2014-09-01 07:16:55 <^7heo> moin 2014-09-01 07:17:27 morning 2014-09-01 07:19:22 <^7heo> Raah damn, I'm tired of that nick change on freenode. 2014-09-01 07:19:47 for your account ? 2014-09-01 07:22:14 <^7heo> yeah 2014-09-01 07:22:18 <^7heo> I have the password on some drive 2014-09-01 07:22:25 <^7heo> but I have to dig I don't know where to find it back 2014-09-01 07:22:33 <^7heo> on various level of virtualization 2014-09-01 07:22:40 <^7heo> it's a bit annoying 2014-09-01 07:23:15 :o it's a irc hostname problem ? 2014-09-01 07:23:57 <^7heo> nah, it's that freenode is less stable than the political regime in ukraine... 2014-09-01 07:24:13 <^7heo> I wonder why the Alpine-Linux chan is on freenode and not OFTC... 2014-09-01 07:24:46 oh, ok ^^ 2014-09-01 07:25:55 freenode is often a ddos target in this time :/ 2014-09-01 07:27:10 <^7heo> Goodbox: by "this time" you mean the last 10 years, right? 2014-09-01 07:28:29 Hihi, I can't answer to that because i'm a newbie, i just hope folks who was using clear pwd auth has finally switch to sasl after the last nickserv usurpation :/ 2014-09-01 07:28:47 http://paste.ee/r/xXBpg haha isn't coreans 2014-09-01 07:58:27 <^7heo> Goodbox: yeah 2014-09-01 08:14:05 my cpu has sse sse2 sse4a or 3DNow flags, which should use whith gcc for better fp perfs, is it selected automaticaly with -O2,3 or -fast ? 2014-09-01 08:14:26 +i 2014-09-01 08:15:17 Goodbox: you could check gentoo, they probably have it on their wiki. 2014-09-01 08:15:41 ok thanks 2014-09-01 08:18:19 Goodbox: here is a handy reference: http://web.archive.org/web/2006/http://www.funroll-loops.org/ 2014-09-01 08:23:08 many thanks, i plan for an environnement based on alpine to build musl-cross. That's yet complex things for me... that i hope i'll pass :) 2014-09-01 08:27:13 I'll put two i5 with alpine plus my desk client distcc-pump mode 2014-09-01 08:31:30 and that for trying to build arm hardfp binaries, which i'm really not sure to get :s 2014-09-01 08:35:50 w 20 2014-09-01 08:35:52 oops 2014-09-01 08:36:21 Have you seen the news about systemd's upcoming “new packaging system”? 2014-09-01 08:36:28 yeah... 2014-09-01 08:36:38 <^7heo> openfbtd: irc != vi 2014-09-01 08:37:13 It seems like we will get separate ecosystems in the near future 2014-09-01 08:37:25 There will be systemd linux and other linux 2014-09-01 08:37:39 I didn't like at all lennart's manner to impose caffein to poeples... 2014-09-01 08:38:09 All of its critical posts begin to "Grab caffein" 2014-09-01 08:38:18 are beginning * 2014-09-01 08:38:43 sorry, i didn't learned english at school, but I'm angry 2014-09-01 08:38:44 heh 2014-09-01 08:38:52 <^7heo> GNU/Linux, systemd/Linux and UNIX-based/Linux? 2014-09-01 08:38:59 hangry * 2014-09-01 08:39:03 <^7heo> hungry :D 2014-09-01 08:39:12 ugh facepalm 2014-09-01 08:39:14 :) 2014-09-01 08:39:18 <^7heo> Hungary? 2014-09-01 08:39:24 If this shit gets adopted by major distros, distros like alpine will _have_ to aither join or separate 2014-09-01 08:39:32 <^7heo> either? :P 2014-09-01 08:39:32 And I hope they separate 2014-09-01 08:39:38 either* 2014-09-01 08:39:43 <^7heo> openfbtd: we will fork anyway. 2014-09-01 08:39:51 <^7heo> We already forked: musl is a sign of that. 2014-09-01 08:40:03 CRUX will probably do the same 2014-09-01 08:40:24 And I will be forced to switch from Arch after all :D 2014-09-01 08:41:32 <^7heo> Arch sucks so hard 2014-09-01 08:41:53 It dioes it you see arch as the base arch systen 2014-09-01 08:42:21 If you see arch as its infrastructure, it's one of the easiest distros to tinker with 2014-09-01 08:42:36 so let's get them pkgctl \o/ 2014-09-01 08:42:41 But if they do adopt the new systemd package system, then fuck it. 2014-09-01 08:42:54 It will stop being worth it to support 2014-09-01 08:43:26 openfbtd: i think systemd's goal is to kill distros 2014-09-01 08:43:31 I currently use this: https://fleshless.org/pages/spark.html 2014-09-01 08:43:43 ncopa, yeah, seems like it 2014-09-01 08:43:45 yeap, i'm agree 2014-09-01 08:43:55 basically, in the end, all distros will be the same 2014-09-01 08:43:59 <^7heo> Please drop the 'm between "I" and "agree" :P 2014-09-01 08:44:14 ok ! 2014-09-01 08:44:14 :p 2014-09-01 08:44:17 at that point lennart will likely say: whats the point it making your own distro? 2014-09-01 08:44:27 which of course will be a valid point at that time 2014-09-01 08:44:31 <^7heo> Nah, he will advetise it. 2014-09-01 08:44:41 <^7heo> and try to leverage more power/money 2014-09-01 08:44:48 <^7heo> advertize it even 2014-09-01 08:44:51 And tehn they will do something that makes separating a distro very hard 2014-09-01 08:44:53 to do 2014-09-01 08:44:56 then* 2014-09-01 08:45:39 i kinda expected lennart to try replace package managers 2014-09-01 08:45:48 Like signing your repos through a system that goes through the systemd dev team 2014-09-01 08:45:50 i didnt think it would come so fast though 2014-09-01 08:46:49 The fucked up thing is they are encouraging dll hell 2014-09-01 08:46:51 Basically 2014-09-01 08:46:53 i'm glad we didnt jump on the systemd bandwagon early... 2014-09-01 08:47:09 And then solving it by separating libs 2014-09-01 08:47:20 That's NOT a good dlolution 2014-09-01 08:47:23 solution* 2014-09-01 08:47:34 it's good for commercial/binary-only things 2014-09-01 08:47:37 bad for embedded 2014-09-01 08:47:51 It's more among the lines of “let's create a problem and herocally solve it” 2014-09-01 08:47:54 fabled: they *claim* its good for embedded 2014-09-01 08:48:04 i think they want tv vendors onboard 2014-09-01 08:48:14 heroically*. Which is pretty much all of systemd 2014-09-01 08:48:15 their embedded is likely '1GB RAM' 2014-09-01 08:48:18 or more 2014-09-01 08:48:22 yeah 2014-09-01 08:48:37 The actual embedded system devs have told lennart to go fuck himself 2014-09-01 08:48:43 Like OpenWRT 2014-09-01 08:48:45 the ram/disk-is-cheap argument... 2014-09-01 08:49:02 Maybe systemd guys found of ginving pate troll... All is troll and control in what their product. 2014-09-01 08:49:17 openwrt has some interesting stuff 2014-09-01 08:49:29 im interested in openwrt's ubus and ubox and all 2014-09-01 08:49:33 Also lennrat keeps taling about boot ilems 2014-09-01 08:49:39 http://www.youtube.com/watch?v=nfd5LV72Apk&list=UUPGK3Ffu-Ldr_I6iBuKQsAA this is one of my vms 2014-09-01 08:49:51 Arch with my own service manager and rc 2014-09-01 08:49:55 <^7heo> I kinda expect lennart to replace everything. 2014-09-01 08:50:03 It's actually faster than systemd out of the box on arch 2014-09-01 08:50:50 really nice fact ! 2014-09-01 08:50:52 <^7heo> It's not hard to get faster to a bloat. 2014-09-01 08:51:03 And yeah, I am an actual system engineer 2014-09-01 08:51:16 i.e an advanced system administrator 2014-09-01 08:51:38 And the fact that Lennart keeps telling people I should love his thing... 2014-09-01 08:51:39 No 2014-09-01 08:51:42 I don't. 2014-09-01 08:51:58 I am one of the audiences you speak for. And I hate your guts. 2014-09-01 08:52:47 <^7heo> openfbtd: he is not here, don't talk to him directly :P 2014-09-01 08:52:53 <^7heo> openfbtd: but same story here. 2014-09-01 08:53:10 ^7heo, I'm drunk and angry. So I'm basically just venting 2014-09-01 08:53:23 If I'm annoying, I'll shut up :) 2014-09-01 08:53:29 <^7heo> You're a System engineer, that's to be expected. 2014-09-01 08:53:35 <^7heo> I'm also a sysadmin. 2014-09-01 08:53:38 its monday morning here... isnt it a bit too early for drinking? :) 2014-09-01 08:53:48 I'm just off work 2014-09-01 08:53:49 <^7heo> ncopa: that's why I'm still civil :P 2014-09-01 08:54:08 We have personell problems, so I work night shifts for a bit >_. 2014-09-01 08:54:25 personnel? 2014-09-01 08:56:11 Actually. I'll go sleep it off 2014-09-01 08:56:30 These news made me extremely agitated. 2014-09-01 08:56:34 Not good. 2014-09-01 08:56:49 Does some debian sysadmins will have the right to strike because of hostage in their workplaces ? 2014-09-01 08:56:55 ACTION go out 2014-09-01 09:04:37 <^7heo> I dunno, but I'm worried by what's to come. 2014-09-01 09:04:51 <^7heo> the so called "industry" doesn't have the required maturity 2014-09-01 09:11:05 Anyway. 2014-09-01 09:11:06 I'm off to sleep. 2014-09-01 09:11:20 <^7heo> Sleep well 2014-09-01 09:11:28 <^7heo> Or, rather, sleep wheel. 2014-09-01 09:11:49 gn openfbtd 2014-09-01 09:12:27 that's physic, a world it oxidizes x) 2014-09-01 09:12:54 rude to swimm against the current 2014-09-01 09:48:45 <^7heo> yeah but where the current leads is usually full of crap 2014-09-01 09:54:15 :( 2014-09-01 09:55:07 hopefully alpine-linux sounds like alpines villages losts in the mountains, and that's their culture i want to perpetuate 2014-09-01 10:00:01 i always thought that redhat and canonical (and to some extent google) were just trying to build on open source os x clone 2014-09-01 10:01:11 which is probably a sensible goal but one that doesn't concern me at all, yet for some reason my systems are getting affected by those efforts 2014-09-01 10:01:46 idem, few years ago i commented the introduction to unity and future canonical goals as it on linuxfr.org, everybody tell me i was wrong... :p 2014-09-01 10:02:28 and now i'm retaining to ask them where they are 2014-09-01 10:03:07 canonical choices totally destroyed ubuntu-fr community 2014-09-01 10:15:00 After formatting arch, I stayed in some BSD places for traveling to the "unix origins as a personnal dekstop". In a certain way, for discovering how deep the gap beetwen licences opinions was, and meet folks running them as for food or hobby. 2014-09-01 10:15:48 The probably most suprizing thing for me wasn't the "out of box" vision in BSD flavours 2014-09-01 10:16:21 but the cathedral it was 2014-09-01 10:18:23 everything goes rights, and it let finally less maniability in term of tuning and modifying systems behavours than a linux distro 2014-09-01 10:18:41 where everything can goes somewhere 2014-09-01 10:18:47 go * 2014-09-01 10:22:29 i'm talking to train myself for generating english sentences and share my feelings about unix in 2014. I found that verry gripping and fascinating what theses concepts engendred in the world 2014-09-01 10:29:07 By experiencing as far i could each major bsd branches, there where systematically delicate compromises for my needs 2014-09-01 10:30:52 power management and wifi suprizingly well supported but no sutable smp for doing a bit of multimedia due to the OpenBSD giant kernel lock old mechanism 2014-09-01 10:34:14 at the opposite, excellent desktop reactivity for the NetBSD scheduler but big problem with my sound device (Wait I was thinking that Open was forking Net... why this disgression ?) 2014-09-01 10:39:12 I finally had few freebsd installs, which tell that support loading drm in the kernel for GC drivers, but in the fact you must use the proprietary one. And two weeks later Xorg collapses saying incomprehensible things without any updrade. 2014-09-01 10:41:50 That's why I choose alpine for my personnal desktop. I love the work that is done for build experienced and innovative things in the unix world. 2014-09-01 10:44:13 At my eyes, thats things like that which will do the science involve, as far for security than performances 2014-09-01 10:44:41 not bloated bundles of *ctl things 2014-09-01 10:45:12 that fit to corporates appliances and unusual pecuniary profit 2014-09-01 10:45:27 useless* 2014-09-01 10:48:28 and i'm really sorry for crying my english training, but it need run out :p 2014-09-01 10:53:50 security and* performances 2014-09-01 16:33:31 Anyone have success building nftables on Alpine? 2014-09-01 16:41:57 hum 2014-09-01 16:42:05 anyone having an issue where apk update hangs? 2014-09-01 16:42:29 it's not a lookup error, and I can access it fine from my computer 2014-09-01 16:44:25 nvm it fetched after like 5 minutes o.O 2014-09-01 16:44:44 Mp5shooter: dead route? 2014-09-01 16:44:56 ACTION shrugs 2014-09-03 12:17:26 salut, anyone using zfs on alpine ? 2014-09-03 12:17:36 & linux containers 2014-09-03 13:48:46 <^7heo> adnauseam: not really. But I should. Zfs rocks. 2014-09-03 13:49:20 <^7heo> on the fly compression and encryption is just as badass as a fs can get 2014-09-03 14:07:35 adnauseam: zfs is not yet compatible with musl libc. 2014-09-03 14:08:04 f15RyWpgGTO6 is working on it though. 2014-09-03 14:16:51 encryption? 2014-09-03 14:17:23 yeah, i'm a bit busy with other projects, but the zfsonlinux guys and me are in dialog to fix this. 2014-09-03 14:35:07 f15RyWpgGTO6: do you take paypal'ed pizza/drinks? 2014-09-03 14:39:37 you can flattr me: https://flattr.com/profile/stf 2014-09-03 14:41:46 flattr needs yet another signup... i'll check out tonight 2014-09-03 14:41:53 if you star things on github that's an automatic flattr if you also do flattr 2014-09-03 14:43:46 darkfaded: i know :/ sorry, i hate registrations myself. but flattr is something that is good for saying thanks. i really like to flattr other people. and they seem to be motivated by it. 2014-09-03 14:47:07 the other benefit is, that you can avoid paypal completely if using flattr. my paypal account has been suspended by paypal, because of stupid automatic moneylaundering detection scripts. 2014-09-03 14:47:28 (wasn't moneylaundering, was crowdfunding) 2014-09-03 14:48:08 lol yeah that's too much for their shitty algs 2014-09-03 15:23:16 discourse seems to be intersting, untill you find out it runs on ruby... 2014-09-03 15:24:47 <^7heo> :P 2014-09-03 15:33:49 Why do you only officially support Docker? 2014-09-03 15:33:50 Hosting Rails applications is complicated. 