most recent documentation has the correct intel 2014-10-27 21:11:28 <^7heo> \o/ 2014-10-27 21:14:36 <^7heo> \o/ 2014-10-27 21:14:48 <^7heo> algitbot: don't die on me, you K9 2014-10-27 21:38:23 <^7heo> algitbot: ping 2014-10-27 21:38:26 <^7heo> =/ 2014-10-27 21:55:08 re 2014-10-28 09:56:40 <^7heo> jomat: wb 2014-10-28 09:57:30 :-) 2014-10-28 10:01:31 <^7heo> yeah sorry for the delay 2014-10-28 10:03:55 hi ! 2014-10-28 10:04:44 good morning all 2014-10-28 10:42:02 morning 2014-10-28 13:12:05 Hi 2014-10-28 13:21:45 hi 2014-10-28 13:23:15 quick question 2014-10-28 13:23:19 how cna I trust a repo ? 2014-10-28 13:23:57 packages are signed 2014-10-28 13:24:37 beyond that, you have to trust us. :) 2014-10-28 13:35:43 eh, that's not waht I meant ^^ 2014-10-28 13:36:11 ERROR: http://dl.2f30.org/alpine/v3.0/main/: UNTRUSTED signature 2014-10-28 13:39:20 is it your own repo? 2014-10-28 13:39:37 nope 2014-10-28 13:40:10 I tried extracting the key in the APKINDEX to /etc/apk/keys, but that's not it 2014-10-28 13:41:05 no 2014-10-28 13:41:11 our keys are in alpine-keys 2014-10-28 13:41:35 but its inside alpine-base, which should be installed by default. 2014-10-28 13:42:21 z3bra: do you have the same issue with nl.a.o? 2014-10-28 13:42:27 I don't 2014-10-28 13:42:42 I just want to use some friendly repo 2014-10-28 13:42:54 rather than recreating the packages 2014-10-28 13:42:57 then it sounds supicious 2014-10-28 13:43:00 or the pkgs are broken 2014-10-28 13:43:06 why ? 2014-10-28 13:43:18 untrusted sig 2014-10-28 13:43:19 I trust the guy who packaged it 2014-10-28 13:43:34 packaged it? 2014-10-28 13:43:44 or the owner of the mirror? 2014-10-28 13:56:41 on the same subject, how do I add an untrusted signature to my keys so that this error doesn't occur? 2014-10-28 13:57:54 i know you can use apk --allow-untrusted, but how to you add the key used to sign packages to my local apk environment? 2014-10-28 13:58:09 add it to /etc/apk/keys 2014-10-28 13:58:13 your pub key 2014-10-28 13:58:17 add what? 2014-10-28 13:58:40 where is the key used to sign packages? :) 2014-10-28 13:58:42 when you build packages, you should have setup a key pair 2014-10-28 13:59:02 check our wiki. it should be documented 2014-10-28 13:59:13 i think abuild will also complain about it and give you a hint 2014-10-28 13:59:31 i couldn't find anything on the wiki 2014-10-28 14:02:44 there is this: 2014-10-28 14:02:46 http://wiki.alpinelinux.org/wiki/Abuild_and_Helpers#abuild-keygen 2014-10-28 14:03:22 "The public key should be distributed and installed into /etc/apk/keys on the alpine box that will install the packages. This basically means that the main developer's public keys should be in /etc/apk/keys on all Alpine boxes" 2014-10-28 14:06:53 clandmeter, it's the same person 2014-10-28 14:07:34 z3bra: i dont understand. this person builds his own pacakges and has his own repo? 2014-10-28 14:07:58 He devel softwares, package them and share them on this repo 2014-10-28 14:08:26 ok 2014-10-28 14:08:33 you can do 2 things 2014-10-28 14:08:38 you can ask his pub key 2014-10-28 14:08:43 and put it manually 2014-10-28 14:08:48 ok, i get it. i'll update the wiki to be a bit more clear 2014-10-28 14:09:09 or he can package it into forinstance: devname-keys.apk 2014-10-28 14:09:19 Isn't the public key shared in the APKINDEX ? 2014-10-28 14:09:47 from what i know, its inside alpine-keys.apk 2014-10-28 14:13:02 clandmeter: chicken and egg though, how can you apk add devname-keys without his key ;) 2014-10-28 14:13:30 answer, you use apk --allow-untrusted 2014-10-28 14:13:54 apk add --allow-untrusted key.apk 2014-10-28 14:13:58 :) 2014-10-28 14:14:18 and if you dont trust it this way, you can ask him to email it to you. 2014-10-28 14:14:50 i'll put a howto together on using third party repros 2014-10-28 14:15:17 i like the idea of 2014-10-28 14:15:27 1) package your key 2014-10-28 14:15:56 2) add your URL to /etc/apk/repositories 2014-10-28 14:16:22 3) apk update && apk --allow-untrusted my_key.apk 2014-10-28 14:16:36 simple 2014-10-28 14:26:09 ScrumpyJack, the process of adding 3rd party key should be more transparent 2014-10-28 14:26:17 than using .apk 2014-10-28 14:26:18 ScrumpyJack: if you add the repo, you dont have to speciffy the .apk 2014-10-28 14:26:32 cd /etc/apk/keys 2014-10-28 14:26:44 wget path/to/new/key 2014-10-28 14:26:47 apk add whatevernameithas 2014-10-28 14:27:25 note that pkgs .apk can contain install pre/post... hooks 2014-10-28 14:27:29 vkrishn: it depends how the dev provides the key's 2014-10-28 14:27:38 scripts, kinda not ok 2014-10-28 14:28:01 yes, I prefer the key addition to be more transparent 2014-10-28 14:28:25 its not your choice, it depends on the developer :) 2014-10-28 14:28:53 but of course you can teach whatever way you guys prefer. 2014-10-28 14:28:58 yes, that's what I am trying to put, something that can be accepted as BCP 2014-10-28 14:29:11 i'd prefer to use apk to add the key than wget. why do you want it to be transparent? i think the user should be aware of what he's doing as he's adding an untrusted source 2014-10-28 14:29:54 with all the implications 2014-10-28 14:30:07 coz, a mitm .apk pgs can be exploited 2014-10-28 14:30:29 and wget cant? 2014-10-28 14:30:30 ncopa: any comment on intial repo key? 2014-10-28 14:31:35 clandmeter: if you had an untrusted repo, you still need to add the key of the untrusted repo 2014-10-28 14:31:36 less likely, and less prone, coz almost anything today is susceptible to mitm 2014-10-28 14:32:03 untrusted repo is not in question 2014-10-28 14:32:15 is method of adding the key 2014-10-28 14:33:05 I'm talking about the method of adding the key of an untrusted repo :) 2014-10-28 14:33:13 well, there can be scripts in the key apk, but then agian, there could be scripts in any of the packages of the untrusted repo. You have to trust the developer. 