2014-09-03 15:33:59 :) 2014-09-03 15:34:07 nice development choice 2014-09-03 15:37:03 it's complicated let's encapsulate it and then the children will be safe 2014-09-03 15:37:32 (unless you look closely and see the errors are stil there, just in yet another layer of debug) 2014-09-03 16:21:05 I like my new nick. 2014-09-03 16:23:29 http://thedailywtf.com/Articles/The-Fizz-Buzz-from-Outer-Space.aspx 2014-09-04 00:50:10 how do you convert a disk from MBR to GPT on alpine? 2014-09-04 00:54:30 I read you can do it with gdisk but I don't see a gdisk package 2014-09-04 01:31:57 dd if=/dev/null of=/dev/sdx 2014-09-04 01:32:07 and then use fdisk to do shit 2014-09-04 01:45:54 I ended up using parted 2014-09-04 01:46:05 all done now though 2014-09-04 01:46:06 wee 2014-09-04 07:09:52 hoi 2014-09-04 07:10:06 ncopa: i'm trying to set up an alpine desktop 2014-09-04 07:10:31 installed xf86-video-ati forthe radeon device 2014-09-04 07:11:06 and when the gui appears it freezes 2014-09-04 07:11:11 do you have any idea? 2014-09-04 07:13:03 kocka, could be grsec sysfs limit thing. does it work with vanilla kernel? 2014-09-04 07:13:30 or perhaps running "find /sys -type d -exec chmod o+rx {} \;" will help? 2014-09-04 07:14:16 there is nothing in syslog 2014-09-04 07:14:47 and shutting down fine for pressing the power button 2014-09-04 07:29:20 kocka: I dont know really. what alpine version is it? 3.0 or edge? 2014-09-04 07:35:36 3.0.4 2014-09-04 07:35:43 latest stable 2014-09-04 07:36:00 does it happen with vanilla kernel too? 2014-09-04 07:37:08 i pushed firefox 32 to testing 2014-09-04 07:37:21 i will keep firefox 31 in main til next ESR release 2014-09-04 07:37:24 which is 38 2014-09-04 07:38:03 https://www.mozilla.org/en-US/firefox/organizations/faq/ 2014-09-04 07:54:02 havent tested with vanilla yet 2014-09-04 07:57:57 mornin' 2014-09-04 10:03:00 hello ncopa, there's been some talks of official support for busybox in #openrc, maybe you'd want to participate? 2014-09-04 10:04:04 CcxCZ: I thought they already had official support for busybox 2014-09-04 10:04:48 I joined #openrc but are going for lunch now 2014-09-04 16:55:43 re 2014-09-04 16:55:49 yes it happen with vanilla kernel too 2014-09-04 16:56:56 what gfx? 2014-09-04 17:02:12 mm 2014-09-04 17:06:47 how can i set that? 2014-09-05 10:13:44 Hi 2014-09-05 10:14:44 I'm impressed with what AL has to offer, what harm could i do by using edge on a production system ? 2014-09-05 10:40:27 <^7heo> define production 2014-09-05 10:41:40 Jarth, we make no guarantee on the stability of 'edge' rolling release 2014-09-05 10:42:03 while we do not intentionally break it, accidents happen sometimes. 2014-09-05 10:42:08 okay, i'm looking forward to work more ( maybe a lot ) with alpine, it has everything i was looking for 2014-09-05 10:42:24 3.0-stable is recommended for production 2014-09-05 10:42:29 Though i appreciate all efforst, i'd like to build my own 2014-09-05 10:42:43 though as it's the first musl release it is considered 'beta' quality to some degree 2014-09-05 10:42:49 version of hiawatha, elasticsearch, logstash, enlightenment 2014-09-05 10:42:59 this however does not seem to be an easy task at times 2014-09-05 10:43:10 yes. build mastering is tricky at times. 2014-09-05 10:43:16 i'd like to have enlightenment too 2014-09-05 10:43:22 musl does seem to hold great promise 2014-09-05 10:43:28 i have it on ubuntu 2014-09-05 10:43:33 not perfect but works well 2014-09-05 10:43:42 yes, we appreciate musl as well. and especially the co-operation we get from the author(s) 2014-09-05 10:43:48 great 2014-09-05 10:44:04 and we've written some feature / improvements to it as well 2014-09-05 10:44:15 i'm used to working with Debian and Sabayon, getting disgusted again with ubuntu 2014-09-05 10:44:46 would it be hard to part software from .deb to the apk format ? 2014-09-05 10:44:53 part = port 2014-09-05 10:46:25 it's easier to port ebuilds 2014-09-05 10:46:44 the abuild format used for building apk is pretty similar to them 2014-09-05 10:48:32 ah great, thanks 2014-09-05 10:48:37 i saw the mention of overlays 2014-09-05 10:48:56 is it related to abuild being like ebuild or not as much ? 2014-09-05 10:52:24 simpler 2014-09-05 10:52:27 abuild is simpler 2014-09-05 11:06:57 is there a 2.x -> 3.x upgrade guide somewhere for prod. machines? 2014-09-05 11:09:31 <^7heo> http://wiki.alpinelinux.org/wiki/Upgrading_Alpine ? 2014-09-05 11:15:34 Thanks, great chat 2014-09-05 11:25:36 ^7heo: thx 2014-09-05 11:26:55 <^7heo> kocka: anytime. 2014-09-06 10:41:20 Hi ! 2014-09-07 03:49:10 is alpine 2.7 the last non-musl release? 2014-09-07 17:07:24 I upgraded from alpine 2.6 branch to 2.7 2014-09-07 17:07:33 now I can't log into my samba shares and I have this in syslog 2014-09-07 17:07:40 Sep 7 07:54:38 obsidian daemon.err smbd[21338]: [2014/09/07 07:54:38.822331, 0] ../lib/util/become_daemon.c:136(daemon_ready) 2014-09-07 17:08:04 anyone know what's going on? 2014-09-07 17:11:08 just fixed it 2014-09-07 17:11:14 had to do smbpasswd -a for my user again.. 2014-09-08 06:19:27 Mp5shooter: yes, v2.6 has samba-3.6, and v2.7 has samba-4.1 2014-09-08 10:41:20 Is anybody successfully running docker on Alpine? I wanted to play with it a little, but get the following Error when starting the Deamon 2014-09-08 10:41:25 http://pastebin.com/FXDZHv0U 2014-09-08 10:43:43 <^7heo_> I didn't try yet. 2014-09-08 10:51:09 tze: you're supposed to work ;p 2014-09-08 10:51:11 *runs* 2014-09-08 10:57:34 darkfaded: I'm paid to look busy, not to actually be busy! 2014-09-08 10:58:36 danG! 2014-09-08 11:24:40 tze: Could not resist https://i.imgflip.com/bwu7k.jpg 2014-09-08 11:32:16 katlogic: I know, I know, but i want to at least have played with it for a bit 2014-09-08 11:36:28 tze: Ok then. contrib/check-config.sh 2014-09-08 11:42:28 warning: /proc/config.gz does not exist, searching other paths for kernel config... 2014-09-08 11:42:50 Where is the Alpine Kernel Conf? Didn't find it in the usual paths 2014-09-08 11:44:13 tze: You might want to bisect grsec features first. 2014-09-08 11:48:45 tze: here's the ones i turn off for chroot stuff, probably docker will need a few more 2014-09-08 11:48:49 http://wiki.alpinelinux.org/wiki/Talk:Running_glibc_programs 2014-09-08 12:21:05 I have an old v2.0 bluetooth dongle, any chance of getting it detected ? 2014-09-08 12:26:15 ok, found the issue :) 2014-09-09 08:45:10 hello 2014-09-09 08:57:19 morning 2014-09-09 09:03:36 algitbot is a bot 2014-09-09 09:03:37 wtf 2014-09-09 09:03:39 hello 2014-09-09 09:03:41 wat 2014-09-09 09:03:42 hello 2014-09-09 09:03:45 that isn't spammy 2014-09-09 09:03:46 :P 2014-09-09 09:21:07 algitbot: hi 2014-09-09 09:31:00 algitbot: what do you say? 2014-09-09 09:31:23 algitbot: who are you to spam us? 2014-09-09 09:31:36 algitbot: thanks 2014-09-09 11:23:37 <^7heo_> debian sucks. 2014-09-09 11:28:14 lately 2014-09-09 11:28:48 we missed our duty to contribute, and now less competent maintainers took over and subverted it. 2014-09-09 11:29:50 <^7heo_> Okay, lemme rephrase. 2014-09-09 11:29:57 <^7heo_> Debian sucks since Debian 7. 2014-09-09 11:30:05 <^7heo_> (maybe a bit before that) 2014-09-09 11:32:24 same is the situation with firefox btw 2014-09-09 11:33:54 <^7heo_> I agree. 2014-09-09 11:33:58 <^7heo_> Firefox sucks as well. 2014-09-09 11:34:51 <^7heo_> Leaving us with no sane browser. 2014-09-09 11:35:44 two great systems, gentrified. 2014-09-09 11:35:56 <^7heo_> not sure if Debian was great. 2014-09-09 11:36:04 <^7heo_> I had a love/hate relation with it before. 2014-09-09 11:36:09 <^7heo_> Now I have a hate/hate one. 2014-09-09 11:36:22 <^7heo_> But it was usable. 2014-09-09 12:03:10 hi 2014-09-09 12:11:56 <^7heo_> quinq: hi 2014-09-09 12:56:29 <^7heo_> is there the ISC dhcp server on alpine? 2014-09-09 12:58:03 <^7heo_> okay it's simply called dhcp 2014-09-09 12:58:06 <^7heo_> sorry for the noise. 2014-09-09 13:27:57 http://wiki.alpinelinux.org/wiki/Alpine_Linux:Overview mentions uClibc being used. I thought you are building on musl libc? 2014-09-09 13:32:15 gr8: i fixed. thanks 2014-09-09 13:32:33 <^7heo_> Wow, that was a big one. 2014-09-09 13:36:52 why don't you build Alpine statically linked by default? 2014-09-09 13:40:21 <^7heo_> good question. 2014-09-09 13:44:20 because sharing is caring 2014-09-09 13:44:41 shared makes maintenance *alot* easier 2014-09-09 13:45:04 otherwise would we need rebuild much more every time a lib got updated 2014-09-09 13:45:14 it also save runtime memory 2014-09-09 13:45:57 having static /sbin is worth a few gigabytes though (and it doesn't need gigabytes :) 2014-09-09 13:46:18 <^7heo_> ncopa: it doesn't save runtime memory. 2014-09-09 13:46:35 ^7heo: 2014-09-09 13:46:38 i think it does 2014-09-09 13:46:41 <^7heo_> ah 2014-09-09 13:46:42 a bit 2014-09-09 13:46:43 <^7heo_> that way it's better. 2014-09-09 13:46:51 <^7heo_> you think it does. 2014-09-09 13:47:00 <^7heo_> But it's something that has to be determined. 2014-09-09 13:47:01 benchmark it! benchmark it! 2014-09-09 13:47:05 <^7heo_> Because I think it doesn't. 2014-09-09 13:47:08 <^7heo_> yeah 2014-09-09 13:47:10 <^7heo_> we should definitely. 2014-09-09 13:47:16 but yeah, it's *supposed to* at least 2014-09-09 13:47:25 <^7heo_> with the overhead? 2014-09-09 13:47:44 <^7heo_> with the "let's put the whole lib in ram while we use only one function out of it?" 2014-09-09 13:47:57 there are already some experimental linux distributions that are statically linked. DSLR Linux, Morpheus, Stali 2014-09-09 13:48:04 well, it's in ram already usually (because of other apps) 2014-09-09 13:48:09 <^7heo_> gr8: I know only the two last ones... 2014-09-09 13:48:15 but you get the linking phase overhead 2014-09-09 13:48:16 <^7heo_> and sta.li isn't real yet. 2014-09-09 13:48:19 the first one is quite new 2014-09-09 13:48:34 but then, dunno how that would play with ASLR anyway 2014-09-09 13:48:46 <^7heo_> I meant overhead as "stub overhead" 2014-09-09 13:48:50 <^7heo_> not as loading overhead. 2014-09-09 13:48:53 there have been statically linked things 20 years ago, and there was quite a few reasons to go away, i bet we just don't know them any more 2014-09-09 13:49:14 <^7heo_> darkfaded: the reasons at the moment was to save memory 2014-09-09 13:49:21 <^7heo_> we don't have that problem anymore. 2014-09-09 13:49:35 that's a desktopy issue anyway 2014-09-09 13:49:54 <^7heo_> at the time it wasn't 2014-09-09 13:50:04 i dont want rebuild every package linked to libfoo when libfoo is updated 2014-09-09 13:50:05 <^7heo_> it was a general issue. 2014-09-09 13:50:11 <^7heo_> ncopa: that's a valid reason, ofc. 2014-09-09 13:50:17 <^7heo_> ncopa: and I perfectly hear you on that, 2014-09-09 13:50:24 <^7heo_> ncopa: you are right. 2014-09-09 13:50:33 <^7heo_> you shouldn't have to compile that, ofc. 2014-09-09 13:50:49 <^7heo_> so as of now, dyn linking IS necessary, but for that reason and that reason only. 2014-09-09 13:51:04 <^7heo_> btw, links or lynx? 2014-09-09 13:51:29 <^7heo_> I just need to configure a stupid cisco device, and X doesn't start... 2014-09-09 13:51:49 elinks maybe 2014-09-09 13:52:04 <^7heo_> k 2014-09-09 13:52:05 <^7heo_> thanks 2014-09-09 13:55:39 <^7heo_> okay 2014-09-09 13:55:45 w3m and elinks are only ones that I know of that have tabs 2014-09-09 13:55:48 ^7heo_: i think for anything like user applications, dynamic linking is nice since you can replace old libs via mapping in ld. so you don't need to recompile... also LD_PRELOAD is another thing in favor of dynamic linking. for emergency maintenace or HA clusters I *LOVE* static linked stuff 2014-09-09 13:55:48 <^7heo_> elinks isn't able to display the cisco pages properly. 2014-09-09 13:56:03 <^7heo_> on the other hand, cisco pages are quite ugly 2014-09-09 13:56:33 <^7heo_> darkfaded: compiling shouldn't be a problem neither for the users nor for the maintainers... 2014-09-09 13:56:54 <^7heo_> other than that, if you do use proper compression in updates, you shouldn't download the same binary code twice. 2014-09-09 13:57:05 <^7heo_> I mean in the package manager update. 2014-09-09 13:57:06 that's a lot of theory 2014-09-09 13:57:23 besides chrome there's no single project that ever applied it to reality 2014-09-09 13:57:25 <^7heo_> So in the end, the ONLY thing that changes it that you unfold many times the same function in various programs when you update. 2014-09-09 13:57:37 <^7heo_> that's a LOT of theory, you are totally right. 2014-09-09 13:57:49 <^7heo_> but there's no proof that this isn't 10000x better than what we have. 2014-09-09 13:57:54 ack :) 2014-09-09 13:58:24 <^7heo_> and the fact that it doesn't exist isn't a valid reason to not make it exist :P 2014-09-09 13:58:32 hehe 2014-09-09 13:58:37 <^7heo_> that said, I have to work, 'cause that cisco switch is a pain in my lower back. 2014-09-09 13:58:58 <^7heo_> and CcxCZ: w3m isn't available on alpine, afaik 2014-09-09 13:59:44 <^7heo_> ncopa: is it expectable that I can't start X in qemu/kvm ? 2014-09-09 14:00:47 ^7heo_: i would expect X to work in qemu/kvm 2014-09-09 14:01:09 <^7heo_> ncopa: straight away, or after some config? 2014-09-09 14:01:36 it *should* work stright away, if you have the proper xorg video driver 2014-09-09 14:01:43 but i would not be surprised if it doesnt 2014-09-09 14:01:53 <^7heo_> hmm 2014-09-09 14:02:00 <^7heo_> I have installed cirrus ofc... 2014-09-09 14:02:03 might also depend on what video chip you emulate 2014-09-09 14:02:07 <^7heo_> and startx fails 2014-09-09 14:02:20 <^7heo_> X also, same error: no screens found. 2014-09-09 14:02:37 Xorg.0.log? 2014-09-09 14:02:47 <^7heo_> moment 2014-09-09 14:02:52 is e v3.0 or edge? 2014-09-09 14:03:00 <^7heo_> oh 2014-09-09 14:03:08 <^7heo_> failed to load module "vesa" 2014-09-09 14:03:14 <^7heo_> brb, slapping myself with a large trout. 2014-09-09 14:03:49 nah, its more cumbersome to get Xorg running that it should be 2014-09-09 14:04:00 been thinking of a setup-desktop or similar script 2014-09-09 14:04:04 that makes all the magic 2014-09-09 14:04:33 <^7heo_> magic != good 2014-09-09 14:05:00 <^7heo_> alright fbdev missing now. 2014-09-09 14:05:25 <^7heo_> works. 2014-09-09 14:05:26 <^7heo_> perfect. 2014-09-09 14:05:31 cool :) 2014-09-09 14:05:36 <^7heo_> now I only need to git clone dwm and stuff. 2014-09-09 14:05:37 <^7heo_> thanks a lot. 2014-09-09 14:05:51 <^7heo_> I forgot about Xorg log and messages wasn't helpful. 2014-09-09 14:07:47 <^7heo_> another problem: can't shut X down, there's no such thing as ^alt-F1 in qemu... 2014-09-09 14:07:55 <^7heo_> and "sendkey ctrl-alt-f1" doesn't work 2014-09-09 14:07:57 <^7heo_> damnit. 2014-09-09 14:10:12 <^7heo_> also, short feedback on apk 2014-09-09 14:10:35 <^7heo_> when you apk add stuff 2014-09-09 14:10:49 <^7heo_> and there's only one archive downloaded 2014-09-09 14:10:57 <^7heo_> it happens in less than a second here, so no time to see anything 2014-09-09 14:11:11 <^7heo_> and then I get the report for the full dependency list, in one size-report 2014-09-09 14:11:13 <^7heo_> like 2014-09-09 14:11:20 <^7heo_> downloaded 250MB 2014-09-09 14:11:24 <^7heo_> and I know I can't downloat THAT fast 2014-09-09 14:12:11 <^7heo_> Live example: 2014-09-09 14:12:18 <^7heo_> # apk add binutils 2014-09-09 14:12:31 <^7heo_> (1/2) Installing binutils-libs (2.24-r2) 2014-09-09 14:12:36 <^7heo_> (2/2) Installing binutils (2.24-r2) 2014-09-09 14:12:51 <^7heo_> Executing busybox-1.22.1-r9.trigger 2014-09-09 14:12:58 <^7heo_> OK: 273 MiB in 101 packages 2014-09-09 14:13:00 <^7heo_> # 2014-09-09 14:13:15 <^7heo_> that's a bit misleading the first times :P 2014-09-09 14:13:57 OK: 273 MiB in 101 packages 2014-09-09 14:14:03 is stats of all installed packages 2014-09-09 14:14:16 <^7heo_> like, the WHOLE system? 2014-09-09 14:14:28 yes 2014-09-09 14:14:45 <^7heo_> alright, that explains why each package I install has a list of dependencies increased by one... 2014-09-09 14:14:49 <^7heo_> :P 2014-09-09 14:14:57 <^7heo_> say, where is stdio.h? 2014-09-09 14:15:17 apk info --who-owns /usr/include/stdio.h 2014-09-09 14:17:06 <^7heo_> awesome, danke. 2014-09-09 14:17:28 <^7heo_> ofc when the package isn't installed... 2014-09-09 14:17:30 <^7heo_> it's not working 2014-09-09 14:17:37 true 2014-09-09 14:17:48 to keep the index small 2014-09-09 14:18:01 <^7heo_> nah but I'm not criticizing that 2014-09-09 14:18:13 <^7heo_> we should have an only search engine for that, however. 2014-09-09 14:18:19 yes 2014-09-09 14:18:24 <^7heo_> if possible with a VERY simple interface, like 2014-09-09 14:18:29 <^7heo_> something you can curl 2014-09-09 14:18:34 hm 2014-09-09 14:18:39 <^7heo_> yeah like 2014-09-09 14:18:56 <^7heo_> http://packages.a.o/search/pattern 2014-09-09 14:19:16 <^7heo_> and we send back json or yaml formatted data. 2014-09-09 14:19:19 <^7heo_> what do you think? 2014-09-09 14:19:21 yeah 2014-09-09 14:19:36 <^7heo_> in what should I write it? 2014-09-09 14:19:38 <^7heo_> php? 2014-09-09 14:19:39 <^7heo_> python? 2014-09-09 14:19:41 <^7heo_> C? 2014-09-09 14:19:51 Lua? 2014-09-09 14:20:01 php should be ok too 2014-09-09 14:20:08 i have some idea of a hook 2014-09-09 14:20:10 <^7heo_> I'm much more fluent in php than lua but... 2014-09-09 14:20:17 <^7heo_> I have nothing against learning a sane thing 2014-09-09 14:20:24 that when package gets uploaded to mirror 2014-09-09 14:20:38 it will generate a package-1.0.filelist file 2014-09-09 14:20:49 could be done with a simple Makefile 2014-09-09 14:21:04 <^7heo_> btw, do you have any idea where I can find the stdio.h header? 2014-09-09 14:21:08 <^7heo_> 'cause it's a bit annoying to not have it... 2014-09-09 14:21:10 i have a local mirror for that task where i'm doing some tar tzf ... > files.txt stuff 2014-09-09 14:21:23 <^7heo_> that'd be great. 2014-09-09 14:21:29 %.filelist: %.apk 2014-09-09 14:21:41 <^7heo_> okay 2014-09-09 14:21:42 tar -ztf $< > $@ 2014-09-09 14:21:43 http://irclogger.com/.alpine-linux/2014-08-25#1408993322 2014-09-09 14:22:07 find . -name \*.apk -exec sh -c 'i=`basename $0`;tar -tzf $0|sed s/^/$i:\\//' {} \; >files.txt 2014-09-09 14:22:46 <^7heo_> jomat: yeah that works but... 2014-09-09 14:22:59 <^7heo_> it's only for the local stuff, again. 2014-09-09 14:23:06 <^7heo_> so --who-owns works the same for that. 2014-09-09 14:23:20 but it generates a file list for non-installed 2014-09-09 14:23:24 musl-dev 2014-09-09 14:23:31 ^7heo_: no 2014-09-09 14:23:32 is what has the libc header 2014-09-09 14:23:43 <^7heo_> thanks 2014-09-09 14:23:48 ^7heo_: you probably want apk add build-base 2014-09-09 14:23:52 <^7heo_> yeah 2014-09-09 14:23:59 <^7heo_> many thanks 2014-09-09 14:24:05 <^7heo_> I wasn't aware of the proper names 2014-09-09 14:24:12 <^7heo_> makes total sense when you think of it 2014-09-09 14:24:41 ^7heo_: i run it on those files: http://dl-3.alpinelinux.org/alpine/v3.0/main/x86_64/ 2014-09-09 14:25:08 ^7heo_: and wget -r them 2014-09-09 14:25:58 <^7heo_> is there something such as apk autoremove ? 2014-09-09 14:26:17 <^7heo_> 'cause I just typed "apk add xorg-server-dev" instead of "libx11-dev" 2014-09-09 14:26:22 <^7heo_> and now I regret it =/ 2014-09-09 14:26:58 type: apk del xorg-server-dev 2014-09-09 14:27:00 :-D 2014-09-09 14:27:20 <^7heo_> it will remove all the deps? 2014-09-09 14:28:42 <^7heo_> also, any idea where to find tic? 2014-09-09 14:29:22 ^7heo_: ncurses-5.9-r1.apk:/usr/bin/tic 2014-09-09 14:29:37 <^7heo_> thanks 2014-09-09 14:33:33 jomat: http://sprunge.us/ThAi 2014-09-09 14:33:42 Makefile to generate file lists 2014-09-09 14:33:49 it does not clean old though 2014-09-09 14:34:25 nice! 2014-09-09 14:36:00 tar -ztf $< > $@.tmp & mv $@.tmp $@ 2014-09-09 14:36:05 might be better 2014-09-09 14:36:09 tar -ztf $< > $@.tmp && mv $@.tmp $@ 2014-09-09 14:36:17 create .tmp firsst 2014-09-09 14:36:38 otherwise it might create corrupt files if you do ctrl-c or if it fails for other reason 2014-09-09 15:03:32 <^7heo_> I keep getting "BAD signature" in xf86-video-modesetting 2014-09-09 15:05:04 <^7heo_> okay one apk del and rm later, it's okay. 2014-09-09 15:09:57 <^7heo_> I wonder why my input isn't working in X... I mean, afterall, I'm having no problem with the time (it changes) 2014-09-09 15:25:08 <^7heo_> damnit, X freezes all the time. 2014-09-09 15:25:16 <^7heo_> And I'm sure it's just the input... 2014-09-09 15:25:43 i had to install udev to get input to work w/ X 2014-09-09 15:25:52 <^7heo_> hmm 2014-09-09 15:25:57 <^7heo_> thanks for the feedback, trying that. 2014-09-09 15:25:59 i used the setup-xorg-base script 2014-09-09 15:26:36 <^7heo_> hmm 2014-09-09 15:26:38 i would be interested in getting xorg to work w/ mdev though 2014-09-09 15:26:46 i just dont have time to look into it at the moment 2014-09-09 15:26:50 <^7heo_> I would be interested in getting X work here. 2014-09-09 15:28:56 <^7heo_> with udev, I get "Exec format error"... 2014-09-09 15:28:59 <^7heo_> very strange. 2014-09-09 15:34:36 I'm using Alpine & X currently as main system on my notebook and have those packages installed for x input: xf86-input-mouse xf86-input-keyboard xf86-input-evdev 2014-09-09 15:34:59 <^7heo_> jomat: I also am using that at home on my laptop 2014-09-09 15:35:02 <^7heo_> and it works mighty fine. 2014-09-09 15:35:08 <^7heo_> but here, with qemu, I have no input... 2014-09-09 15:35:10 <^7heo_> sucks. 2014-09-09 15:59:24 <^7heo_> works now. 2014-09-09 15:59:28 <^7heo_> I did resetup everything once. 2014-09-09 15:59:30 <^7heo_> works 2014-09-09 15:59:36 <^7heo_> I'm doing to delete the old "drive" 2014-09-09 15:59:40 <^7heo_> I must have done shit. 2014-09-09 16:00:50 <^7heo_> on the other side, using the setup script must have been what saved me. 2014-09-09 16:48:58 why the ada-no-pie patch? 2014-09-10 09:08:38 <^7heo_> moin 2014-09-10 09:08:45 <^7heo_> how does one list all the installed packages? 2014-09-10 09:15:43 apk info 2014-09-10 09:16:02 or cat /etc/apk/workd 2014-09-10 09:16:08 world* 2014-09-10 09:18:14 <^7heo_> thanks 2014-09-10 09:28:02 <^7heo_> tired of debian 2014-09-10 09:28:27 <^7heo_> why do I still have that tftp-hpa script in /etc/init.d/ after I "sudo apt-get remove --purge tftp-hpa"?? 2014-09-10 09:28:30 <^7heo_> v_v 2014-09-10 09:29:45 <^7heo_> and now apt complains 2014-09-10 09:29:49 <^7heo_> broken system... 2014-09-10 09:30:46 <^7heo_> dpkg: error processing tftpd-hpa (--remove): subprocess installed post-removal script returned error exit status 1 2014-09-10 09:30:50 <^7heo_> So awesome. 2014-09-10 17:17:28 a lot of software i use is in "unmaintained"! 2014-09-10 17:19:14 what happened to those? 2014-09-10 17:19:33 pulseaudio? 2014-09-10 17:19:50 ok kde is big and maybe a bitch to build 2014-09-10 17:20:08 anyway... 2014-09-10 17:22:02 idk there is a process to take over maintainership 2014-09-10 17:22:31 for kde it would be good to have a group that can tend to it i guess? 2014-09-10 17:22:43 (i just have 3 pesky packages i know anything about) 2014-09-10 17:26:57 <^7heo_> Anaphaxet0n: if you use shitty software, that's the reason why it's unmaintained. 2014-09-10 17:27:06 <^7heo_> I mean, pulseaudio? 2014-09-10 17:27:19 :) 2014-09-10 17:27:35 <^7heo_> You know how I fix every single audio problem in Linux? 2014-09-10 17:27:41 hehe 2014-09-10 17:27:44 my guess: uninstall pulse 2014-09-10 17:27:47 <^7heo_> I remove pulseaudio. 2014-09-10 17:27:49 <^7heo_> yeah] 2014-09-10 17:27:50 :) 2014-09-10 17:27:53 <^7heo_> All freaking time! 2014-09-10 17:27:54 ok 2014-09-10 17:28:11 <^7heo_> and about KDE, it's not THAT shitty on the UI, but technically... 2014-09-10 17:28:21 <^7heo_> and it's soooo bloated. 2014-09-10 17:28:31 <^7heo_> I bet it's even bigger than the default windows UI 2014-09-10 17:28:38 that doesn't say a lot 2014-09-10 17:28:56 other popular desktop-environment crap is also bigger than windows 2014-09-10 17:28:57 <^7heo_> right, and there's no way to know 2014-09-10 17:29:11 <^7heo_> 'cause we can't know how big the windows UI source is. 2014-09-10 17:29:26 <^7heo_> I dunno, lxde isn't that big 2014-09-10 17:29:31 <^7heo_> and it's rather popular. 2014-09-10 17:29:39 pulseaudio should have been the authors last project, after that we should have banned him from all commit access. instead? we're forced to emmigrate because he's given more powers, not less. 2014-09-10 17:30:39 i mean avahi should've been a warning sign already. 2014-09-10 17:34:57 <^7heo_> yeah 2014-09-10 17:35:23 <^7heo_> and now systemd 2014-09-10 17:35:48 <^7heo_> https://upload.wikimedia.org/wikipedia/commons/1/18/Avahi-Architecture-Overview.svg 2014-09-10 17:35:54 <^7heo_> v_v 2014-09-10 18:56:47 so the systemd man is the avahi and puple man 2014-09-10 18:57:00 and what is so wrong about avahi and systemd? 2014-09-10 18:57:22 s/puple/pulse 2014-09-10 19:49:12 Anaphaxet0n: avahi is that thing that happily tears down embedded systems based on centos/rhel 2014-09-10 19:49:37 like air traffic etc., integrating a distro is about 20% checking avahi does never ever get into play 2014-09-10 19:49:48 so people don't die from this desktop shit 2014-09-10 19:53:37 Anaphaxet0n: i worked on 3 systems like that. and the most dangerous components of the systems today are avahi, networkmanager and in the future also systemd 2014-09-10 19:54:18 like, you already have redundant networks into every host. network fails. and some bullshit for laptop users tears down the network config but doesn't bring it back up again because it's bad code anyway 2014-09-10 19:54:47 now put that whole system on a unmanned lighthouse in the ocean 2014-09-10 19:57:00 I've hated network manager ever since it mysteriously broke my wpasupplicant setup 2014-09-10 19:57:26 wpasupplicant simply couldn't connect, because network manager was running in the background 2014-09-10 20:01:02 if you do crypto you'll also start hating the gnome-keyring-manager ;) 2014-09-10 20:24:05 they use centos/rhel for air traffic???!! 2014-09-10 20:24:45 they use linux in unmanned stuff?! 2014-09-10 20:29:11 Anaphaxet0n: go on lkml and ask people if they think linux is the right thing. they'll all say yes. they'll just react like you if some manager decides to really do it :) 2014-09-10 20:29:17 but i reacted just the same 2014-09-10 20:29:33 "this is running for 10 years... and you use *linux*?" 2014-09-10 20:30:03 i'd have guessed qnx or something 2014-09-10 20:30:17 or at least some hardcore linux like carrier grade linux 2014-09-10 20:30:18 but no. 2014-09-10 20:30:29 commercial off the shelf 2014-09-10 20:30:36 dev friendly > sanity 2014-09-10 20:30:49 I wouldn't trust linux my life 2014-09-10 20:31:14 if you have any modern car... you do :) 2014-09-10 20:31:55 and it's all workable if you let a few people iron out the bugs. just it's a classic example for why people hate avahi/networkmanager/systemd 2014-09-10 20:32:00 nah airbags dont use logic 2014-09-10 20:32:16 this stuff is written for laptops of lkml smartasses in a cafe 2014-09-10 20:32:19 not for servers 2014-09-10 20:43:57 and who talked about servers? 2014-09-10 20:44:14 smartasses of all kinds need a decent desktop 2014-09-11 05:17:01 morning 2014-09-11 06:08:27 morning 2014-09-11 06:08:45 there was someone who reported that Xorg hangs in qemu with curry gfx 2014-09-11 06:08:53 did anyone find out why it happens? 