2014-10-28 14:33:14 hi 2014-10-28 14:33:33 like I said "untrusted repo" is not related here 2014-10-28 14:34:09 how pub key + repo works: 2014-10-28 14:34:14 you get a pub key 2014-10-28 14:34:24 you add it to your /etc/apk/keys 2014-10-28 14:34:25 vkrishn: read back to the begining 2014-10-28 14:34:34 another method is submitting you key to trusted site like AL, but AL would only sign it not confirm to trustworthiness of repo 2014-10-28 14:34:38 your key 2014-10-28 14:35:10 now apk will check that the packages you use in repositories are from the same person that gave you the pub key 2014-10-28 14:35:42 that is what it does, nothing more nothing less 2014-10-28 14:36:38 so if you get the key in insecure manner, then you cannot really trust the repo either 2014-10-28 14:37:09 so i think it makes sense to distribute the key separately from the repo itself 2014-10-28 14:37:10 worst, you already executed that .apk with --allow-untrusted 2014-10-28 14:37:41 with possible scripts in it 2014-10-28 14:38:08 then if you want attack the repo, you'll have to attack both places 2014-10-28 14:38:20 transparent key allows some scope of cross verification 2014-10-28 14:38:34 yes 2014-10-28 14:39:07 the key only guarantee you that the package you install is from the same person that gave you the key 2014-10-28 14:40:00 yes, that's what I am trying 2014-10-28 14:40:01 to say 2014-10-28 14:40:10 or more precise, the key only proves that the package is built by someone who has the private key matching the pub key 2014-10-28 14:40:12 vkrishn: then dont install our iso 2014-10-28 14:40:15 a well co-ordinated mitm 2014-10-28 14:40:18 it has alpine-keys 2014-10-28 14:40:29 same process, but runs in the background 2014-10-28 14:40:51 clandmeter, AL is not in question 2014-10-28 14:41:03 the keys on the iso only guarantees that the packages you install later are from same person that created the iso 2014-10-28 14:41:07 its method of 3rd party key 2014-10-28 14:41:23 does it matter? 2014-10-28 14:41:29 3rd party or not 2014-10-28 14:41:34 hmmm... 2014-10-28 14:41:52 like i told you 2014-10-28 14:42:00 you can copy it 2014-10-28 14:42:11 its *transparant* 2014-10-28 14:42:22 no scripts involved 2014-10-28 14:42:22 it's all about trust. you will not be able to remove the need for a certain level of trust. 2014-10-28 14:42:32 yes 2014-10-28 14:42:32 one is just trying to get the key to the person that one things belong to, and not key jacked by mitm 2014-10-28 14:42:34 but ones you apk add, you add packages from the repo 2014-10-28 14:42:40 thinks 2014-10-28 14:42:50 in the alpine iso case 2014-10-28 14:42:58 if you trust http://alpinelinux.org 2014-10-28 14:42:58 so the only thing we discuss here, how do you add the key. 2014-10-28 14:43:10 then you can download iso from mirror and verify checksum 2014-10-28 14:43:28 yes, cross verify 2014-10-28 14:43:49 there needs to be a way to cross verify when in doubt 2014-10-28 14:44:13 i think it makes sense to ship repo key from different channel than the repository itself 2014-10-28 14:44:46 yes thats probably the case 2014-10-28 14:44:52 it will help detect if one of the channels are bad 2014-10-28 14:44:53 maye send to the ml signed 2014-10-28 14:45:05 AL can accept them , sign them, but not guarantee them (necessarily) 2014-10-28 14:45:15 hmm 2014-10-28 14:45:33 are we talking about AL shipping 3rd party repo keys? 2014-10-28 14:45:33 hmmm... :) 2014-10-28 14:45:52 they could with repo, not asking 2014-10-28 14:46:18 AL shouldn't ship 3rd party keys, surely 2014-10-28 14:46:25 but not necessarily with repo 2014-10-28 14:46:27 ok good. thanks :) 2014-10-28 14:47:08 providing with repo, carries responsibility 2014-10-28 14:47:17 however small 2014-10-28 14:47:30 and i don't think AL should sign keys or give any level of confidence to a user, as it could be a false confidence 2014-10-28 14:47:40 i'd like to postpone the 3rd party repo discussion til after v3.1 2014-10-28 14:47:46 does anyone else do that? 2014-10-28 14:47:47 ok, 2014-10-28 14:47:51 gtg 2014-10-28 14:47:52 for now i'd like to focus on getting v3.1 out 2014-10-28 14:48:06 ScrumpyJack: i agree. i dont want do that 2014-10-28 15:01:50 Hi, I have a issue 2014-10-28 15:02:17 file ownership from apkovl is not preserved when extracting... 2014-10-28 15:03:32 I guess that busybox tar is the culprit 2014-10-28 15:04:25 have you tried with gnu tar? 2014-10-28 15:04:32 apk add tar 2014-10-28 15:05:40 hmm, but at boot busybox is used always, is not? 2014-10-28 15:05:41 i'd surpised if it were a busybox tar :) 2014-10-28 15:06:03 ah, so this is on boot? you are not trying manually? 2014-10-28 15:06:28 no, at boot 2014-10-28 15:06:35 have copied the apkvol from somewhere else? 2014-10-28 15:06:53 no, it is created by 'lbu' 2014-10-28 15:07:15 onto a local HDD or a memorysitck? 2014-10-28 15:07:59 I have an diskless installation 2014-10-28 15:08:09 boot from usb and runs in RAM 2014-10-28 15:08:24 and the apkvol is on the USB stick? 2014-10-28 15:08:29 yes 2014-10-28 15:08:36 which version? 2014-10-28 15:08:39 of AL 2014-10-28 15:08:44 3.0.6 2014-10-28 15:08:56 let me try that out 2014-10-28 15:12:55 I made a simple test creating and extracting a few files 2014-10-28 15:13:07 result: GNU-tar preserves ownership of files, busybox-tar not 2014-10-28 15:14:25 ooo, good. i don't need to test it :) 2014-10-28 15:14:43 at boot, apkvol is extracted by 'root' so all files have 'root:root' ownership 2014-10-28 15:14:45 would you mind creating a bug report with your findings? 2014-10-28 15:15:06 I do 2014-10-28 15:15:21 oh 2014-10-28 15:15:45 I mean 'I will do' 2014-10-28 15:16:00 ah :) thank you very much 2014-10-28 15:16:14 ;) 2014-10-28 15:49:12 By the way, why is the cgroup memory controller disabled? 2014-10-28 15:54:07 i can't remember ;) 2014-10-28 15:55:55 afaik it has to be enabled by a kernel boot parameter even if it's compiled in/available as module 2014-10-28 15:59:20 ScrumpyJack: I have created the bug report: http://bugs.alpinelinux.org/issues/3492 2014-10-28 16:00:44 ok , I think I found a way to cross-verify key.apk, kinda manual but ok 2014-10-28 16:01:04 ScrumpyJack, so ok with key.apk 2014-10-28 16:40:17 <^7heo> I have a funny bug 2014-10-28 16:40:25 <^7heo> with dual screen, xrandr. 2014-10-28 16:40:40 <^7heo> the cursor isn't disappearing from the first screen when I go to the second... 2014-10-28 16:40:52 <^7heo> can anyone confirm that? 2014-10-28 16:45:59 no 2014-10-28 16:46:12 intel driver here 2014-10-28 16:46:24 <^7heo> hmm 2014-10-28 16:46:29 <^7heo> I have to check my driver, you're right. 2014-10-28 16:54:53 <^7heo> with intel drivers it doesn't happen anymore. 2014-10-28 16:54:57 <^7heo> Thanks for the heads up 2014-10-28 17:00:04 <^7heo> okay, login manager sucks. 2014-10-28 17:00:10 <^7heo> there's only lxdm and sim 2014-10-28 17:00:12 <^7heo> slim even 2014-10-28 17:00:35 <^7heo> the former behaves poorly and the latter depends on dbus... 2014-10-28 17:51:34 ^7heo: try xdm 2014-10-28 17:51:37 @testing 2014-10-28 17:51:38 ;-) 2014-10-28 17:51:51 no dbus dependency 2014-10-28 17:52:23 i think both, slim and lxdm depend also on polkit or so 2014-10-28 17:53:36 ^7heo: Hm... xorg-server also has dbus as dependency 2014-10-28 18:10:55 <^7heo> ?! 