2014-09-11 09:20:29 <^7heo_> moin @ 2014-09-12 11:21:10 Hi, i'm trying to build http://vice-emu.sourceforge.net/ on alpine, inspired from it's ebuild and pkgbuild, my one start compiling but fails, here's info i pasted: https://bpaste.net/raw/951943fec390 . So, i'm a little bit lost :/ 2014-09-12 11:22:42 Goodbox: it is due to our hardened toolchain 2014-09-12 11:22:46 you cannot use ebx register 2014-09-12 11:24:46 regbx ? hmm ok 2014-09-12 11:25:12 we have more cpuid patches 2014-09-12 11:25:32 ok, i'll look to add another define 2014-09-12 11:25:37 http://git.alpinelinux.org/cgit/aports/tree/main/cpufrequtils/0001-Support-for-Position-Independet-Code-by-aboid-clobbe.patch 2014-09-12 11:28:02 Thanks, i'm gonna read the pic fix guide and try to add a define. One last question, my procedure is ok ? 2014-09-12 11:28:22 this should work: 2014-09-12 11:28:24 #define cpuid(func, ax, bx, cx, dx) \ 2014-09-12 11:28:24 __asm__ __volatile__ ("movl %%ebx, %%esi; cpuid; xchg %%ebx, %%esi;": \ 2014-09-12 11:28:24 "=a" (ax), "=S" (bx), "=c" (cx), "=d" (dx) : " 2014-09-12 11:29:12 ok ncopa, i'll try in few minutes ! 2014-09-12 11:29:22 the cpuid instruction clobbers ebx registry 2014-09-12 11:30:14 so we save original ebx value in esi 2014-09-12 11:30:21 then call cpuid 2014-09-12 11:30:45 swap ebx <-> esi, so original ebx is restored 2014-09-12 11:31:02 and tell gcc that we return value from esi instead of ebx 2014-09-12 11:31:19 hmm that's begining to takes sence to me, then I can remove -fPIC ? 2014-09-12 11:31:28 no 2014-09-12 11:31:34 well 2014-09-12 11:31:45 i think we add -fPIC by default 2014-09-12 11:32:35 the rest of the c code needs be compiled with -fPIC, but it might be you dont need manually add the flag 2014-09-12 11:33:36 ah ok, i'm gonna try 2014-09-12 11:34:17 i'll tell you whats going on, but I first have to eat something 2014-09-12 11:34:21 many thanks 2014-09-12 11:34:27 food helps :) 2014-09-12 11:34:29 np 2014-09-12 11:48:55 <^7heo_> brain needs food. 2014-09-12 11:48:57 <^7heo_> and you need brain. 2014-09-12 11:49:01 <^7heo_> so... 2014-09-12 11:49:27 Dain bramaged. 2014-09-12 11:59:26 Ok, then we're attempting to hide or avoid using the PIC register 2014-09-12 12:01:05 i carefully patched as suggested, now that's parser which cry, i take a look 2014-09-12 12:02:56 yeah i see 2014-09-12 12:03:28 does we need the last : ? 2014-09-12 12:08:42 all right ncopa, tried ("movl %%ebx, %%esi; cpuid; xchg %%ebx, %%esi;" : "=a" (ax), "=S" (bx), "=c" (cx), "=d" (dx)) seem ok :) 2014-09-12 12:17:31 fixed last things, and all right it passed successfully, again thank you ncopa 2014-09-12 12:17:39 nice afternoon all 2014-09-12 12:49:31 Goodbox: feel free to send a patch with the apkbuild to the alpine-devel mailing list ;) 2014-09-12 13:29:22 ncopa: with pleasure, i'll share all my packages, some of them need some work (girara and mumble's qmake whitout qt-designer...). Then my repo need even more maturity, i'll share when it is clean :) 2014-09-14 15:48:24 hi, i have an issue with xorg-1.16 and xf86-video-ati (alpine linux edge x86-64), i got an unresolved symbol error in /var/log/Xorg.0.log too: http://sprunge.us/IIDL 2014-09-14 15:49:01 i can start X with fbdev, but it hangs when i launch st (a terminal emulator) 2014-09-14 16:36:24 Hello 2014-09-14 16:36:44 Evil_Bob: does the kernel module is loaded ? 2014-09-14 16:37:06 yea the "radeon" kernel module is loaded 2014-09-14 16:38:30 hmm well, sorry I never tried with an ati card, maybe the Xorg configuration which can't use radeon 2014-09-14 16:41:00 the link to my Xorg.0.log i posted has an error which seems to indicate a problem with glamor-egl, but im not sure whats wrong 2014-09-14 16:42:08 i ran "apk fix" on xf86-video-ati, xorg-server etc too just in case 2014-09-14 16:46:39 effectively, a shared object seem miss, what about libdrm version installed ? 2014-09-14 16:47:44 yea libdrm is installed too 2014-09-14 16:47:58 ok strange... 2.4.56-r0 ? 2014-09-14 16:49:02 yea exact version 2014-09-14 16:49:24 well, let my take a try with alpine edge on a ati laptop 2014-09-14 16:49:33 5 mins 2014-09-14 16:49:42 awesome, appreciate it man :) 2014-09-14 17:00:08 yw! Then I installed xfce4 to test, this is a HD 5650 Mobility; loaded relatives modules are: radeon, drm_kms_helper and ttm. I just have a /etc/X11/20-modules.conf with some modules, you should have the same: dri, dri2, glx plus useless defaults. The X11 log tell about AIGLX error, missing /usr/lib/xorg/modules/dri/swarst_dri.so however the server and xfce starts well :/ 2014-09-14 17:02:26 i have a radeon 5770 card, i have mesa-dri-ati installed 2014-09-14 17:04:46 Okay, hmm as we see, your AIGLX seem works, but thinking of corrupted glamor_egl binary appears really improbable 2014-09-14 17:04:48 hmm i dont have drm_kms_helper or ttm loaded, i didnt before with xorg-1.15 though and it worked fine 2014-09-14 17:07:12 possibly a symbol conflict, i saw glamor-egl pkg depends on wayland which provide a libdrm.so too 2014-09-14 17:09:21 If you have ideas where we can investigate between our apline edge ati box 2014-09-14 17:09:52 I guess your packages are the lasts 2014-09-14 17:12:37 This message talk about GLX and glamor name space conflicts http://lists.freedesktop.org/archives/glamor/2013-January/000292.html 2014-09-14 17:46:22 this recent commit added the dependency glamor-egl for xf86-video-ati http://git.alpinelinux.org/cgit/aports/commit/?id=1c6ba8ca3d5e6419437cf6ea8f76306a8df6ca58 is it possible its a bug? 2014-09-14 17:47:27 --enable-dri and --enable-kms is also removed 2014-09-14 18:00:00 Yes, actually on my box, xf86-video-ati depends on these things: https://bpaste.net/show/071081aa3601 2014-09-14 18:11:22 i will try to rebuild xf86-video-ati with aports and bisect the problem 2014-09-14 18:37:48 ok i have a working x server again, i added --enable-dri, --enable-kms and --disable-glamor: http://www.codemadness.nl/downloads/patches/alpine/0001-xf86-video-ati-disable-glamor-and-enable-dri-and-kms.patch 2014-09-14 18:38:40 Goodbox: thanks for helping out! 2014-09-14 18:40:29 Hope rebuilding helps, may the force be with you :P 2014-09-14 18:42:04 yea it helped :) 2014-09-14 18:42:25 yay nice, it works ? 2014-09-14 19:37:33 question, is apk going to support ssl in the feature? 2014-09-14 19:41:04 Note: Currently there are no public https or ftp repositories. The protocols are available for local repositories. 2014-09-14 19:41:08 woops 2014-09-14 22:47:53 I'm trying to compile something in Alpine, and getting http://sprunge.us/TddF 2014-09-14 22:47:58 am I missing a package? 2014-09-14 22:49:55 ah I was missing musl-dev 2014-09-15 05:53:44 morning ! 2014-09-15 06:32:24 morning 2014-09-15 06:32:42 Evil_Bob: is http://www.codemadness.nl/downloads/patches/alpine/0001-xf86-video-ati-disable-glamor-and-enable-dri-and-kms.patch needed to fix xorg with ati chips? 2014-09-15 06:47:37 Sorry, I got a dumb question... I'm packaging a software and its tag is "0.9.8-rc3", should-i use this string for pkgver ? 2014-09-15 06:54:18 well, for the moment I took the 0.9.7 tag 2014-09-15 06:56:48 time to launch for me, bon appêtit 2014-09-15 06:57:33 i think (can't check at the moment) there's one for the upstream software release (0.9.8-rc3) and one for *your* package version 2014-09-15 07:22:02 darkfader: okay, thanks 2014-09-15 07:37:07 ncopa: i only tested it on my radeon 5770, but yea it works very well now again 2014-09-15 07:37:37 Evil_Bob: ok, I'll apply it then i think 2014-09-15 07:37:44 thanks! 2014-09-15 07:37:56 ty too :) 2014-09-15 07:38:57 Evil_Bob: feel free to hang around in #alpine-devel 2014-09-15 07:57:59 re 2014-09-15 07:58:01 hmm 2014-09-15 07:58:06 what is this about? 2014-09-15 07:58:14 i mean the patch 2014-09-15 07:58:55 what was the issue? 2014-09-15 08:08:39 i guess i had the same issue :( 2014-09-15 08:08:54 thats why xfce freezed 2014-09-15 08:19:21 kocka: symbol error with glamor-egl + xf86-video-ati, if you can test the patch on your machine too and report if it fixes your problem aswell that would be great :) 2014-09-15 08:31:05 oumm 2014-09-15 08:31:20 ive installed another system since then 2014-09-15 08:31:40 should create a livecd based on xfce :/ 2014-09-15 13:06:45 hey, are there vim users ? I have an incomprehensible problem calling :inoremap test it reacts not normally. 2014-09-15 13:07:13 The only way I got it work is to call vim with vim -i NONE -U NONE -i NONE -N 2014-09-15 13:35:55 Goodbox, http://chezmoicamarche.com 2014-09-15 13:36:14 <^7heo_> v_v 2014-09-15 13:36:21 <^7heo_> that's so lame it can only be french. 2014-09-15 13:36:35 How is it lame ? 2014-09-15 13:36:43 <^7heo_> First, "it works" is the default apache page. 2014-09-15 13:36:51 I know 2014-09-15 13:37:00 <^7heo_> Then the proper translation would be "It works for me", or "It works here" 2014-09-15 13:37:09 <^7heo_> but don't ask french people to translate *properly* 2014-09-15 13:37:18 <^7heo_> and then, apache... 2014-09-15 13:37:22 <^7heo_> really? 2014-09-15 13:37:40 hahah :D 2014-09-15 13:39:31 seriously your inoremap with vim-7.4-r2 on alpine works as attempted ? 2014-09-15 13:39:39 I'm french, you know :) Goodbox is, too 2014-09-15 13:39:45 :) 2014-09-15 13:39:47 Goodbox, yeah it works 2014-09-15 13:40:05 When in insert mode, pressing t-e-s-t goes back to normal mode 2014-09-15 13:40:06 <^7heo_> Damn, French infestation here too... 2014-09-15 13:41:49 then :inoremap alone returns what i was expecting for, but typing :inoremap test and then "test" writes me "" on the buffer editing :/ 2014-09-15 13:43:29 haha 2014-09-15 13:43:41 hmm well, rebuild :D 2014-09-15 13:43:51 Try to insert a "real" escape 2014-09-15 13:44:02 press "" 2014-09-15 13:44:05 and then escape 2014-09-15 13:44:13 it will insert "^[" 2014-09-15 13:44:27 which is the escape char 2014-09-15 13:44:33 yeah we're ok 2014-09-15 13:44:53 do you have any plugin ? 2014-09-15 13:45:36 Nope no one, hmm i first thougt my termcaps 2014-09-15 13:47:19 Goodbox: I use vim here 2014-09-15 13:47:34 and does this work for you too ? 2014-09-15 13:47:46 how do i reproduce the issue? 2014-09-15 13:47:57 :inoremap test 2014-09-15 13:48:08 go in insert mode, type "test" 2014-09-15 13:48:17 see if it brings you back to normal mode 2014-09-15 13:48:42 Goodbox, trying to map "jk" / "kj" to escape ? 2014-09-15 13:48:43 yeah, that's it 2014-09-15 13:48:48 same 2014-09-15 13:49:10 test goes back to navigation mode here 2014-09-15 13:49:13 or what it is called 2014-09-15 13:49:19 noraml :à 2014-09-15 13:49:25 normal* 2014-09-15 13:49:29 k thanks 2014-09-15 13:50:20 >< 2014-09-15 13:50:41 I wiped all confs 2014-09-15 13:50:52 the only way I got it works is: 2014-09-15 13:51:31 Did you use the "original" APKBUILD file ? 2014-09-15 13:51:36 export TERM=xterm; vim -i NONE -U NONE -N 2014-09-15 13:52:03 Actually this is the generic binary for i386 2014-09-15 13:52:21 I don't thing -i NONE is relevant 2014-09-15 13:53:38 Hmmm 2014-09-15 13:53:44 and "-U" is not available in the alpine version 2014-09-15 13:54:02 then that's juste the nocompatible mode 2014-09-15 13:54:13 and for the -N switch, well, as long as you have a vimrc, it is set 2014-09-15 13:54:22 you DO have a vimrc, don't you ? 2014-09-15 13:54:58 I deleted it ... 2014-09-15 13:55:08 aaah ! 2014-09-15 13:55:09 but same with 2014-09-15 13:55:13 here we are ^^ 2014-09-15 13:55:17 I tried wit 2014-09-15 13:55:23 with :set compatible 2014-09-15 13:55:34 and it effectively insert "" 2014-09-15 13:55:39 :> ~/.vimrc 2014-09-15 13:55:42 \o/ 2014-09-15 13:55:43 and try again 2014-09-15 13:56:23 yeah really strange, I never enabled nocompatible in the past 2014-09-15 13:56:34 I told ya 2014-09-15 13:56:38 but let's be nocompatible from now :) 2014-09-15 13:56:56 'compatible' 'cp' boolean (default on, off when a |vimrc| or |gvimrc| file is found) 2014-09-15 13:57:22 Most distro provide a default vimrc, and set nocompatible for you 2014-09-15 13:57:25 alpine don't 2014-09-15 13:57:43 so you use a totally vanilla vim 2014-09-15 13:57:52 which IS by default, vi-compatible :)à 2014-09-15 13:58:09 That takes sence :) thanks you! 2014-09-15 13:58:15 no prob 2014-09-15 13:58:24 if you plan on using vim without a vimrc, 2014-09-15 13:58:33 you'd better alias vim to vim -N 2014-09-15 13:58:47 though I don't see how you could use vim without a vimrc :P 2014-09-15 13:59:10 i do not :p 2014-09-15 13:59:16 ;) 2014-09-15 14:00:02 you could package https://github.com/tpope/vim-sensible I guess 2014-09-15 14:00:32 and nvi for me, none of that blinkenlights stuff! 2014-09-15 14:02:37 hmm, i'm ok for package and try out nvi ! 2014-09-15 14:03:26 after that, going in town to buy coffee 2014-09-15 14:05:52 ow it seem uses lisp and lacks unicode support 2014-09-15 14:06:04 finally I prefer buying coffee first 2014-09-15 14:06:50 brb 2014-09-15 14:08:40 hehe 2014-09-15 14:08:42 good choice 2014-09-15 15:08:51 Just remember now that vi/vim starts nvi on bsd's; By the nvi/vi default disturbing backspace behaviour ^^ 2014-09-15 15:11:39 vi -> nvi there, vim is vim, but maybe not the rhellish vim-enhanced? 2014-09-15 15:12:18 maybe this disturbing thing is what i always miss... will look into that 2014-09-15 15:14:43 ha do not annoy you, I'm just a joung newbie... 2014-09-15 15:16:30 Goodbox: nah, you just told me something new, that's good 2014-09-15 15:16:46 i just know vim sometimes does things i don't understand when i'm in really fast type mode 2014-09-15 15:16:56 now i can maybe track them down :) 2014-09-15 15:20:37 haha anyway basics vi gestures is good to know 2014-09-15 15:22:31 I can't browse without vimperator 2014-09-15 15:22:42 and sometimes uses ranger, 2014-09-15 15:23:10 but never used the shell vi mode for long time 2014-09-15 15:26:39 Goodbox: i use the shell mode all day, the things that trip me are just the vim "addons" 2014-09-15 15:26:57 nothing beats esc-v turning a oneliner into a proper script :) 2014-09-15 15:27:57 that's a shell function which uses command line history ? 2014-09-15 15:29:06 terminal vim? shell-mode sounds like 'ex' to me 2014-09-15 15:29:40 yeah just for the line editing 2014-09-15 15:30:25 i'll be harsh: anyone thinking emacs mode is en par with vi mode doesn't know vi mode well (or is an emacs god) 2014-09-15 15:39:03 I found, entering editing to the wanted position vi intuitivity is really shorter than to learn many emacs commands which can't do theses quick actions 2014-09-15 17:33:56 question, how can I check the contents of a package/port using apk? 2014-09-15 17:38:26 Amnesia: apk info -L $pkgname 2014-09-15 17:40:43 pnutzh4x0r: thanks 2014-09-15 18:35:25 is anyone using bitlbee over here? 2014-09-15 18:35:52 the reason I'm asking, is because it's crashing when I connect to it:/ 2014-09-15 18:38:23 i haven't tried bitlbee on alpine, but i know it needs some minor patching to work with musl 2014-09-15 18:38:44 they're misusing iconv and have some nonstandard charset names 2014-09-15 18:42:04 got any idea how far away it's on the roadmap? 2014-09-15 18:44:07 ACTION is about to try and give alpine a shot as desktop os:-) 2014-09-15 18:45:06 i'm using it on a laptop fairly successfully. obviously there's a lot less packaged desktop software than big dists 2014-09-15 18:46:00 I'm used to that, previously I ran crux 2014-09-15 18:47:52 so how do you folks react on known vulns? 2014-09-15 18:48:14 fixing them, i think :) 2014-09-15 18:49:18 ff 29.0 is kinda outdated though:/ 2014-09-15 18:55:07 Amnesia: yea i use bitlbee, it gives some glib errors but it works for me 2014-09-15 18:55:20 Evil_Bob: teh interesting.. 2014-09-15 18:55:30 cause It's unusable for me:/ 2014-09-15 18:56:34 makes me wonder whether it's a pebcak now:p 2014-09-15 18:57:15 does grsec/pax intervene when you want to connect to localhost? 2014-09-15 18:59:09 ? 2014-09-15 19:01:23 nvm, it doesnt 2014-09-15 19:21:16 im also using bitlbee 2014-09-15 19:37:57 pnutzh4x0r: it's working with the newest version in edge 2014-09-15 19:37:59 \o/ 2014-09-15 19:43:06 is there a file somewhere containing a list of files owned per package? (even if not installed), similar to the slackware MANIFEST file? 2014-09-15 19:47:11 Evil_Bob: u can check an individual installed package w/ apk info -L $pkgname 2014-09-15 19:47:40 otherwise, i think u have to extract the list from each package if u want a global list 2014-09-15 21:10:06 question, when you forgot to supply the --purge argument to a shitload of "apk del" commands, your system will be cluttered with orphaned conf files. does alpine support some sort of mechanism to remove them? 2014-09-15 21:16:11 reinstall and deinstall with purge is a bit of a bruteforce approach, but might work 2014-09-15 21:17:59 :/ 2014-09-15 21:20:22 isn't there some kind of tool that lists all files that aren't handled by the package manager? 2014-09-15 21:40:05 i think apk audit can also just show the ones that came with a package 2014-09-15 21:40:12 but give it a shot, just to be sure 2014-09-15 21:40:44 that aside, what is going on on your system hehe 2014-09-16 02:25:06 Does alpine support the GMA500? 2014-09-16 02:50:57 Does AL have packages for Enlightenment? 2014-09-16 03:52:07 ACTION was reading briefly about Intel Edison today. Running an Atom processor. Thoughts on running alpine on it? 2014-09-16 03:53:14 heading to bed now. maybe dream of projects. :) 2014-09-16 05:00:25 is there a latex package for alpine? 2014-09-16 06:13:32 morning 2014-09-16 06:13:43 UserError: kernelconfig.x86:CONFIG_DRM_GMA500=m 2014-09-16 06:13:43 kernelconfig.x86_64:CONFIG_DRM_GMA500=m 2014-09-16 06:13:50 we have the kernel module there 2014-09-16 06:14:14 does it need special userland tools? 2014-09-16 06:14:23 or special x86-video-* driver? 2014-09-16 06:14:28 i havent used it myself 2014-09-16 06:23:34 UserError: we have some outdated enlightenment packages in testing 2014-09-16 06:23:48 niceeee 2014-09-16 06:24:06 and that is gma500 2d acceleration not just modeset 2014-09-16 06:24:30 i'm not sure i know intel has a driver? 2014-09-16 06:24:35 emgd 2014-09-16 06:27:01 modesetting 2014-09-16 06:27:08 i think you can use xf86-video-modesetting 2014-09-16 06:29:22 this is a nice wiki: https://wiki.archlinux.org/index.php/poulsbo 2014-09-16 06:33:34 darkfader: the packages aren't installed anymore, there's just some residue left 2014-09-16 06:36:57 Amnesia: apk info -L $(apk info) 2014-09-16 06:37:06 should give a list of all files managed by package manager 2014-09-16 06:37:18 i suppose you could out that to a tmpfile 2014-09-16 06:37:36 then do: cd /; find usr/ etc/ bin/ sbin/ lib/ 2014-09-16 06:37:47 and out that to other file 2014-09-16 06:37:48 yeah, and diff it 2014-09-16 06:37:50 sort and compare 2014-09-16 06:37:59 guess I'll do something like that yes 2014-09-16 06:37:59 you need to sort too i think 2014-09-16 06:38:58 ok 2014-09-16 06:39:29 I guess I'll fix my other issues first:P 2014-09-16 06:39:40 migrating always sucks^^ 2014-09-16 06:39:52 tend to be some work yes :-/ 2014-09-16 06:40:11 many things on alpine could have use of some polish 2014-09-16 06:40:20 there was issue with bitlee in v3.0? 2014-09-16 06:41:03 on the main branch yes 2014-09-16 06:41:18 I've upgraded to edge and now it's working like a breeze 2014-09-16 06:45:02 ok 2014-09-16 06:45:09 do you know what was wrong? 2014-09-16 06:46:31 ncopa: nope 2014-09-16 06:46:59 but the crash occured when I connected to the daemon from localhost 2014-09-16 06:47:00 edge has 3.2.2 and uses libotr (4.0.0) instead of libotr3 (3.2.1) 2014-09-16 06:47:11 ok, so it is a segfault 2014-09-16 06:47:16 yep 2014-09-16 06:48:15 Amnesia: do you think you could file a bug for it? http://bugs.alpinelinux.org/projects/alpine/issues/new 2014-09-16 06:48:37 write that it segfaults in v3.0, but works in edge 2014-09-16 06:49:49 sure 2014-09-16 06:54:16 ncopa: I'll be porting some packages from crux to alpine soon 2014-09-16 06:54:22 nice 2014-09-16 06:55:05 hopefully It won't required a lot of patching^.^ 2014-09-16 06:55:20 depends on how portable the package is 2014-09-16 06:55:45 most packages are pretty good at it, and needs no or very little patching 2014-09-16 06:55:49 others needs lots... 2014-09-16 06:56:04 https://privatepaste.com/d234a68a64 2014-09-16 06:56:32 https://privatepaste.com/7c4cbdb127 2014-09-16 06:56:40 that second one's a little bit clearer:-) 2014-09-16 06:58:11 avr-* is going to be a pita though^^ 2014-09-16 07:03:36 we have lots of those 2014-09-16 07:04:10 yeah, I'll start with the ones I actually use myself:) 2014-09-16 07:04:24 lua-bitop? 2014-09-16 07:04:27 i think we have bitlib 2014-09-16 07:04:53 acpi, we have busybox acpi 2014-09-16 07:05:09 bitlbee we have 2014-09-16 07:05:15 ca-certificates 2014-09-16 07:05:38 (infact we have a much nicer certificate update script than most others) 2014-09-16 07:06:01 dwm i think someone sent a patch recently 2014-09-16 07:06:07 ffmpeg we have 2014-09-16 07:06:13 firefox we have 2014-09-16 07:06:24 luaexpat -> lua-expat 2014-09-16 07:06:29 luasocket -> lua-socket 2014-09-16 07:06:36 busybox acpi? 2014-09-16 07:06:55 yup, an acpi implementation 2014-09-16 07:06:58 shipped with busybox 2014-09-16 07:07:30 http://sprunge.us/YFEN 2014-09-16 07:07:45 we have mpv 2014-09-16 07:07:50 ah 2014-09-16 07:07:53 qemu 2014-09-16 07:08:13 i actually use virt-manager/libvirt/qemu on my work desktop 2014-09-16 07:08:26 rdesktop i think we have, but I kinda prefer remmina 2014-09-16 07:08:52 regarding acpid, how can I check the percentage of capacity left in my battery? 2014-09-16 07:09:07 find /sys -iname "capacity" -exec cat {} \; is ugly 2014-09-16 07:09:10 good question :) 2014-09-16 07:09:11 and needs to be ran as root:P 2014-09-16 07:12:59 i actually think busybox acpid can be used to monitor battery 2014-09-16 07:13:06 need config /etc/acpid.map 2014-09-16 07:13:16 how the format there is, i dont know 2014-09-16 07:13:21 but it looks like it is possible 2014-09-16 07:14:08 http://lists.busybox.net/pipermail/busybox/2010-September/073462.html 2014-09-16 07:14:44 hm I'll take a look at that later on.. 2014-09-16 07:15:17 btw, since xorg-server 1.16.0 it's impossible to go to tty1 when X's started:/ 2014-09-16 07:15:24 am I the only one where that's occuring? 2014-09-16 07:15:45 i can do alt-f1 2014-09-16 07:16:04 but i think you need fbcon 2014-09-16 07:16:18 dafuq, then that's new.. 2014-09-16 07:16:29 could be yes 2014-09-16 07:16:37 i think they use kms by default 2014-09-16 07:16:44 what's the package called ? 2014-09-16 07:16:47 and afaik, kms requires fbcon 2014-09-16 07:16:50 its a kernel module 2014-09-16 07:16:56 so: modprobe fbcon 2014-09-16 07:17:00 ah 2014-09-16 07:17:09 echo 'fbcon >> /etc/modules' 2014-09-16 07:17:17 echo 'fbcon' >> /etc/modules 2014-09-16 07:17:27 yep, that does the job 2014-09-16 07:17:28 t 2014-09-16 07:17:29 y 2014-09-16 07:32:55 are you running alpine on a desktop/laptop aswell? 2014-09-16 07:32:57 as well* 2014-09-16 07:38:31 yes 2014-09-16 07:38:43 both my work desktop and my private laptop 2014-09-16 07:39:21 how did you setup the chronyd init script? 2014-09-16 07:39:35 cause without network connectivity it blocks the boot process 2014-09-16 07:39:48 I could edit the init scripts,but that's ugly 2014-09-16 07:40:30 server pool.ntp.org 2014-09-16 07:40:30 initstepslew 10 pool.ntp.org 2014-09-16 07:40:30 commandkey 10 2014-09-16 07:40:30 keyfile /etc/chrony/chrony.keys 2014-09-16 07:40:30 driftfile /etc/chrony/chrony.drift 2014-09-16 07:40:31 makestep 1000 -1 2014-09-16 07:41:02 but i always have net on my desktop 2014-09-16 07:43:28 but not on your laptop right:P? 2014-09-16 07:44:12 i'm checking 2014-09-16 07:45:20 http://sprunge.us/YFPF 2014-09-16 07:50:07 yes, that's the default config 2014-09-16 07:50:43 should we enable glamor in xorg server? 2014-09-16 07:52:37 sesecrete 2014-09-16 08:08:29 <^7heo_> what is glamor? 2014-09-16 08:09:02 <^7heo_> ACTION will maybe set an ignore mask on "proxad.net" one of those days. Might save a lot of time. 2014-09-16 08:13:25 glamor is 2d acceleration via opengl or similar 2014-09-16 08:16:23 old style is to do 2d acceleration via ddx 2014-09-16 08:16:45 but driver developers are tired of writing 2 set of drivers: ddx and gl 2014-09-16 08:17:14 and since they have some overlapping functionality, they want do the 2d acceleration using the already written gl driver 2014-09-16 08:17:17 and drop ddx 2014-09-16 08:18:06 sounds lke we want glamor support 2014-09-16 09:42:06 glamor === gallium? 2014-09-16 09:57:49 What are the compile targets for 32 and 64bit alpine packages? 2014-09-16 09:58:45 ? 2014-09-16 09:58:54 compile targets? 2014-09-16 09:58:58 You use gcc no? 2014-09-16 09:59:02 yes 2014-09-16 09:59:31 for example ubuntu targets the pentium pro~ or there abouts no mmx etc 2014-09-16 09:59:46 ok 2014-09-16 09:59:49 on 64bit they do up to sse2 whatever intel or amd model that is 2014-09-16 10:00:07 for 64bit its gcc default 2014-09-16 10:00:35 on x86 i think we use i486 2014-09-16 10:28:15 <[xming]> how do I find out which pkgs does a (installed) file belongs to? 2014-09-16 10:32:38 [xming]: apk info -W $filename 2014-09-16 10:37:09 <[xming]> I am trying to find out which package installed /etc/terminfo/* 2014-09-16 10:37:21 <[xming]> apk info -W /etc/terminfo/ansi 2014-09-16 10:37:30 <[xming]> ERROR: /etc/terminfo/ansi: Could not find owner package 2014-09-16 10:40:43 <[xming]> it seems that I am missing /etc/terminfo/* in some containers 2014-09-16 10:46:07 <[xming]> ah found it it's ncurses-base 2014-09-16 12:09:48 http://youtu.be/bmav517MQJc 2014-09-16 12:12:26 <^7heo_> Amnesia: what is this all about? 2014-09-16 12:14:59 check it out for yourself 2014-09-16 12:18:39 <^7heo_> Amnesia: I can't, I'm at work 2014-09-16 12:18:51 ah k 2014-09-16 12:20:21 Do people still actually use pentium 1's (non and +mmx) & ppros? 2014-09-16 12:20:35 i think so 2014-09-16 12:21:00 there was a MB with i486 cpu architecture, running on 200Mhz or so 2014-09-16 12:21:20 ALIX maybe 2014-09-16 12:21:22 WinChip? 2014-09-16 12:21:25 ah 2014-09-16 12:23:00 Secondly, is busybox compatible with link time optimization? 2014-09-16 12:23:27 dunno 2014-09-16 12:28:03 ALIX is i586 w/ mmx(ext) I believe, but it has some other oddity that makes it incompatible with pentium pros 2014-09-16 12:28:47 interesting. is it still made today? 2014-09-16 12:28:54 it is 2014-09-16 12:29:03 have a bunch of them here right now 2014-09-16 12:29:33 I just wasn't aware of x86 other than winchip, cyrix/via/whatever, amd, intel 2014-09-16 12:29:42 600MHz i586 isn't all that bad and it has lots of perihepals 2014-09-16 12:30:08 Is it mainly for headless operation? 