2014-10-28 18:11:01 <^7heo> but it's not started 2014-10-28 18:12:26 <^7heo> oh wait 2014-10-28 18:14:01 <^7heo> nop, nothing running with dbus 2014-10-28 18:22:44 rc-update add xdm ; rc-service xdm start 2014-10-28 18:23:02 Ah... dbus 2014-10-28 18:23:05 no idea 2014-10-28 18:23:29 cups and wpa_supplicant also bring dbus :-( 2014-10-28 18:23:43 and wireshark needs bash 2014-10-28 19:10:59 pretty sure you can use wpa_supplicant and cups without dbus 2014-10-28 19:14:00 You can 2014-10-28 19:14:23 I use wpa_supplicant with a custom bash wifi scanner on my lappie 2014-10-28 19:14:32 And cups on my work machine. 2014-10-28 19:14:37 Both have no dbus 2014-10-28 19:34:12 fr0stb1te: Do you have the wifi scanner script online somewhere? 2014-10-28 20:22:23 <^7heo> I'm wondering... I get"/usr/include/xf86drm.h:40:17: fatal error: drm.h: No such file or directory" when I compile. Shouldn't xf86drm.h contain a valid path to drm.h? 2014-10-28 20:22:39 <^7heo> (it's missing the leading "libdrm/") 2014-10-28 20:23:33 <^7heo> same for xf86drmMode.h 2014-10-28 20:33:35 <^7heo> also, does anyone know where strdup is defined? because I don't find it in the included headers... 2014-10-29 00:09:30 <^7heo> okay, I built tmux. 2014-10-29 00:09:36 <^7heo> sorry, redshift 2014-10-29 00:09:44 <^7heo> I still have a slight problem however. 2014-10-29 00:10:16 <^7heo> abuild -r fails after the build, because it tries to package, and I didn't do anything for a package... 2014-10-29 00:10:41 <^7heo> it's my first APKBUILD so I'm not too sure what to do with the package section. 2014-10-29 00:11:45 <^7heo> any idea, anyone? :P 2014-10-29 00:26:39 <^7heo> okay, I think I've found the way to make it happen. 2014-10-29 00:31:21 <^7heo> redshift packaged. \o/ 2014-10-29 00:31:29 <^7heo> ACTION is looking forward to package more stuff. 2014-10-29 00:54:39 <^7heo> well, good night 2014-10-29 00:54:40 <^7heo> o/ 2014-10-29 09:09:30 <^7heo> moin 2014-10-29 09:21:40 <^7heo> ncopa: ping 2014-10-29 09:22:05 hi 2014-10-29 09:22:12 ^7heo: pong 2014-10-29 09:22:29 <^7heo> ncopa: :) 2014-10-29 09:22:32 <^7heo> ncopa: did you get my email 2014-10-29 09:22:36 <^7heo> ? 2014-10-29 09:22:52 probably... 2014-10-29 09:22:57 :) 2014-10-29 09:23:05 my boss does that same 2014-10-29 09:23:13 <^7heo> I don't remember what was the subject 2014-10-29 09:23:23 <^7heo> clandmeter: I can paste the patches in the chanel if you prefer ;D 2014-10-29 09:23:25 send me an email, and then starts calling me if i received it :p 2014-10-29 09:23:37 sure 2014-10-29 09:23:43 i dont handle ml patches. 2014-10-29 09:23:51 it is probably mingling with the other 287 unread emails in my inbox 2014-10-29 09:23:51 its troublesome to apply them 2014-10-29 09:24:07 <^7heo> ncopa: can you search for "redshift" in your emails? :P 2014-10-29 09:24:16 sprunge or github is fine for me. 2014-10-29 09:24:20 or any other repo 2014-10-29 09:24:29 found it 2014-10-29 09:24:33 <^7heo> super :) 2014-10-29 09:24:41 <^7heo> think you can try it? 2014-10-29 09:24:45 ^7heo: at this point i am mostly interested in wrapping up v3.1 2014-10-29 09:24:49 not adding new things 2014-10-29 09:25:13 source="https://github.com/jonls/redshift/releases/download/v1.9/redshift-1.9.tar.xz 2014-10-29 09:25:19 <^7heo> is that wrong? 2014-10-29 09:25:35 <^7heo> it's the source he/she/they provide on the website 2014-10-29 09:25:44 replace 1.9 with $pkgver 2014-10-29 09:25:49 <^7heo> oh right 2014-10-29 09:25:55 then i only need change version one place when upgrading 2014-10-29 09:26:04 <^7heo> well, and the patches. 2014-10-29 09:26:13 oh you want maintain it yourself 2014-10-29 09:26:24 <^7heo> why not. 2014-10-29 09:26:32 <^7heo> unless you want more work 2014-10-29 09:28:37 <^7heo> my goal is to help in order to have a linux distribution that can be used as industrial standard. 2014-10-29 09:28:57 <^7heo> So I'm not going to stop at maintaining one cosmetic package... 2014-10-29 09:29:07 nice 2014-10-29 09:29:08 thanks 2014-10-29 09:29:19 i appreciate help with maintaining 2014-10-29 09:29:25 <^7heo> thank me when I will have found the time to help. 2014-10-29 09:29:28 it looks like a cool app 2014-10-29 09:29:35 <^7heo> it's very useful. 2014-10-29 09:29:58 <^7heo> I actually wrote the makefile myself, and patched the code to remove all the unnecessary c99 dependencies 2014-10-29 09:30:16 <^7heo> 'cause the c99 std was causing much more errors than it was solving. 2014-10-29 09:30:32 <^7heo> so it was easier to revert to ANSI C. 2014-10-29 09:30:53 i wonder if it sould be possible to do something like that, but that checks the brightness/colortemp with the webcam 2014-10-29 09:30:59 oh, you patched it 2014-10-29 09:31:03 try send the patches upstream 2014-10-29 09:31:16 <^7heo> meh, not sure if they will accept it. 2014-10-29 09:31:27 <^7heo> the for (int i = 0... is probably there on purpose. 2014-10-29 09:32:18 <^7heo> and I did the makefile to avoid relying on autohell. 2014-10-29 09:32:26 <^7heo> 18 lines of makefile to avoid autohell 2014-10-29 09:32:34 <^7heo> I'd say that it's a good tradeoff. 2014-10-29 09:32:54 <^7heo> and btw, I wondered... 2014-10-29 09:33:09 <^7heo> I did get: "/usr/include/xf86drm.h:40:17: fatal error: drm.h: No such file or directory" when I compiled. 2014-10-29 09:33:16 <^7heo> Shouldn't xf86drm.h contain a valid path to drm.h? 2014-10-29 09:33:27 oh 2014-10-29 09:33:29 <^7heo> (i.e. #include ) 2014-10-29 09:33:49 not necesarily 2014-10-29 09:34:07 you normally set -I/usr/include/libdrm with pkg-config --cflags 2014-10-29 09:34:19 <^7heo> the includes should all be relative to /usr/include/ ? 2014-10-29 09:34:32 gcc -I 2014-10-29 09:34:40 <^7heo> yeah, it's what I did add in the makefile 2014-10-29 09:34:51 <^7heo> but I was still wondering if that behavior was desired. 2014-10-29 09:35:14 you normally use pkg-config --cflags xf86drm 2014-10-29 09:35:22 <^7heo> mokay 2014-10-29 09:35:25 witch should have Depends: drm 2014-10-29 09:35:34 so it should just work 2014-10-29 09:35:39 <^7heo> okay 2014-10-29 09:35:42 with pkg-config 2014-10-29 09:35:45 <^7heo> yeah 2014-10-29 09:35:59 <^7heo> I have to get more xp with the toolstack. 2014-10-29 09:36:11 but really 2014-10-29 09:36:16 we want stay close with upstream 2014-10-29 09:36:23 <^7heo> so, no patches? 2014-10-29 09:36:27 if upstream are not good enough, try fix it with upstream 2014-10-29 09:36:32 preferible no patches 2014-10-29 09:36:37 <^7heo> makes sense 2014-10-29 09:36:41 <^7heo> it's less work to maintain. 2014-10-29 09:36:45 yes 2014-10-29 09:36:50 <^7heo> however 2014-10-29 09:36:57 it makes huge difference when you have 1000+ packages to maintain 2014-10-29 09:37:05 <^7heo> we can't educate people and tell them that "autohell" suck, can we? 2014-10-29 09:37:13 <^7heo> yeah I know. 2014-10-29 09:37:30 <^7heo> my solution for that is to distribute the package maintainance. 