2014-09-16 12:30:10 it's AMD Geode to be precise, which I think is based on cyrix 2014-09-16 12:30:24 oh i remember the geode 2014-09-16 12:30:46 actually HMI touchscreens here 2014-09-16 12:31:03 but it used to be popular for wifi routers 2014-09-16 12:31:19 This whole time I was wondering why most distributions kept their 32bit compile targets so low even for X11+ packages 2014-09-16 12:32:15 I guess it's gonna be eventually replaced now that they're making http://pcengines.ch/apu.htm 2014-09-16 12:33:38 That's cool, they recycled the bobcat cores 2014-09-16 12:35:18 still waiting for video-enabled one, also got cubietruck for testing, but didn't get around to turn it on yet 2014-09-16 12:35:36 allwinner is a nice bit of kit 2014-09-16 12:35:56 not a fan of binary blobs 2014-09-16 12:36:01 linux sunxi development has far accelerated past where i thought it would be 2014-09-16 12:36:08 for what? the mali? 2014-09-16 12:36:09 <^7heo_> or you would be on #Gnome 2014-09-16 12:36:11 <^7heo_> or #ubuntu 2014-09-16 12:36:13 <^7heo_> or #KDE 2014-09-16 12:36:14 <^7heo_> or... 2014-09-16 12:36:31 what? 2014-09-16 12:37:19 UserError: yeah, fwih lima-driver is far from stable yet. and there might be other stuff on the soc 2014-09-16 12:46:57 Yeh, CedarX has been a pain. Not sure if they stuck with it after the A10/s/13 though 2014-09-16 12:48:53 main difference i586/ppro as far as compat goes is having or not having PAE i think 2014-09-16 12:49:48 The first one or two P-M's I don't believe have PAE either 2014-09-16 12:50:46 only the last ones had it probably (>1.2GHz) 2014-09-16 12:51:27 i still have one laptop with a 266mhz mmx, didn't manage to install alpine on it yet :/ 2014-09-16 12:51:46 it has a godly keyboard, so it's something i really want to use 2014-09-16 13:34:54 ncopa: I see mpv is still on 0.5.0, although 0.5.3 was released 2014-09-16 14:16:36 wm4: ok, i'll upgrade it 2014-09-16 14:16:49 we need a 'flag-outdated-package' feature 2014-09-16 15:10:41 <^7heo_> btw, I have a "can't open dev/X" 2014-09-16 15:10:49 <^7heo_> or more exactly, "can't open /dev/X" 2014-09-16 15:10:56 <^7heo_> I did put startx in my /etc/inittab 2014-09-16 15:11:02 <^7heo_> should I change something? 2014-09-16 15:11:12 <^7heo_> because /dev/X isn't a usual node, is it? 2014-09-16 15:20:07 ncopa: I was just wondering whether there's a better method to let package maintainers know of new version of a specific software 2014-09-16 15:23:31 <^7heo_> I guess that my /dev/X problem has to do with rootless X? 2014-09-16 15:27:23 <[xming]> how do I do this with awall? iptables -t nat -A PREROUTING -p tcp --dport 5060 -j DNAT --to 192.168.1.6:5070 2014-09-16 15:28:14 Hello 2014-09-16 15:28:34 <[xming]> hi box :D 2014-09-16 15:28:44 [xming]: adding -i ? 2014-09-16 15:29:15 <[xming]> with "awall" and its json config 2014-09-16 15:29:42 ah :) 2014-09-16 15:30:30 <[xming]> maybe I can't ditch shorewall yet 2014-09-16 15:31:01 sorry never tried 2014-09-16 15:31:09 <[xming]> np 2014-09-16 15:31:46 <^7heo_> ncopa: any idea on my /dev/X problem? 2014-09-16 15:58:55 <[xming]> oh I didn't know that alpine is a fork of LEAF/LRP 2014-09-16 17:27:02 Is there anywhere a collection of Alpine Linux artwork? Something like logos, backgrounds, etc 2014-09-16 18:40:23 is there a latex package for alpine? 2014-09-16 18:44:43 <[xming]> this is the only match abiword-plugin-latex-3.0.0-r3 :p 2014-09-16 18:57:04 yeah I noticed, but sometimes there's a package with a sneaky name 2014-09-16 19:19:55 <[xming]> maybe you can port teTeX? :D 2014-09-16 21:26:18 well, it's worth the effort to me, we'll see what comes of it 2014-09-17 06:13:24 morning 2014-09-17 06:13:32 we havent packaged latex 2014-09-17 06:14:24 jomat: we don't have many alpine linux designers 2014-09-17 06:14:30 we have a new logo coming up 2014-09-17 06:17:01 http://bugs.alpinelinux.org/issues/2882 2014-09-17 06:17:35 but we have not made any collection of it yet 2014-09-17 07:15:49 <^7heo_> moin 2014-09-17 08:06:54 <^7heo_> ncopa: any plans for getting webkit-1.0 in the packages? 2014-09-17 08:11:36 <^7heo_> okay, please lemme reformulate: is there any webkitgtk-2.0 somewhere? 2014-09-17 08:11:40 <^7heo_> (in alpine packages) 2014-09-17 08:13:12 <^7heo_> the "search file in all packages" function would be VERY great. 2014-09-17 08:13:16 i think it is 2014-09-17 08:13:24 <^7heo_> or 'serach parttern in all packages' 2014-09-17 08:13:29 yeah, there is no file db 2014-09-17 08:13:30 <^7heo_> katlogic: where? 2014-09-17 08:13:51 webkitgtk? 2014-09-17 08:13:54 + -dev 2014-09-17 08:14:02 <^7heo_> 'cause i built suckless surf with webkitgtk-3.0 and now it crashes on load, saying that the versions of the symbol don't match 2014-09-17 08:14:06 <^7heo_> which is logic. 2014-09-17 08:14:14 <^7heo_> katlogic: that gets you webkitgtk-3.0 2014-09-17 08:14:19 <^7heo_> not the one for gtk-2.0 2014-09-17 08:14:49 <^7heo_> in other words, if I want surf on Alpine, I need to port surf to GTK-3 first? 2014-09-17 08:15:08 <^7heo_> that'd be too bad, 'cause alpine's great, surf's great, but I don't want to port it =/ 2014-09-17 08:15:40 yeah, looks like theres been nothing deping on webkitgtk2 so far 2014-09-17 08:15:51 <^7heo_> yeah 2014-09-17 08:15:58 <^7heo_> I really wonder if I should port surf 2014-09-17 08:15:59 <^7heo_> I mean 2014-09-17 08:16:05 <^7heo_> gtk-2.0 is ugly and deprecated... 2014-09-17 08:16:07 <^7heo_> so... 2014-09-17 08:16:15 well yeah 2014-09-17 08:16:33 might as well port it to webkit-qt :) 2014-09-17 08:16:37 which is neat and awesome! 2014-09-17 08:16:54 <^7heo_> I don't like qt 2014-09-17 08:16:58 hehe 2014-09-17 08:17:30 <^7heo_> GTK's written in C, Qt in C++ 2014-09-17 08:17:35 <^7heo_> all's being said here. 2014-09-17 08:17:37 ^7heo_: I dont think anyone would object if you post webkitgtk2-dev package along surf, its a reasonable dependency. 2014-09-17 08:17:42 ^7heo_: 2014-09-17 08:17:50 Irrational fear of C++ I see 2014-09-17 08:17:57 you do realize webkit is written in C++ too :) 2014-09-17 08:17:59 <^7heo_> not irrational, not fear. 2014-09-17 08:18:07 <^7heo_> when there are no alternatives... 2014-09-17 08:18:17 <^7heo_> please link a web engine written in pure C. 2014-09-17 08:18:25 links 2014-09-17 08:18:31 <^7heo_> ? 2014-09-17 08:18:38 links, a browser :) 2014-09-17 08:18:38 <^7heo_> that's not a web engine. 2014-09-17 08:18:42 <^7heo_> it's a browser. 2014-09-17 08:18:54 what the hell is a web engine 2014-09-17 08:19:02 <^7heo_> gecko, webkit, etc. 2014-09-17 08:19:10 yes; a browser. css/html renderer. 2014-09-17 08:19:21 <^7heo_> Gecko is a web browser engine 2014-09-17 08:19:23 <^7heo_> https://en.wikipedia.org/wiki/Gecko_%28software%29 2014-09-17 08:19:32 <^7heo_> It's not the same as a browser. 2014-09-17 08:19:43 I'm not here to sperg about it. 2014-09-17 08:19:43 <^7heo_> It's like if you call ncurses an application. 2014-09-17 08:19:50 If you really insist, there is Arachne. 2014-09-17 08:20:03 which had its rendered prepared out to be used in other projects. 2014-09-17 08:20:09 <^7heo_> link? 2014-09-17 08:20:10 *renderer 2014-09-17 08:20:18 <^7heo_> 'cause I find a lot of different stuff with that name, but no browser engine. 2014-09-17 08:20:33 ^7heo_: No need to, you seem to be competent with google and wiki. 2014-09-17 08:20:47 <^7heo_> yeah, but I'm not competent at finding stuff you make up. 2014-09-17 08:21:26 Sorry, but arachne is real. 2014-09-17 08:21:37 as well as is embrowser derived from it 2014-09-17 08:21:39 <^7heo_> yeah, as a web browser. 2014-09-17 08:21:44 <^7heo_> https://en.wikipedia.org/wiki/Arachne_%28web_browser%29 2014-09-17 08:21:50 <^7heo_> Now, as a web engine... 2014-09-17 08:22:59 <^7heo_> Anyway, to cut the crap... 2014-09-17 08:23:03 ^7heo_: As for a web engine, thats just fancy name for renderer separated out of browser. 2014-09-17 08:23:20 <^7heo_> I would like to know if it's possible for ncopa to package webkit for gtk-2.0 or not... 2014-09-17 08:23:35 <^7heo_> katlogic: exactly. It's a component that is isolated in its code. 2014-09-17 08:23:36 Again, just do it :) 2014-09-17 08:23:40 <^7heo_> So the work hasn't to be done by me. 2014-09-17 08:23:44 <^7heo_> or by someone else. 2014-09-17 08:23:52 <^7heo_> And no, I'm not going to re-invent the wheel. 2014-09-17 08:23:54 thats what mailing list archive is for. 2014-09-17 08:24:01 <^7heo_> Especially for something as absurd as the web. 2014-09-17 08:24:13 <^7heo_> mailing list requires mail. 2014-09-17 08:24:16 <^7heo_> I have only IRC here. 2014-09-17 08:24:19 <^7heo_> So I ask on IRC. 2014-09-17 08:24:21 <^7heo_> simple, right? 2014-09-17 08:25:16 > archive 2014-09-17 08:25:39 <^7heo_> that's nice music alright. 2014-09-17 08:26:09 spergcore :) 2014-09-17 08:26:36 <^7heo_> is that music? 2014-09-17 08:36:39 ^7heo: everything is possible. the impossible just takes longer time to get done... 2014-09-17 08:37:08 gtk2 is no longer maintained upstream really iirc 2014-09-17 08:37:27 so i'd prefer not add new gtk2 stuff 2014-09-17 08:37:36 <^7heo_> ncopa: makes sense, thanks for your answer. 2014-09-17 08:37:53 <^7heo_> on the other hand, porting surf to GTK-3.0 would make sense I think. 2014-09-17 08:38:05 might, or might not 2014-09-17 08:38:18 seems like many prefers port gtk2->qt rather than gtk2->gtk3 2014-09-17 08:38:19 I'd prefer not to add gtk3 to my systems, because crapload of dependencies and dbus-based spaghetti :-( 2014-09-17 08:38:41 <^7heo_> ncopa: yes, ofc, but Qt isn't an option for those who want C instead of C++ 2014-09-17 08:38:46 yeah, qt sounds like a saner choice there. that or EFL 2014-09-17 08:39:27 but yes, the entire gtk3 situation is sad 2014-09-17 08:40:17 "let's make libgnome the new version of gtk" esentially 2014-09-17 08:40:38 I'd be thrilled to see uzbl-compatible variant that uses the qt webkit 2014-09-17 08:41:07 indeed 2014-09-17 08:41:49 <[xming]> does gtk3 needs those? I thought it was just gnome3 2014-09-17 08:43:45 honestly i dont see the appeal of "lightweight" webkit ui chromes, the moment you link webkit, things kinda cease being lightweight. 2014-09-17 08:44:03 :) 2014-09-17 08:46:16 <[xming]> :D 2014-09-17 08:46:37 <^7heo_> katlogic: it used to be lightweight. 2014-09-17 08:46:41 <^7heo_> that's the point. 2014-09-17 08:47:10 ^7heo_: I dont notice any differences between midori and surf in terms of lightweightness. 2014-09-17 08:47:38 Except surf being this pretentious suckless thing (i mean dwm is awesome at all, but dwm links only to already resident xlibs, not goddamn hunk of foreign C++ code) 2014-09-17 08:48:51 <^7heo_> katlogic: surf has one major difference with midori. It works. 2014-09-17 08:49:39 In terms of "works", as in conforms to latest web fads, i'm afraid we're stuck with firefox and chromium for a long time :( 2014-09-17 08:50:22 Seee, thats the problem with your layout engines. Those alone are not enough to cobble up a modern browser it seems. 2014-09-17 08:50:25 <[xming]> lol IE6 used to be *the* browser which works :p 2014-09-17 08:52:07 <^7heo_> katlogic: I meant works in the sense: doesn't randomly segfault. 2014-09-17 08:52:30 <^7heo_> [xming]: nah, the web used to be made for IE6. That's slightly different :) 2014-09-17 08:52:57 ^7heo_: I'm sure there is another webkit ui chrome linking to gtk3 which does not segfault as much :) 2014-09-17 08:53:52 though I never really managed to crash any of those. they just sucked horribly as a browser :( 2014-09-17 08:53:59 might as well use links 2014-09-17 08:55:06 ... which does not support ActiveG and MozillaRTC either. 2014-09-17 08:56:19 <[xming]> elinks FTW 2014-09-17 09:02:41 i see that arch linux builds both webkit for gtk3 and gtk2 2014-09-17 09:03:04 it should not be too hard to do similar 2014-09-17 09:03:18 and provide a webkitgtk2 package 2014-09-17 09:03:26 it just takes ages to compile 2014-09-17 09:04:25 <^7heo_> yeah 2014-09-17 09:04:30 <^7heo_> but if you can provide it, it's better :) 2014-09-17 09:06:55 could you please file a request on bugs.a.o 2014-09-17 09:06:57 so i dont forget 2014-09-17 09:07:08 i wont be able to it right now 2014-09-17 09:07:20 so i will forget it unless you file a feature request on bugs.a.o 2014-09-17 09:11:43 <^7heo_> ncopa: in aconf I guess? 2014-09-17 09:11:59 <^7heo_> or in alpine-conf? 2014-09-17 09:13:06 ? 2014-09-17 09:13:13 <^7heo_> in bugs.a.o 2014-09-17 09:13:21 <^7heo_> I have to open a project to submit a bug, right? 2014-09-17 09:13:30 in "alpine linux" 2014-09-17 09:13:49 http://bugs.alpinelinux.org/projects/alpine/issues/new 2014-09-17 09:13:54 <^7heo_> danke 2014-09-17 09:14:13 i am trying to figure why xulrunner does not build on x86 2014-09-17 09:14:15 <^7heo_> ooh, I had to click on the very little "projects" link in the top 2014-09-17 09:14:23 <^7heo_> mozilla, that's why 2014-09-17 09:14:42 <^7heo_> damn, I have to create a login/pass 2014-09-17 09:15:44 <^7heo_> I guess that a new account has to be activated with a link on the mail? 2014-09-17 09:17:22 i think so yes 2014-09-17 09:17:35 but i think you can send an email to bugs@alpinelinux.org too 2014-09-17 09:17:41 and it will create an anonymous issue 2014-09-17 09:31:38 <^7heo_> okay, but I can't send/receieve any email for now. 2014-09-17 09:31:43 <^7heo_> could you please validate the user 7heo? 2014-09-17 09:31:47 <^7heo_> so I can login? 2014-09-17 09:31:54 <^7heo_> I just created it. 2014-09-17 09:36:04 ok i activated you 2014-09-17 09:39:52 <^7heo_> vielen dank 2014-09-17 09:40:23 <^7heo_> damn, forgot passwd now v_v 2014-09-17 09:40:37 <^7heo_> f**k 2014-09-17 09:40:55 <^7heo_> okay, well, I'll put a reminder on my phone for sending you an email tonight. 2014-09-17 09:47:11 <^7heo_> http://bugs.alpinelinux.org/issues/3370 2014-09-17 09:47:55 <^7heo_> ncopa: ^ 2014-09-17 09:50:16 very well. thanks! 2014-09-17 10:58:26 <^7heo_> ncopa: is webkit unstable in 3.0.4? 2014-09-17 10:58:38 <^7heo_> because I get a segfault each time I try to execute surf... 2014-09-17 10:58:58 <^7heo_> (I did gdb it, but it was giving me warnings and no useful info) 2014-09-17 10:59:08 <^7heo_> (and the host is down for now, so I can't try it again) 2014-09-17 11:04:44 ^7heo: you probably need paxmark the binary due to jit 2014-09-17 11:05:00 paxctl -c -m /usr/bin/surf 2014-09-17 11:05:02 or similar 2014-09-17 11:15:58 <^7heo_> hmm 2014-09-17 11:16:12 <^7heo_> ncopa: in order to make it work, or in order to get more info on the segfault? 2014-09-17 11:16:44 in order to make it work 2014-09-17 11:16:49 does dmesg give any hints? 2014-09-17 11:17:16 <^7heo_> nah 2014-09-17 11:17:23 <^7heo_> I did shutdown the host, and didn't check it. 2014-09-17 11:17:31 <^7heo_> I could try again, but I have to debug some sshd first. 2014-09-17 14:05:46 hi, i have a problem with webkitgtk i wonder if someone has the same issue. when i open midori and inspect element the screen is blank and i get some warnings to stderr 2014-09-17 14:06:05 with surf i get the same error 2014-09-17 14:09:20 <^7heo_> Evil_Bob: I can tell you as soon as I get surf working. 2014-09-17 14:10:18 ^7heo_: can you install midori and try to inspect element, i get the same error 2014-09-17 14:10:29 its a webkitgtk bug in combination with musl maybe 2014-09-17 14:10:42 i have surf working 2014-09-17 14:11:50 if you g_object_set(G_OBJECT(settings), "enable-developer-extras", FALSE, NULL); in surf it doesnt crash 2014-09-17 14:11:58 <^7heo_> Evil_Bob: I can try, but I have to configure a switch first. 2014-09-17 14:12:06 <^7heo_> So moment. 2014-09-17 14:12:06 ok 2014-09-17 14:12:12 no rush 2014-09-17 14:14:28 Hi. 2014-09-17 14:14:59 I plan to use alpine witn openvswitch and qemu-kvm. Is it the right thing to do ? ;) 2014-09-17 14:15:23 ans it will be client to a CEPH cluster 2014-09-17 14:15:48 <^7heo_> alpine works well in kvm, if that's the question. 2014-09-17 14:16:12 part of it ;) second half is about openvswitch ? 2014-09-17 14:17:28 <^7heo_> never used, no idea. 2014-09-17 14:17:47 I plan to have alpine run from a usb stick with a second one for the customisation. The real HDD used for storage in a CEPH cluster launched from qemu-kvm machines on top of alpine... 2014-09-17 14:19:04 openvswitch is the network part of this scheme... 3 virtual networks for each VM: WAN, LAN and STORAGE. 2014-09-17 14:19:34 <^7heo_> so many buzzwords... 2014-09-17 14:19:37 lol 2014-09-17 14:19:57 I'm preparing this setup for last couple of weeks 2014-09-17 14:19:59 <^7heo_> but you're french, I understand the need of buzzwords, it's a metric for your salary, isn't it? 2014-09-17 14:20:18 But is it synergy cloud responsive distributed enabled? 2014-09-17 14:21:08 Jean-Scotch: it should work. the rbd integration can be a problem, since we don't have the 'rbd' command or ceph tools so far 2014-09-17 14:21:12 <^7heo_> UserError: not sure if it's a best practice, wouldn't it shatter the elasticity of the symmetrical REST API? 2014-09-17 14:21:24 so you'd need to map disks via /sys/bus/rbd/... 2014-09-17 14:21:33 ugly but if you have a few days it's solveable 2014-09-17 14:21:47 or you just attach the storage directly to each vm and not to the host 2014-09-17 14:21:49 then no issue 2014-09-17 14:22:00 note a full ceph port to alpine is *not* in the works 2014-09-17 14:22:01 Only if it's nodejs powered with nosql 2014-09-17 14:22:26 <^7heo_> everybody uses mangodb now, it's a database that's not needing any administration, so admins can do it on their free time! 2014-09-17 14:22:52 That's why I'm plannig ti run CEPH nodes in debian qemau launched from alpine... 2014-09-17 14:22:53 Jean-Scotch: http://bugs.alpinelinux.org/issues/3241 you might wanna bump that 2014-09-17 14:23:49 <^7heo_> debian sucks. 2014-09-17 14:24:08 darkfader: I had read that. but it seems to be resolved in the dev tree 2014-09-17 14:24:08 Jean-Scotch: it's not just about the nodes. 2014-09-17 14:24:22 <^7heo_> sorry, it's a reflex. 2014-09-17 14:24:35 Jean-Scotch: the rbd module is now available, i dont think there's a ceph rbd program to call though? 2014-09-17 14:24:43 Jean-Scotch: did you see a ceph-client package?? 2014-09-17 14:25:57 ^7heo_: when i was really burnt out from job[-2] i tested health by trying to fix some debian vm 2014-09-17 14:26:00 no, but if rbd is enabled in the kernel, qemu should be fine launching VMs from RBD devices 2014-09-17 14:26:19 took about 1 month till brain didn't hurt from touching debian 2014-09-17 14:26:21 Jean-Scotch: ok 2014-09-17 14:26:43 if you have success with that, please /msg me since that would be really a good thing to know 2014-09-17 14:27:53 I'm just waiting for the USB keys to arrive to mount a small lab to test the idea... 2014-09-17 14:28:39 ^7heo_: still configuring that switch? :P 2014-09-17 14:28:41 and for the pure CEPH nodes, I will run those atop a voyage linux install as there is no plan to do it with alpine... 2014-09-17 14:29:20 gonna call it a day, or an afternoon 2014-09-17 14:30:07 <^7heo_> Evil_Bob: well, it's longer than expected, yes. 2014-09-17 14:30:25 <^7heo_> Evil_Bob: some people asked for helpdesk in the meantime, and I discovered another switch I have to provision. 2014-09-17 15:01:19 <^7heo_> Evil_Bob: I have no X in the machine anymore, for some reason. 2014-09-17 15:01:34 <[xming]> http://linux.slashdot.org/story/14/09/17/1258203/torvalds-no-opinion-on-systemd 2014-09-17 15:02:22 <^7heo_> vesa: Ignoring device with a bound kernel driver 2014-09-17 15:02:29 <^7heo_> I have that already... how did I fix it? 2014-09-17 15:04:47 <^7heo_> sudo apk add $(apk search xf86-video* | grep -v doc | rev | cut -d- -f3- | rev) 2014-09-17 15:04:50 <^7heo_> fixed. 2014-09-17 15:05:04 <^7heo_> (nah I don't care to have ALL xorg drivers installed) 2014-09-17 15:08:49 <^7heo_> ncopa: in the end, the `paxctl -c -m /usr/local/bin/surf` did fix the issue. Thanks a bunch. 2014-09-17 15:09:01 <^7heo_> ncopa: btw, I was wondering, is there anything planned for provisionning in alpine? 2014-09-17 15:16:50 ^7heo_: so surf doesnt crash anymore for you? does the web inspector work too? 2014-09-17 15:20:11 <^7heo_> Evil_Bob: do you know how to actiave it from the command line? 2014-09-17 15:20:20 <^7heo_> (i.e. when invoking surf?) 2014-09-17 15:20:49 surf -N (case sensitive) 2014-09-17 15:21:46 then open it with ctrl-shift-o 2014-09-17 15:22:31 <^7heo_> thanks 2014-09-17 15:22:34 <^7heo_> oh wait 2014-09-17 15:24:11 <^7heo_> it has to be done in surf, right? 2014-09-17 15:24:14 <^7heo_> ^O I mean 2014-09-17 15:24:17 yea 2014-09-17 15:24:29 <^7heo_> the X running surf has no hid connected. 2014-09-17 15:24:38 <^7heo_> so wait, checking xdotool 2014-09-17 15:24:44 lol ok 2014-09-17 15:24:52 <^7heo_> not existing in alpine it seems 2014-09-17 15:25:03 <^7heo_> lemme grab a keyboard. 2014-09-17 15:25:34 ^7heo_: btw, so do you use the default alpine kernel (with pax enabled)? 2014-09-17 15:25:55 i use a custom kernel without it 2014-09-17 15:26:40 <^7heo_> okay, the keyboard isn't working 2014-09-17 15:26:53 <^7heo_> I use the default alpine everything 2014-09-17 15:27:04 <^7heo_> excepted the suckless surf that I got via git ofc 2014-09-17 15:27:10 <^7heo_> and same for dwm 2014-09-17 15:28:17 <^7heo_> but the keyboard doesn't work. 2014-09-17 15:28:22 <^7heo_> in any way. 2014-09-17 15:29:17 <^7heo_> as I expected. 2014-09-17 15:29:23 <^7heo_> I have an asus keyboard, it works. 2014-09-17 15:29:28 <^7heo_> the cherry one doesn't. 2014-09-17 15:29:37 <^7heo_> I have to investigate that. 2014-09-17 15:32:32 <^7heo_> Evil_Bob: ^O works if I press it twice, but then the web inspector is empty (white area). 2014-09-17 15:35:19 yea same for me, and i get messages to stderr: ** Message: console message: undefined @0: RangeError: Maximum call stack size exceeded. 2014-09-17 15:35:47 can you try to open: surf -N codemadness.nl/blog/rss.xml 2014-09-17 15:35:57 and open the inspector? it crashes for me 2014-09-17 15:37:39 <^7heo_> I don't need to open the inspector. 2014-09-17 15:37:43 <^7heo_> it segfaults directly. 2014-09-17 15:38:24 <^7heo_> but I get a lot of "** Message: console message: undefined @0: RangeError: Maximum call stack size exceeded" in any case. 2014-09-17 15:38:36 <^7heo_> $ DISPLAY=:0 surf -N codemadness.nl/blog/rss.xml 2014-09-17 15:38:36 <^7heo_> ** Message: console message: undefined @0: RangeError: Maximum call stack size exceeded. 2014-09-17 15:38:40 <^7heo_> ** Message: console message: undefined @0: RangeError: Maximum call stack size exceeded. 2014-09-17 15:38:43 <^7heo_> Segmentation fault 2014-09-17 15:38:46 <^7heo_> that's what happens here. 2014-09-17 15:40:50 ok, yea it segfaults directly too for me, but with surf -n it doesnt, so it makes me think its the webinspector 2014-09-17 15:41:09 atleasts its reproducable, ty for testing :) 2014-09-17 15:41:51 <^7heo_> anytime. 2014-09-17 15:42:42 i dunno if i cba to crawl through the mountains of shit that is glib+gtk+webkit though :) 2014-09-17 15:50:22 <^7heo_> well, glib shouldn't be concerned, as alpine runs on musl 2014-09-17 15:50:24 <^7heo_> or? 2014-09-17 15:52:23 dunno :) 2014-09-17 15:59:30 <[xming]> glib not glibc :D 2014-09-17 16:01:12 <^7heo_> ah 2014-09-17 16:01:17 <^7heo_> my bad 2014-09-17 17:07:12 <^7heo_> oh, a netsplit on freenode! 2014-09-17 17:07:17 <^7heo_> it's been SO long! 2014-09-17 17:57:48 <^7heo_> ncopa: ping 2014-09-17 21:00:44 <^7heo> ncopa: still not here? 2014-09-17 21:58:57 nope 2014-09-18 06:53:07 hoi 2014-09-18 06:53:21 morning 2014-09-18 07:04:39 ^7heo: pong 2014-09-18 07:09:30 <^7heo_> ncopa: :) 2014-09-18 07:09:45 <^7heo_> ncopa: so, my question was: does alpine have a provisionning system? 2014-09-18 07:11:52 <^7heo_> 'cause if not, I was thinking about a simple mechanism (but it would have to be shipped on the iso for maximum efficiency): a sh script that downloads another (say from sprunge), and executes it. 2014-09-18 07:12:00 <^7heo_> the script would be called provision, ideally 2014-09-18 07:12:16 <^7heo_> and would take the sprunge "path" part of the URL as a parameter. 2014-09-18 07:12:28 <^7heo_> with an optional flag to specify the source you want, with a FULL URL. 2014-09-18 07:12:35 <^7heo_> how does it sound? 2014-09-18 07:12:50 <^7heo_> it's not extremely complicated to do, it's basically a curl | sh 2014-09-18 07:13:34 curl | sh sounds risky 2014-09-18 07:14:00 been thinking of pxe boot 2014-09-18 07:14:06 especiallyon http without ssl 2014-09-18 07:14:10 dunno if that goes as provisioning system 2014-09-18 07:14:16 <^7heo_> pxe is VERY teddious to setup. 2014-09-18 07:14:26 <^7heo_> sprunge is https afaik 2014-09-18 07:14:39 <^7heo_> ah no it's http my bad 2014-09-18 07:15:03 <^7heo_> but why is http a problem? I mean it's for a script you just put online 2014-09-18 07:15:38 1) can you trust that sprunge.us really is sprunge.us? without https not really 2014-09-18 07:16:04 <^7heo_> ncopa: okay, in that case, let's use nc. 2014-09-18 07:16:11 <^7heo_> and local network. 2014-09-18 07:16:15 with https, you at least have (the CA's i have installed says sprunge.us is who he claims to be) 2014-09-18 07:16:23 <^7heo_> this is very simple to setup also. 2014-09-18 07:16:25 is anyone over here using perl scripts with irssi? 2014-09-18 07:16:29 <^7heo_> and all you need is an IP and a port. 2014-09-18 07:16:38 <^7heo_> Amnesia: I was. Or have been. Why? 2014-09-18 07:16:56 <^7heo_> ncopa: SSL can easily be spoofed. 2014-09-18 07:17:03 <^7heo_> ncopa: so it's not really secure either. 2014-09-18 07:17:13 <^7heo_> if you want something secure, local http is more secure than remote https. 2014-09-18 07:17:20 ^7heo_: cause Irssi's telling me "script" is an unknown command 2014-09-18 07:17:25 well the CA is supposed to guarantee it 2014-09-18 07:17:30 Good Morning everyone! 2014-09-18 07:17:33 but yes, i am generally sceptic 2014-09-18 07:17:39 need think how to do it properly 2014-09-18 07:17:50 got irssi-0-.8-16-r1 / irssi-perl installed 2014-09-18 07:17:52 <^7heo_> certificate pinning is how to do it properly. 2014-09-18 07:17:59 <^7heo_> but then you have a LOT of problems. 2014-09-18 07:18:03 sounds like I just missed something awesome! 2014-09-18 07:18:36 hi ginjachris, we're talking about provisioning system 2014-09-18 07:18:37 <^7heo_> because MANY bullshit CDNs are exchanging certificates faster than teens are exchanging corporal fluids. 2014-09-18 07:18:59 ^7heo: i think https is overrated yes 2014-09-18 07:19:42 oh i had a drunken rant on here about https being ruined a while back :S 2014-09-18 07:22:19 <^7heo_> it's basically ruined yes. 2014-09-18 07:22:20 <^7heo_> So 2014-09-18 07:22:24 <^7heo_> what about that nc solution? 2014-09-18 07:22:37 <^7heo_> you just cat "script" | nc -l 8080 2014-09-18 07:22:52 <^7heo_> and then on the new-host side you just "provision " 2014-09-18 07:22:58 <^7heo_> and bam. 2014-09-18 07:23:05 <^7heo_> what 'bout that idea? 2014-09-18 07:23:11 <^7heo_> it's kept local, so fine. 2014-09-18 07:23:20 <^7heo_> and if you want to do it remotely, you can tunnel via ssh 2014-09-18 07:23:25 <^7heo_> which is secure too. 2014-09-18 07:23:29 <^7heo_> so? 2014-09-18 07:23:51 ^7heo_: which version do you have installed of irssi? 2014-09-18 07:24:15 <^7heo_> Amnesia: Irssi: Client: irssi 0.8.15 (20100403 1617) 2014-09-18 07:24:21 <^7heo_> Amnesia: on debian, currently. 2014-09-18 07:24:37 ah, that explains:) 2014-09-18 07:24:53 Amnesia: i updated perl to 5.20.1 yesterday 2014-09-18 07:24:54 <^7heo_> (yeah I had to use debian since alpine doesn't work with the keyboards we have here, for a reason I have to elucidate) 2014-09-18 07:25:11 i didnt check if everything linked to libperl needs rebuild 2014-09-18 07:25:17 i assume it doesnt 2014-09-18 07:25:53 <^7heo_> irssi might need :) 2014-09-18 07:26:07 my binary hasn't been linked to libperl 2014-09-18 07:26:20 nc looks cool :) But yeah, you'd want to encrypt somehow (like SSH as you say) 2014-09-18 07:27:03 btw, when I upgrade, wayland gets installed... while it doesn't have any reverse dependencies at all.. 2014-09-18 07:28:28 Amnesia: what happens if you apk del wayland? 