2014-10-29 09:37:39 <^7heo> not to patch upstream. 2014-10-29 09:38:00 <^7heo> it's easier to distribute than patch upstream, because you don't have to convince the people in charge of upstream. 2014-10-29 09:38:30 the alpine policy is: try patch as little as possible 2014-10-29 09:38:41 <^7heo> yeah, that was clear already :P 2014-10-29 09:39:07 autotools is not my favorite, but i respect if upstream want use it 2014-10-29 09:39:15 <^7heo> and I try to give the feedback: I think it's going to make it harder in the end. Politics aren't really a good way to go for being efficient. 2014-10-29 09:39:42 its normally not worth patching out autotools just because we dont like it 2014-10-29 09:40:10 if you want do it, take it up with upstream 2014-10-29 09:41:01 <^7heo> mokay. 2014-10-29 09:41:18 <^7heo> I try to judge what's better on the long run... 2014-10-29 09:41:35 <^7heo> 'cause for sure, politics aren't a good option. 2014-10-29 09:41:45 <^7heo> but at the same time, forking everything isn't an option at all. 2014-10-29 09:42:34 <^7heo> ncopa: so, to be clear: you're not going to use that APKBUILD; because it patches too much, right? 2014-10-29 09:42:41 <^7heo> or? 2014-10-29 09:43:48 I can look at it 2014-10-29 09:43:57 but i will likely not use it as is 2014-10-29 09:44:03 <^7heo> okay 2014-10-29 09:44:22 <^7heo> then I'm going to contact upstream and discuss the use of autohell and c99 2014-10-29 09:46:11 c99 might be okish 2014-10-29 09:46:23 but i think you can also add -std=c99 to gcc 2014-10-29 09:46:26 or similar 2014-10-29 09:46:28 <^7heo> I did. 2014-10-29 09:46:41 <^7heo> and it caused a LOT of errors with "implict declaration" of functions 2014-10-29 09:49:42 ugh, now i see why you patched out autotools 2014-10-29 09:49:49 they just tag it 2014-10-29 09:49:54 <^7heo> yeah 2014-10-29 09:49:55 autotools has a: make dist 2014-10-29 09:50:00 to generate tarballs 2014-10-29 09:50:03 <^7heo> yeah 2014-10-29 09:50:05 lazy devs 2014-10-29 09:50:08 <^7heo> it's not really usesd. 2014-10-29 09:50:25 <^7heo> a makefile is really better here. 2014-10-29 09:50:35 <^7heo> and I'm not sure if they will change anything... 2014-10-29 09:50:45 <^7heo> 'cause 'autohell is the industrial standard' right? 2014-10-29 09:50:56 yeah... :-/ 2014-10-29 09:51:22 normally the configure blah takes longer than the acutal compile 2014-10-29 09:51:29 <^7heo> yes it does. 2014-10-29 09:51:39 its so stupid that i lack words for it 2014-10-29 09:51:40 <^7heo> actual compilation time on my laptop: 0.6 2014-10-29 09:52:02 and often the configure script is bigger than the actual application 2014-10-29 09:52:04 <^7heo> the PRINTING only from autohell would take more than the double of that. 2014-10-29 09:52:09 <^7heo> yes. 2014-10-29 09:52:19 they use macros instead of functions 2014-10-29 09:52:25 <^7heo> that's okay, I prefer that style. 2014-10-29 09:52:30 <^7heo> it's less jumps in the end. 2014-10-29 09:52:39 <^7heo> but yeah, granted, we're not counting jumps anymore :P 2014-10-29 09:52:41 due to support for some legacy system that didnt support functions 2014-10-29 09:52:58 and i bet they dont support that legacy system nowdays anyways 2014-10-29 09:53:05 <^7heo> yeah well 2014-10-29 09:53:27 have been thinking of a autotools killer 2014-10-29 09:53:32 <^7heo> if my memory is good, autohell is searching for like 27 different fortran compilers, even if it's 100% unrelated to the project at hand 2014-10-29 09:53:58 some configury script that has the useful stuff 2014-10-29 09:54:03 like check pkg-config packages 2014-10-29 09:54:13 check symbols in libs 2014-10-29 09:54:16 check defines 2014-10-29 09:54:19 check headers 2014-10-29 09:54:24 <^7heo> well, if you would replace autohell by a self-sustainable script, I'd jump for it 2014-10-29 09:54:32 <^7heo> but ideally 2014-10-29 09:54:35 <^7heo> you git clone 2014-10-29 09:54:35 <^7heo> make 2014-10-29 09:54:36 <^7heo> done. 2014-10-29 09:54:39 <^7heo> right? 2014-10-29 09:54:46 not that easy 2014-10-29 09:54:50 <^7heo> (provided that you have all the deps) 2014-10-29 09:54:52 so much to do, so little time... 2014-10-29 09:54:56 you might want do a preflight check 2014-10-29 09:55:06 <^7heo> ncopa: make can do that. 2014-10-29 09:55:06 like, does my libc have strlcpy? 2014-10-29 09:55:14 <^7heo> ah that kind of check. 2014-10-29 09:55:24 <^7heo> clandmeter: yeah, but I want to help, so I'm learning. 2014-10-29 09:55:25 you might want set prefix=/usr 2014-10-29 09:55:40 so you dont need set it when you do make install 2014-10-29 09:55:50 thats conveninet 2014-10-29 09:55:55 so a generated config.h 2014-10-29 09:56:00 and a generated config.mk 2014-10-29 09:56:02 aint bad 2014-10-29 09:56:34 you might want have optional features or alternative implementations 2014-10-29 09:56:55 like --with-ssl=openssl --with-ssl=polarssl 2014-10-29 09:59:44 <^7heo> yeah 2014-10-29 11:29:10 <^7heo> ncopa: btw, when using cryptsetup + lvm, the cherry keyboards don't work with the initramfs 2014-10-29 12:29:30 dalias: i think i fixe your xf86-video-modesetting issue: http://git.alpinelinux.org/cgit/aports/commit/main/xf86-video-modesetting?id=81ed1c9da8eb6d5a1e1d29710caa603e1a876a1e 2014-10-29 13:09:30 ncopa, even without that (i binary patched the call to the undefined symbol to a call to abort, and X didn't abort, so I can conclude it wasn't being used), i was still had failures 2014-10-29 13:09:57 screen mode getting utterly corrupted, unusable on console, after running X 2014-10-29 13:12:00 hum 2014-10-29 13:12:42 maybe i should just remove the fedora patch? 2014-10-29 13:12:53 i was hoping it fixed issue in qemu 2014-10-29 13:12:57 is that the only change between the version in 3.0 and the version in edge? 2014-10-29 13:13:12 i'm guessing not 2014-10-29 13:13:20 no 2014-10-29 13:13:27 i think xorg server was updated 2014-10-29 13:13:44 and xf86-video-modesetting in 3.0 is 0.8.1 2014-10-29 13:13:51 i edge it is 0.9.0 2014-10-29 13:17:41 so it very well could be something unrelated that's broken :( 2014-10-29 13:29:34 <^7heo> ncopa: you said you wanted to push 3.1 out. What's missing? 2014-10-29 13:30:39 build servers 2014-10-29 13:31:13 there are some changes in abuild that should be pushed as release 2014-10-29 13:31:18 apk-tools too 2014-10-29 13:34:17 <^7heo> when you say that build servers are missing, you mean: you're missing machines!? 2014-10-29 13:34:35 <^7heo> or are you talking about software servers? 2014-10-29 13:35:16 LXC containers 2014-10-29 13:35:40 <^7heo> =/ 2014-10-29 13:36:00 <^7heo> fabled: it could be the software to run build servers, I dunno. 2014-10-29 13:36:25 it's the 'virtual machine' instances 2014-10-29 13:36:42 <^7heo> I know what a LXC container is, thanks. 2014-10-29 13:37:04 <^7heo> however, I can't contribute as easily if I have to provide machines... 2014-10-29 13:37:05 i am updating icu now 2014-10-29 13:37:10 needs rebuild lots of things 2014-10-29 13:37:15 likw webkit, firefox 2014-10-29 13:37:24 so im trying to clean up firefox while at it 2014-10-29 13:37:32 remove xulrunner split 2014-10-29 13:37:46 <^7heo> which version? 34? 2014-10-29 13:41:03 31esr 2014-10-29 14:11:42 <^7heo> ok 2014-10-29 14:17:30 ncopa, no further findings on why firefox behaves weird? 2014-10-29 14:17:40 (like things on the blank tab page not working right) 2014-10-29 14:17:42 <^7heo> dalias: please define "weird" :) 2014-10-29 14:17:53 <^7heo> dalias: mozilla I said :P 2014-10-29 14:17:58 <^7heo> I stand by my point. 2014-10-29 14:18:10 ^7heo, sorry, we'd talked about it before and i figured ncopa would know the context 2014-10-29 14:18:54 i still just want to track down whether it's an issue in musl, firefox, or some other library it's using 2014-10-29 14:21:08 dalias: here firefox is working fine 2014-10-29 14:21:22 dalias: there were problems with earlier versions... 2014-10-29 14:21:32 <^7heo> AmatCoder: what version are you using? 2014-10-29 14:21:40 <^7heo> 'cause I still have some memory problems with 29 2014-10-29 14:21:48 33 2014-10-29 14:21:55 <^7heo> yeah okay. 2014-10-29 14:23:00 dalias: which version fo ff? 2014-10-29 14:23:04 dalias: ...but it was fixed when xulrunner was removed as dependency 2014-10-29 14:23:23 i think it is due to the splitted xulrunner build 2014-10-29 14:26:49 31.1.1 2014-10-29 14:27:14 buttons at bottom of new tab page don't work 2014-10-29 14:27:18 and various other glitches too 2014-10-29 14:28:06 in site rendering and functionality in some web apps (e.g. usps checkout page won't accept my payment) 2014-10-29 14:28:31 yes, that version was buggy 2014-10-29 14:28:49 i think ff-33 might work better 2014-10-29 14:28:57 i dropped the separate xulrunner build 2014-10-29 14:29:17 i am trying to do the same in edge/main for v3.1 2014-10-29 14:29:24 <^7heo> I now have sound in firefox. 2014-10-29 14:29:25 <^7heo> :D 2014-10-29 14:29:28 <^7heo> \o/ 2014-10-29 14:29:37 <^7heo> thanks algitbot, but don't die this time. 2014-10-29 14:29:42 <^7heo> \o/ 2014-10-29 14:39:53 Hi all. I'm looking after nbd with alpine (both client and server side). the nbd-client is busybox. it mounts a nbd easyly but I do not find how to take the link down cleanly. kill -9 nbd-client.pid works but it is not what I call clean... 2014-10-29 14:42:54 <^7heo> hey Jean-Scotch 2014-10-29 14:44:01 hi ^7heo 2014-10-29 14:46:41 and it seems that qemu cannot use nbd natively ( qemu-system-x86_64 -drive file=nbd:$IP:$PORT gives errors and do nothing) :( 2014-10-29 14:47:48 <^7heo> http://repos.e-x-a.org/ 2014-10-29 14:47:59 <^7heo> v_v 2014-10-29 14:51:10 k0r10n: hi 2014-10-29 14:51:19 you have an apkbuild for chromium? 2014-10-29 14:51:32 <^7heo> nice way of greeting people. I should try too. :D 2014-10-29 14:52:23 Jean-Scotch: does kill with -1 -2 work? 2014-10-29 14:52:35 * -1 OR -2 2014-10-29 14:53:27 buckley310: no 2014-10-29 14:54:46 <^7heo> okay so SLiM is dependent on dbus because of consolekit. 2014-10-29 14:56:09 ^7heo: Do you want to try xdm@testing and give support if it works? :-) 2014-10-29 14:56:26 that's not dependant on consolekit 2014-10-29 14:56:35 s/support/feedback/ 2014-10-29 14:56:38 <^7heo> I just replaced my login manager. 2014-10-29 14:56:55 <^7heo> echo "pgrep ^dwm$ || startx" >> ~/.profile 2014-10-29 14:57:10 <^7heo> easy, actually. 2014-10-29 14:57:16 ^7heo: 2014-10-29 14:57:22 ::respawn:/bin/su -c "xinit 2>/dev/null >&2" -l 7heo 2014-10-29 14:57:22 <^7heo> I don't understand why people still use window managers in those conditions :D 2014-10-29 14:57:34 that's what I use :-) 2014-10-29 14:57:43 in /etc/inittab 2014-10-29 15:00:32 ^7heo: you have to mention the fun part: ConsoleKit is deprecated 2014-10-29 15:05:03 clandmeter, uselessd-6 :3 2014-10-29 15:07:26 fr0stb1te: done 2014-10-29 15:11:04 <^7heo> darkfader: that's not fun =/ 2014-10-29 15:11:33 <^7heo> darkfader: I mean, I'm glad it's deprecated, but... I'm not glad it has been replaced by pileoshitd 2014-10-29 15:12:01 <^7heo> jomat: problem is, your line is bypassing the login totally. 2014-10-29 15:13:06 ^7heo: Yeah, do you have more users on your pc? 2014-10-29 15:13:42 it seems that the nbd-server to be used with alpine is qemu-nbd. is that right? 2014-10-29 15:13:46 <^7heo> it's a company-owned PC, I can't allow auto-logins. 2014-10-29 15:14:19 ^7heo: because you need to choose users or because of security? 2014-10-29 15:14:30 <^7heo> both 2014-10-29 15:14:48 <^7heo> brb, testing my system 2014-10-29 15:15:27 ok ... I start i3lock from my .xinitrc 2014-10-29 15:16:16 <^7heo> I start dwm from my .xinitrc too 2014-10-29 15:16:23 <^7heo> but I need to startx ONLY after proper login. 2014-10-29 15:16:39 <^7heo> I don't mind if it's not starting at boot tho. 2014-10-29 15:16:45 clandmeter:hi, not yet, i've got temporary Makefile 2014-10-29 15:16:46 <^7heo> since it asks me to login first. 2014-10-29 15:16:49 <^7heo> problem is... 2014-10-29 15:16:50 clandmeter: but working on it 2014-10-29 15:17:42 clandmeter: anyway, i should fix problems with starting 2014-10-29 15:17:43 <^7heo> I want to invoke the startx command and THEN exit the vt. 2014-10-29 15:18:18 clandmeter: u just reconfigure my environment (i had 32 bit env with 64 bit kernel), 2014-10-29 15:18:22 now it's full 64 2014-10-29 15:18:29 so now i can test chrome 2014-10-29 15:18:36 s/u/i 2014-10-29 15:23:37 clandmeter, nice 2014-10-29 15:25:21 <^7heo> okay, I have my login manager. 2014-10-29 15:25:44 <^7heo> echo "start-stop-daemon -b -S xinit 2>/dev/null && clear && exit" >> ~/.profile 2014-10-29 15:26:02 <^7heo> done. 2014-10-29 15:29:53 Jean-Scotch: i dont think busybox can disconnect nbd devices 2014-10-29 15:36:59 i've got an error when try to run:# 2014-10-29 15:36:59 1349 # Fatal error in , line 0 2014-10-29 15:36:59 1350 # Can't override v8::Platform when using default implementation 2014-10-29 15:36:59 1351 # 2014-10-29 15:37:20 so should fix it first.... 2014-10-29 15:47:44 k0r10n: nice, would like to try it when possible. 2014-10-29 15:50:38 ^7heo, exec xinit :) 2014-10-29 15:50:55 <^7heo> fr0stb1te: exec is detaching? 2014-10-29 15:51:04 <^7heo> in my experience, it never did. 2014-10-29 15:51:15 exec xinit & ? :D 2014-10-29 15:51:24 <^7heo> then xinit & is the same. 2014-10-29 15:51:30 Ah. Right. 2014-10-29 15:51:41 <^7heo> but how do you xinit & && echo "paf"? 2014-10-29 15:52:00 I'm actually doing the same as you: starting xinit with a service manager 2014-10-29 15:52:21 <^7heo> I just use start-stop-daemon to be able to have the -b switch. 2014-10-29 15:52:24 <^7heo> the rest is irrelevant. 2014-10-29 15:52:33 <^7heo> I could totally test it differently. 2014-10-29 15:52:40 <^7heo> pgrep xinit for example 2014-10-29 15:53:56 That's ugly. 2014-10-29 15:55:12 <^7heo> what do you think your "nice" programs do in the bg? 2014-10-29 15:55:38 It's much nicer to at least use a pidfile 2014-10-29 15:55:56 Or even just a lockfile 2014-10-29 15:56:05 pgrep is so not the tool for the job 2014-10-29 15:57:19 <^7heo> pgrep works very well, it's not like it could miss an existing process for the current user. 2014-10-29 16:00:11 <^7heo> having a pidfile is just redundant information, and is prone to errors. 2014-10-29 16:00:50 As you wish. 2014-10-29 16:01:11 <^7heo> you're not a native English speaker, are you? 2014-10-29 16:01:37 I am not. 2014-10-29 16:01:40 <^7heo> Alright :) 2014-10-29 16:03:54 <^7heo> I have to say that I like your last blog post, 2014-10-29 16:04:00 <^7heo> https://github.com/landondyer/kasm/blob/master/LICENSE That license is fabulous. 2014-10-29 16:05:19 <^7heo> another comment: you really, really plagiated the suckless theme... One day you might want to find your own identity ;) 2014-10-29 16:07:13 ^7heo: i3lock is a screen locker, dwm a windowmanager 2014-10-29 16:07:44 <^7heo> jomat: i3 is a wm, sorry for mixing it with i3lock. 2014-10-29 16:12:38 Because the suckless.org is so damn original, right? 2014-10-29 16:13:02 <^7heo> I think it has its own identity yes. 2014-10-29 16:13:59 Dude. That layout is literally the example in almost ever css tutorial 2014-10-29 16:14:05 every* 2014-10-29 16:14:16 For good reason — it's simple and practical 2014-10-29 16:17:16 <^7heo> I dunno, I did a quick search on Google for "css tutorial" 2014-10-29 16:17:28 <^7heo> didn't find anything as close as you say to the suckless.org theme. 2014-10-29 16:17:32 <^7heo> http://www.htmldog.com/guides/css/beginner/conclusion/ 2014-10-29 16:17:34 <^7heo> see that 2014-10-29 16:17:45 <^7heo> I really like the theme, but it has nothing to do with the suckless one :) 2014-10-29 16:18:47 Heh. Not css tutorials, div layout tutorials: http://www.w3schools.com/html/html_layout.asp 2014-10-29 16:19:16 http://www.tutorialspoint.com/html/html_layouts.htm 2014-10-29 16:19:18 <^7heo> yeah, the left menu, I totally agree. 2014-10-29 16:19:36 <^7heo> but anyway 2014-10-29 16:19:38 <^7heo> back to work. 2014-10-29 16:20:13 <^7heo> My point was: your blog looks VERY close to suckless. And if I had to choose ANY website on the world that looks like suckless, I'd chose your blog. Nothing else :) 2014-10-29 16:20:40 It's just werid to me that you think of suckless and not, say, early livejournal 2014-10-29 16:20:55 Which is where my colour palette is actually stolen from almost directly 2014-10-29 16:20:58 <^7heo> I've never seen early livejournal :) 2014-10-29 16:21:16 <^7heo> and I've never been too much of a color-sensitive person, I'm a man. 2014-10-29 16:22:09 I could actually make my blog look even more like suckless.org 2014-10-29 16:22:16 With a light header :D 2014-10-29 16:32:19 fr0stb1te: did you try uselessd on alpine? 2014-10-29 16:32:51 Yep. It boots and works as expected for not being configured 2014-10-29 16:33:38 nice 2014-10-29 16:38:40 <^7heo> what is uselessd for, aside replacing runit? 2014-10-29 16:41:48 isn't it for fixing crap at the application level that wants the systemd public apis? 2014-10-29 16:42:27 yes 2014-10-29 16:43:32 um. uselessd doesn't implement all of systemd's API 2014-10-29 16:43:44 You're confusing it with systemd-shim and systembsd 2014-10-29 16:43:54 oh 2014-10-29 16:44:02 uselessd is an init system based on systemd-208 code 2014-10-29 16:44:10 sorry then. i'm happy to have none of any of it 2014-10-29 16:45:52 oh sorry 2014-10-29 16:45:57 i get confused over all this useless bs :) 2014-10-29 16:46:05 <^7heo> yeah, it's a lot of crap for nothing 2014-10-29 16:46:36 fixing this mess is simply a matter of the noyfb principle (none of your fucking business) 2014-10-29 16:47:22 i.e. fake libraries as close as possible to the applications that use these interfaces that expose a fake view of the stuff the application has no business trying to peek at in the first place 2014-10-29 16:47:23 <^7heo> I prefer the kiwf method. 2014-10-29 16:47:40 that doesn't mean just a shim on the dbus 2014-10-29 16:47:53 it means a fake dbus that exposes a fake bus with nothing on it but fake services that do nothing 2014-10-29 16:48:29 <^7heo> http://devopsreactions.tumblr.com/post/101246603377/the-mtu-problem ROFL. 2014-10-29 16:48:33 mm, systemdemu 2014-10-29 16:48:33 ideally at the preprocessor layer so all of the useless code gets optimized out of the apps trying to use it 2014-10-29 16:48:34 dbusemu 2014-10-29 16:48:39 i like the sound of that 2014-10-29 16:49:10 :) 2014-10-29 17:10:15 now i have this problem again ... i need gcc 4.8.3, that's in @edge, and apk just ignores my wish: apk add gcc@edge 2014-10-29 17:12:20 playing with -i -v and -u doesn't change anything 2014-10-29 17:12:48 well, actually -i has one extra line of text: After this operation, 0 KiB of additional disk space will be used. 2014-10-29 17:13:19 what version is in 3.0? 2014-10-29 17:13:26 i thought it was 4.8.3 2014-10-29 17:13:42 or maybe it's 4.8.2 2014-10-29 17:14:27 4.8.2 2014-10-29 17:16:29 btw how are kernel upgrades supposed to work on alpine? it seems that upgrading the kernel package causes the modules for the old kernel to be removed 2014-10-29 17:16:48 meaning you have a broken system until reboot (anything that would load a module that's not already loaded fails) 2014-10-29 17:16:52 is this correct? 2014-10-29 17:18:15 http://sprunge.us/VKXO <- is my apk problem 2014-10-29 17:29:58 hmm 2014-10-29 17:30:12 jomat, does "apk add -u gcc@edge" work? 2014-10-29 17:30:36 if not, what does "apk upgrade --latest" say? 2014-10-29 17:30:56 it would be great to have an "apk cheat sheet" 2014-10-29 17:31:31 for common tasks like "query installed version of a package", "find owner of a file", "list package contents", ... 2014-10-29 17:31:54 i've found apk hard to use in that lots of the common operations like this are buried inside options to apk subcommands 2014-10-29 17:32:11 yeah 2014-10-29 17:32:22 and we had discussion to fix them to a more logical set of commands 2014-10-29 17:32:58 fabled: no 2014-10-29 17:33:23 fabled, that's a big change and certainly would be welcome, but i'd be happy just with a quick reference for finding the commands i need 2014-10-29 17:33:48 dalias: write a bro page for it :-) 2014-10-29 17:34:49 it's mostly due to legacy reasons. 6 years ago, apk-tools was a collection of shell scripts. 2014-10-29 17:34:50 "apk upgrade --latest" now upgrades my firefox: (1/1) Upgrading firefox@testing (33.0-r1 -> 33.0.2-r1) 2014-10-29 17:35:25 but yeah - should write a 'cheat sheet' page 2014-10-29 17:36:03 http://bropages.org/ is a nice project 2014-10-29 17:36:10 for such things :-) 2014-10-29 17:38:17 fabled: http://sprunge.us/TLQX is a transcript of the steps you've asked 2014-10-29 17:39:26 weird. what's in your /etc/apk/world ? 2014-10-29 17:39:42 <^7heo> omg bropages. 2014-10-29 17:39:47 <^7heo> That's so sucky 2014-10-29 17:39:53 fabled: http://sprunge.us/JGAF 2014-10-29 17:41:14 I've never edited that by hand 2014-10-29 17:42:45 sounds like a conflict of installed dependencies or similar. 2014-10-29 17:42:47 hmm 2014-10-29 17:43:08 what does "apk --simulate add gcc@testing>=4.8.3" say ? 2014-10-29 17:43:30 probably, yeah 2014-10-29 17:44:02 "apk --simulate add gcc@testing>=4.8.3"|wc 2014-10-29 17:44:07 0 0 0 2014-10-29 17:44:13 no output at all 2014-10-29 17:44:23 you meant @edge? 2014-10-29 17:44:28 yes 2014-10-29 17:44:30 sorry 2014-10-29 17:44:38 nothing again 2014-10-29 17:44:59 that sounds like a bug 2014-10-29 17:48:25 irclogger logs this channel anyways, but is it ok if I quote the essential lines to the bug report? 2014-10-29 17:51:04 sure 2014-10-29 17:55:13 Jean-Scotch: still stuck on that NBD problem? 2014-10-29 18:10:00 <^7heo> damn I think i have a clone of Lennart at work 2014-10-29 18:30:35 fabled: http://bugs.alpinelinux.org/issues/3493 <- adding g++@edge also upgrades gcc 2014-10-29 18:37:33 Oh wow, now I managed to make apk inoperable by deleting my local repo m) 2014-10-29 18:37:54 resulting that abuild doesn't work anymore, too :-) 2014-10-29 18:44:19 use the --force 2014-10-29 18:54:31 jomat, g++ depends on gcc with specific version 2014-10-29 18:54:34 that kinda explains it 2014-10-29 18:54:44 but it's a bug that you did not get proper error 2014-10-29 19:10:46 fabled: can I also specify a dev dependency like this: gcc>=4.8.3 ? 2014-10-29 19:10:53 my package has a bug with 4.8.2 … 2014-10-29 19:11:00 yes 2014-10-29 19:26:05 <^7heo> does anyone here know why glx-* needs to create a udev device 2014-10-29 19:26:05 <^7heo> ? 2014-10-29 19:43:27 <^7heo> damnit, webgl works only with root rights... 2014-10-29 19:46:15 webgl is a security nightmare i've heard.. 2014-10-29 19:46:44 think: one site uses webgl to read from gpu memory and steal contents that were left there from another site/tab rendering stuff 2014-10-29 19:49:56 <^7heo> yeah 2014-10-29 19:50:19 <^7heo> but it's not helping to make webgl work only with root rights 2014-10-29 19:52:09 <^7heo> hmm 2014-10-29 19:52:59 <^7heo> seems like the error comes from the rights access on /sys/devices/pci0000:00/ 2014-10-29 19:55:36 <^7heo> okay it really seems to be a rights problem. 2014-10-29 20:03:16 <^7heo> I made glxgears work as my normal user (member of group video, and by changing several directories in /sys to root:video and 770) 2014-10-29 20:03:25 <^7heo> but webgl still refuses to work 2014-10-29 20:19:07 is qemu broken in edge? I've got errors on a fresh install complaining about symbol not found related to libusb... 2014-10-29 21:26:23 I just redo the exact same sequence with a 3.0.4 with no errors 2014-10-29 23:17:18 Ho! rsyslog crash with segfault on fresh install... 2014-10-30 07:00:04 Jean-Scotch: i use qemu here? 2014-10-30 08:30:54 ncopa: it went fine when starting again from scratch: burning a usb key, starting it, setup-alpine, apk add qemu then only change repo to edge; apk updateapk upgrade; install openvswitch. and now qemu is working fine. 2014-10-30 08:31:01 thanks for asking 2014-10-30 08:32:13 now, i tried to add screen collectd ssmtp and rsyslog to the mix. everything seems ok so far bur rsyslog who chashes with segfault... 2014-10-30 08:45:06 hm 2014-10-30 08:45:07 ok 2014-10-30 08:45:20 i can add a rsyslog-dbg subpackage 2014-10-30 08:45:27 so you can create a backtrace 2014-10-30 09:00:34 Jean-Scotch: dbg pkg should be ready 2014-10-30 09:01:37 ok. thks. I will give it a try 2014-10-30 09:07:40 morning 2014-10-30 09:16:03 hello all 2014-10-30 09:20:45 good morning 2014-10-30 09:20:52 http://sprunge.us/PhMY shows the output of "rsyslogd -dn" 2014-10-30 09:21:30 http://sprunge.us/PhWY sorry 2014-10-30 09:25:43 and just after this output, it shows "Segmentation fault" 2014-10-30 09:26:37 I tried several variations for the config file to ne extend 2014-10-30 09:26:49 s/ne/no/ 2014-10-30 09:34:51 can we see an strace? 2014-10-30 09:35:11 strace -f rsyslog 2>&1 >/dev/null |sprunge 2014-10-30 09:35:56 http://sprunge.us/EJbj 2014-10-30 09:36:17 arf! nedd ti install strace first ;) 2014-10-30 09:37:09 alpine:~# strace -f rsyslogd -nd 2>&1 >/dev/null | sprunge 2014-10-30 09:37:09 Segmentation fault 2014-10-30 09:37:09 http://sprunge.us/eOMA 2014-10-30 09:53:17 can you seperate the child process traces? 2014-10-30 09:53:30 strace -ff rsyslogd -dn 2014-10-30 09:53:51 then sprunge them 2014-10-30 09:54:07 http://sprunge.us/DeED 2014-10-30 09:54:32 that's the same 2014-10-30 09:54:55 strace -ff rsyslogd -dn -o 2014-10-30 09:55:33 http://sprunge.us/WaBg 2014-10-30 09:55:42 strace -o rsyslog_trace -ff rsyslogd -dn 2014-10-30 09:56:16 http://sprunge.us/UOjQ 2014-10-30 09:56:42 that's the stdout 2014-10-30 09:56:51 arf! sorry 2014-10-30 09:57:04 look at your fs 2014-10-30 09:57:14 it has rsyslog_trace.pid files 2014-10-30 09:57:20 can we see those? 2014-10-30 09:58:01 ok 2014-10-30 09:59:57 it generates 3 files 2014-10-30 10:00:08 alpine:/etc# cat rsyslog_trace.2559 | sprunge 2014-10-30 10:00:08 http://sprunge.us/gSRX 2014-10-30 10:00:08 alpine:/etc# cat rsyslog_trace.2560 | sprunge 2014-10-30 10:00:08 http://sprunge.us/PJRi 2014-10-30 10:00:08 alpine:/etc# cat rsyslog_trace.2561 | sprunge 2014-10-30 10:00:08 http://sprunge.us/KPSD 2014-10-30 10:09:13 i can get it to segfault with -dn but it works fine without 2014-10-30 10:10:34 not here :( "/etc/init.d/rsyslog start" gives ERROR: rsyslog failed to start 2014-10-30 10:13:50 backup your /etc/rsyslog.conf file 2014-10-30 10:14:01 remove the package 2014-10-30 10:14:15 add the package (this will add a clear conf file) 2014-10-30 10:14:28 type rsyslogd 2014-10-30 10:14:33 then ps -ef 2014-10-30 10:14:38 see if it runs 2014-10-30 10:16:08 it is 2014-10-30 10:17:35 but "logger test" does show anything in syslog 2014-10-30 10:20:11 ok, so rsyslogd runs with simple config 2014-10-30 10:20:59 if launched from shell but fron init scripts 2014-10-30 10:21:11 *not from 2014-10-30 10:24:18 are you doing this in a lxc container? 2014-10-30 10:24:41 what does /etc/init.d/rsyslog -d start say? 2014-10-30 10:25:31 no. i'm doing this on bare metal from usb-key 2014-10-30 10:27:08 http://sprunge.us/NfLB 2014-10-30 10:27:40 I've to go to meeting room. I will be back after lunch... 2014-10-30 12:31:48 sorry, meetings. what happens when you run /usr/sbin/rsyslogd -- -c5 -i /var/run/rsyslogd.pid -f /etc/rsyslog.conf 2014-10-30 14:14:32 ScrumpyJack: does not start with those params; sorry but have to go to another meeting... 2014-10-30 18:44:35 Hello all 2014-10-30 18:44:48 I'm looking for a way to monitor Internet usage 2014-10-30 18:44:59 I'd like to be able to break down usage by day/hour/type 2014-10-30 18:45:26 by type I mean "youtube" vs "twitch" vs "gaming" vs "web" vs "other" 2014-10-30 18:46:07 tcpdump 2014-10-30 18:46:10 ! 2014-10-30 18:46:32 yeah... 2014-10-30 18:46:51 you mean "tcpdump + a load of coding and guesswork" 2014-10-30 18:47:03 I was hoping someone had already done the coding and removed the guesswork 2014-10-30 18:52:00 Leolo_3: You could use snort and define your own rules 2014-10-30 18:52:13 again, that's work :-D 2014-10-30 18:53:01 sure, you have to install snort :-) 2014-10-30 18:53:54 and define the rules 2014-10-30 18:54:25 I ask because I went past my bandwidth usage last month 2014-10-30 18:54:30 and I want to know why 2014-10-30 18:54:51 as in "I'm watching to much twitch" or "my neighbour has 'hacked' my wifi" 2014-10-30 18:55:08 I see 2014-10-30 19:00:57 answer seems to be ntop 2014-10-30 19:03:15 service ntop start 2014-10-30 19:03:15 -ash: service: not found 2014-10-30 19:03:24 what is the equiv for alpine? 2014-10-30 19:04:57 rc-service, gotcha 2014-10-30 19:07:53 bleh 2014-10-30 19:08:14 right, so I have alpine linux on my router, booting from CDROM 2014-10-30 19:08:19 saving state to a USB key 2014-10-30 19:15:42 .... and I can't find where it remembers exactly what packages I have installed 2014-10-30 19:17:52 Leolo_3 if you enabled cache then it should be like here written http://wiki.alpinelinux.org/wiki/Local_APK_cache , but to persist beween reboots 2014-10-30 19:18:24 you need to use lbu 2014-10-30 19:18:47 yeah, I did lbu commit 2014-10-30 19:18:57 and that created ja-pinkerton.apkovl.tar.gz 2014-10-30 19:19:16 but I look in there and don't see a file that contains a list of packages 2014-10-30 19:21:24 lbu list|ls http://wiki.alpinelinux.org/wiki/Alpine_local_backup 2014-10-30 19:21:50 sometimes you need to include path 2014-10-30 19:23:12 ah 2014-10-30 19:23:28 etc/apk/world is what I'm looking for 2014-10-30 19:27:53 tbh, alpine package repos need to be renamed to APK World 2014-10-30 19:54:39 Leolo_3 yes, but anything under /etc/ should and will be saved 2014-10-30 20:23:04 Leolo_3: what router hardware have you installed AL on? 2014-10-31 03:58:15 hey all anyone on? 2014-10-31 07:55:51 hi 2014-10-31 08:18:15 hey guy :) 2014-10-31 08:18:21 *guys :) 2014-10-31 08:42:20 hello 2014-10-31 08:54:46 <^7heo> moin moin 2014-10-31 09:31:01 ScrumpyJack: hey 2014-10-31 09:31:07 fuck my network connection right now 2014-10-31 12:06:46 systmkor2: hey. wassup? 2014-10-31 13:40:46 hi all 2014-10-31 13:42:08 if I prepare an usb-key with 2 partitions or if I boot with 2 usb-keys pluged, in both cases the boot process fails with "/sbin/init not found in new root. Launching emergency recovery shell". 2014-10-31 13:43:51 maybe I can fix it changing the "alpine_dev=usbdisk:vfat" parameter in syslinux.cfg? with some UUID? does it seems the right path to follow? 2014-10-31 13:45:16 yes 2014-10-31 13:45:19 I"m looking after that kind of solution because I want to migrate some bare metal at 1000 Km away from my place and my "hand and eyes" is not really competent... 2014-10-31 13:45:53 ok. thks, will look in that direction ;) 2014-10-31 13:53:24 I will migrate debian boxes (with a lot of LVS) to alpine that way. then launch the previous debian in qemu directly from the locals HDD from alpine and then migrate services and data to the new architecture I'm deploying (CEPH cluster + kvm). Do tou think it's worth a report on the wiki? I will also deploy a kind of proxy (kvm accessed locally trough openvswitch)between RBD and NBD for the transitionnal phase when CEPH is not di 2014-10-31 13:54:00 quality documentation is always welcome on the wiki :) 2014-10-31 13:54:38 ok. will do ;) I'm a great fan of wikis for the last 15 years :D 2014-10-31 14:15:18 <^7heo> is nl down? 2014-10-31 14:24:50 /dev/disk/by-uuid doesn't exist in the recovery shell. /dev/usbdisk does but points to sdb1 on my lab netbook. On the target it will not be sdb as there will be more HDD. So I cannot put alpine_dev=sdb1:vfat in syslinux.cfg nor apline_dev=disk/by-uuid/:vfat. If I put alpine_dev=sdb1:vfat, It boots but fails to start services defined in the overlay and /dev/usbdisk points to sdb2, not sdb1... Is this a bug? an (unwanted? 2014-10-31 14:35:22 <^7heo> clandmeter: nl is down, right? 2014-10-31 14:35:40 <^7heo> (and www, and rsync, and all the cnames to it) 2014-10-31 14:37:03 yes it was 2014-10-31 14:37:07 <^7heo> was? 2014-10-31 14:37:08 should be online now 2014-10-31 14:37:09 <^7heo> ok 2014-10-31 14:37:16 yeah http://nl.alpinelinux.org/ is fine 2014-10-31 14:37:29 if you see me timeout, its probably down :) 2014-10-31 14:37:39 although its the first time it happend here 2014-10-31 14:38:03 we use kpn, NL biggest provider. 2014-10-31 14:38:08 even my mobile has issues. 2014-10-31 14:38:51 <^7heo> yeah it's on 2014-10-31 14:49:56 anyone have alpine-xen-3.0.6 work? 2014-10-31 14:50:07 i dont seem to be able to boot any domu 2014-10-31 14:50:14 libxl: error: libxl_dom.c:281:libxl__build_post: Failed to set event channel limit to 1023 (-1) 2014-10-31 14:51:40 <^7heo> I didn't try 2014-10-31 14:51:42 <^7heo> I should. 2014-10-31 14:51:48 <^7heo> but kvm is conflicting with it. 2014-10-31 14:51:59 <^7heo> So I had to remove xen to properly instal kvm 2014-10-31 14:53:12 ^7heo: used not to 2014-10-31 14:53:22 i had kvm as a fallback ~1.5 years ago 2014-10-31 14:54:04 <^7heo> darkfader: it shouldn't, but some files (network bridging if my memory's right) in the kvm packages are directly taken from the xen one 2014-10-31 14:57:01 <^7heo> with the same path ofc. 2014-10-31 16:37:25 what's involved in setting up alpine on arm? 2014-10-31 16:37:57 i have one of these for testing but haven't booted anything on it yet: http://elinux.org/Dragonboard/APQ8074