2014-09-18 07:28:39 it will tell you what holds it back 2014-09-18 07:28:40 ncopa: it just gets removed 2014-09-18 07:28:47 there isn't anything that holds it back 2014-09-18 07:28:50 ok, problem solved then i suppose :) 2014-09-18 07:29:43 ^7theo: i am a bit interested if you have ideas for *simple* provisioning 2014-09-18 07:29:50 ssh might be ok 2014-09-18 07:30:01 if you go for nc, you might as well use http 2014-09-18 07:30:08 ncopa: it should pop up if I execute "apk upgrade" when it isn't installed right..? 2014-09-18 07:30:12 and do the signature checking etc on other layer 2014-09-18 07:30:33 Amnesia: it should not get reinstalled if you do apk upgrade 2014-09-18 07:30:42 installed* 2014-09-18 07:31:06 I didn't have wayland installed before the upgrade 2014-09-18 07:31:14 weird 2014-09-18 07:31:20 yeah.. 2014-09-18 07:31:26 and if you do apk ugprade again? 2014-09-18 07:31:30 the same happens for llvm-libs, mesa-gbm and libepoxy 2014-09-18 07:31:42 somethign migh pull those in 2014-09-18 07:31:53 well, that's why I checked using apk info -r 2014-09-18 07:31:54 libepoxy is needed by recent xorg server 2014-09-18 07:32:14 llvm-libs, is likely some xorg driver (ati?) 2014-09-18 07:32:18 apk info -r libepoxy doesn't return anything 2014-09-18 07:32:45 ncopa-desktop:~$ apk info -r libepoxy 2014-09-18 07:32:45 libepoxy-1.2-r0 is required by: 2014-09-18 07:32:45 xorg-server-xephyr-1.16.0-r2 2014-09-18 07:32:45 xorg-server-1.16.0-r2 2014-09-18 07:33:41 which repositories are you using? 2014-09-18 07:33:48 I'm only using edge now 2014-09-18 07:33:59 i have edge/main and edge/testing 2014-09-18 07:34:45 k I'll enable testing as well 2014-09-18 07:35:03 same results.. 2014-09-18 07:35:05 shouldnt make any difference really 2014-09-18 07:35:29 apk version -U 2014-09-18 07:35:36 <^7heo_> ncopa: I'm thinking nc is simpler than http 2014-09-18 07:35:57 <^7heo_> ncopa: on the server: cat proviscript.sh | nc -q0 -lp 4242 2014-09-18 07:36:10 <^7heo_> ncopa: on the client: nc 4242 | sh 2014-09-18 07:36:12 <^7heo_> done. 2014-09-18 07:36:16 ncopa: that segfaults:-).. 2014-09-18 07:36:32 hm 2014-09-18 07:36:43 try apk fix apk-tools 2014-09-18 07:36:45 <^7heo_> all you need is a script, or an alias that does replace "provisionning " with "nc 4242 | sh 2014-09-18 07:36:57 <^7heo_> and works. 2014-09-18 07:37:14 Amnesia: try just: apk fix 2014-09-18 07:37:23 something is broken somehow 2014-09-18 07:37:30 same result 2014-09-18 07:37:39 after apk fix has been executed * 2014-09-18 07:38:10 x86? 2014-09-18 07:38:15 or x86_64 2014-09-18 07:38:21 x86_64 2014-09-18 07:38:38 'apk version' 2014-09-18 07:38:41 does it segfault? 2014-09-18 07:38:49 nope 2014-09-18 07:39:08 apk version -U only segfaults 2014-09-18 07:39:24 ha 2014-09-18 07:39:26 indeed 2014-09-18 07:39:31 thats a bug 2014-09-18 07:49:53 <^7heo_> damn, how to pipe to exec... 2014-09-18 07:51:32 <^7heo_> for now, the "provision" alias or script is: 2014-09-18 07:51:45 <^7heo_> tmpfile=$(mktemp); nc $ip 4242 > $tmpfile; chmod a+x $tmpfile; $tmpfile; rm $tmpfile; 2014-09-18 07:52:22 <^7heo_> on the provision seed, I still have "cat $script | nc -q0 -lp 4242" 2014-09-18 08:49:30 <^7heo_> on the other hand... 2014-09-18 08:49:41 <^7heo_> before "setup-alpine", the machine doesn't even have an IP... 2014-09-18 08:50:20 <^7heo_> so that would have to be changed too... 2014-09-18 08:51:15 its just to: udhcpcd eth0 2014-09-18 08:51:37 <^7heo_> oh 2014-09-18 08:51:40 <^7heo_> so it can have an IP. 2014-09-18 08:51:49 <^7heo_> that might work then. 2014-09-18 08:53:40 <^7heo_> ncopa: what do you think of my solution? 2014-09-18 08:54:24 <^7heo_> I would personally rewrite it as: tmpfile=$(mktemp); nc $ip 4242 > $tmpfile && chmod a+x $tmpfile && $tmpfile; rm $tmpfile 2014-09-18 08:54:47 <^7heo_> so it would avoid errors if the connection fails. 2014-09-18 08:56:47 <^7heo_> however, I would use the port 4810 rather than 4242 2014-09-18 08:57:10 <^7heo_> https://en.wikipedia.org/wiki/Mont_Blanc 2014-09-18 08:57:22 <^7heo_> more exactly: https://en.wikipedia.org/wiki/Mont_Blanc#Elevation 2014-09-18 08:57:55 i think I'd prefer http 2014-09-18 08:58:00 <^7heo_> meh =/ 2014-09-18 08:58:08 <^7heo_> it's harder to setup... 2014-09-18 08:58:11 <^7heo_> MUCH harder. 2014-09-18 08:58:14 <^7heo_> I mean 2014-09-18 08:58:26 <^7heo_> "cat file | nc -q0 -np 4810" 2014-09-18 08:58:30 <^7heo_> it's kinda easy. 2014-09-18 08:58:37 <^7heo_> no packages to install, nothing else to do. 2014-09-18 08:58:48 and what happens after first client is done? 2014-09-18 08:59:01 <^7heo_> "press up, press enter" 2014-09-18 08:59:04 or if 2 clients tries to connect at the same time? 2014-09-18 08:59:25 <^7heo_> hmm, for that one okay. 2014-09-18 08:59:39 httpd -p 4810 -h . 2014-09-18 08:59:48 <^7heo_> yeah 2014-09-18 08:59:58 <^7heo_> does alpine contain httpd by default? 2014-09-18 08:59:59 wget -q -O - http://IP/file | sh 2014-09-18 09:00:02 yup 2014-09-18 09:00:06 and wget 2014-09-18 09:00:10 <^7heo_> the | sh is a bad idea btw. 2014-09-18 09:00:10 no packages to install 2014-09-18 09:00:19 <^7heo_> if people are using anything else than sh script to do the provisionning 2014-09-18 09:00:22 <^7heo_> it'll break. 2014-09-18 09:00:32 true 2014-09-18 09:00:42 <^7heo_> when you explicitely specify sh, you won't get the shebang interpreted. 2014-09-18 09:00:48 <^7heo_> that's why I came to my mktemp solution. 2014-09-18 09:00:56 <^7heo_> as I found no way to execute /dev/stdin 2014-09-18 09:22:13 <^7heo_> also, the problem is that the httpd has to be on the host that serves the provisionning file. 2014-09-18 09:23:04 <^7heo_> it would be best to have something that can be called directly from the command line 2014-09-18 09:23:07 <^7heo_> and http isn't ideal. 2014-09-18 09:23:15 <^7heo_> especially since it's VERY often used already. 2014-09-18 09:23:23 <^7heo_> (but changing the port solves that problem) 2014-09-18 09:26:19 <^7heo_> or wait 2014-09-18 09:26:43 <^7heo_> nah putting the server on the installer side would require the user to know the IP. 2014-09-18 09:26:58 <^7heo_> on the other hand, this can be retreived from the script. 2014-09-18 09:27:13 <^7heo_> you would just need to invoke "provision" 2014-09-18 09:27:35 <^7heo_> that would listen on port 4810, and print the command line you have to type on your computer to "POST" the script to it. 2014-09-18 09:27:49 <^7heo_> you just have to copy it, done. 2014-09-18 09:27:59 <^7heo_> what do you think about that? 2014-09-18 09:28:33 <^7heo_> problem is, you can't configure a machine to serve provisionning and just provision 2014-09-18 09:29:51 <^7heo_> or, more complicated (more sucky also), "provision" could connect to a host to notify that host to "push" the provision file. 2014-09-18 09:30:03 <^7heo_> that way it can be automatized 2014-09-18 09:30:25 <^7heo_> but at what cost... 2014-09-18 09:30:53 <^7heo_> the provision script would basically create an httpd, plus potentially another notification connection. 2014-09-18 09:31:13 <^7heo_> that would have the advantage of not requiring any specific software package on the other computer 2014-09-18 09:31:24 <^7heo_> (provided that wget or curl are there) 2014-09-18 09:31:39 <^7heo_> and allowing to automatically provision stuff from a pre-configured server. 2014-09-18 09:32:06 <^7heo_> 'cause if I would have to summary the provisionning requirements, I would write: 2014-09-18 09:32:08 i still get the feeling that netboot with pxe is the way to go 2014-09-18 09:32:22 solves many of the problems you describe 2014-09-18 09:32:30 <^7heo_> 1. one command to type on the provisionning host if it's a one time setup 2014-09-18 09:32:44 <^7heo_> 2. a pre-configured host has to be possible too, in the event of many setups 2014-09-18 09:33:07 <^7heo_> 3. configuring more than one client at the same time 2014-09-18 09:33:11 <^7heo_> must be possible 2014-09-18 09:33:23 <^7heo_> it solves everything but point 1. 2014-09-18 09:33:30 <^7heo_> and that's my problem with PXE. 2014-09-18 09:33:34 <^7heo_> it does 66% of the job. 2014-09-18 09:33:36 <^7heo_> not 100%. 2014-09-18 09:33:41 <^7heo_> I'm after a solution that works with 100%. 2014-09-18 09:34:58 <^7heo_> granted, PXE is something that has to be possible too. 2014-09-18 09:35:21 <^7heo_> because if you have a high volume of clients, you don't even want to login and type a command on each. 2014-09-18 09:35:28 <^7heo_> so it has to be possible. But not a requirement. 2014-09-18 09:35:52 i think we need to provide a full pxe solution too 2014-09-18 09:35:57 its on my list 2014-09-18 09:36:09 <^7heo_> I agree with that. 2014-09-18 09:36:15 <^7heo_> it's not solving my problem, but I agree. 2014-09-18 09:36:35 <^7heo_> and since my problem would work in tandem with PXE, I'd like to discuss that before you implement the PXE 2014-09-18 09:36:45 <^7heo_> or I risk to have much more problems implementing my stuff after. 2014-09-18 09:36:49 so once thats ineplace, you only need a one-command to set up the tftp server and dhcp config 2014-09-18 09:37:06 yeah 2014-09-18 09:42:53 <^7heo_> yeah 2014-09-18 09:42:57 <^7heo_> should we open a ticket? 2014-09-18 09:43:09 <^7heo_> so I can describe the requirements, for various use cases? 2014-09-18 09:43:21 <^7heo_> because provisionning 2k clients is one thing 2014-09-18 09:43:28 <^7heo_> but provisionning 10 is different 2014-09-18 09:43:36 ok 2014-09-18 09:43:44 <^7heo_> and you might want to go a no-config on the provisionning host for 10 clients 2014-09-18 09:43:53 <^7heo_> and you can afford to type 10 times 2014-09-18 09:43:54 <^7heo_> root 2014-09-18 09:43:58 <^7heo_> provision 2014-09-18 09:44:11 <^7heo_> it's faster than setting up PXE. 2014-09-18 09:44:22 <^7heo_> (in average situations) 2014-09-18 09:54:56 <^7heo_> Shall I assign the ticket to you or me? 2014-09-18 09:55:58 you can assign me if you want 2014-09-18 09:56:03 <^7heo_> okay 2014-09-18 09:56:10 <^7heo_> I estimate the work to 30h, including research 2014-09-18 09:56:32 <^7heo_> feel free to correct me, if you think that it's way off. 2014-09-18 09:59:20 i wonder, to make file search of packages possible, maybe the buildbot can be extended to tar -tvf file > dir/$packagename-$version, then generate a manifest from it and make it available per mirror? 2014-09-18 10:00:07 the file would contain: "$packagename:" then just the contents per pkg, it would be sufficient 2014-09-18 10:02:03 the manifest script would only have to cat the files > MANIFEST, so it wouldnt be too inefficient 2014-09-18 10:02:12 need maintenance 2014-09-18 10:02:15 clean up old 2014-09-18 10:02:42 yea 2014-09-18 10:05:10 <^7heo_> damn 2014-09-18 10:05:12 <^7heo_> how to edit a ticket? 2014-09-18 10:05:37 <^7heo_> http://bugs.alpinelinux.org/issues/3372 2014-09-18 10:05:40 <^7heo_> I would need to edit that. 2014-09-18 10:05:49 <^7heo_> because of the format, mostly. 2014-09-18 10:10:37 <^7heo_> (sorry for not checking the preview) 2014-09-18 10:45:04 <^7heo_> okay I can easily provision stuff with nc 2014-09-18 10:45:09 <^7heo_> it works great. 2014-09-18 10:53:01 <^7heo_> ncopa: there's no dhclient on the iso here... 2014-09-18 10:53:10 <^7heo_> using 3.0.4 2014-09-18 10:56:31 <^7heo_> which is a bit annoying when I have to provision the host... 2014-09-18 10:56:35 <^7heo_> but just a bit :P 2014-09-18 10:59:02 <^7heo_> when you get used to it it's fine. 2014-09-18 10:59:17 <^7heo_> and the installation of alpine takes around 1 minute with the provision script. 2014-09-18 10:59:21 <^7heo_> from the nc command to poweroff. 2014-09-18 11:06:41 udhcpcd 2014-09-18 11:06:51 udhcpc 2014-09-18 11:10:17 <^7heo_> hmm 2014-09-18 11:10:27 <^7heo_> I will look into that one next provisionning. Thanks. 2014-09-18 13:06:35 dalias: i got in touch with rsmarples, the author of openrc and dhcpcd-ui 2014-09-18 13:06:46 he is interested in feedback on dhcpcd-ui 2014-09-18 13:06:51 i think he want make itusable 2014-09-18 13:07:10 adhdcpd 2014-09-18 15:53:40 ncopa, nice 2014-09-18 15:54:05 ncopa, my biggest feedback is that the ui only allows adding AP's, not selecting an already-added one to use 2014-09-18 15:54:37 and doesn't have any way to edit their priorities for autoselection or most per-AP properties you might want to set 2014-09-18 15:54:59 i realize a lot of this is on the wpa-supplicant side, but imo a frontend still needs to handle it to be very useful 2014-09-18 16:37:51 <^7heo_> ncopa: udhcpc isn't working because the iso boots with eth0 down. 2014-09-18 16:38:11 <^7heo_> if I issue "ip link set eth0 up" before, it works. 2014-09-18 16:38:25 <^7heo_> should I submit a bug for that? 2014-09-18 18:07:49 <^7heo_> how to change the autoboot delay? 2014-09-18 18:10:31 <^7heo_> okay, it was multiplied by 10, that's why I didn't find it. 2014-09-18 18:10:36 <^7heo_> good evening everyone 2014-09-19 09:12:39 <^7heo_> ncopa: just tried the cherry keyboard with another computer (in case it was something ELSE than the keyboard having problems), it is stalled the same way. 2014-09-19 09:12:55 <^7heo_> So conclusion, the cherry keyboards don't work with alpine linux. 2014-09-19 09:13:09 <^7heo_> neither the installer nor the system. 2014-09-19 09:34:22 <^7heo_> there are ONLY 5 developers in alpine?! 2014-09-19 09:34:31 <^7heo_> (based on the number of rsa keys) 2014-09-19 09:35:04 <^7heo_> or are multiple keys used by multiple user groups? 2014-09-19 09:35:13 <^7heo_> s/multiple/different/ 2014-09-19 10:07:52 which rsa keys? 2014-09-19 10:11:25 https://www.openhub.net/p?ref=homepage&q=alpine+linux 2014-09-19 10:11:31 ^7heo: where do you get your statistics from? 2014-09-19 10:11:55 <^7heo_> /etc/apk/keys 2014-09-19 10:12:04 <^7heo_> I have no clue what they are for, yet 2014-09-19 10:12:06 <^7heo_> I didn't read any doc' 2014-09-19 10:13:23 checking the signature for the packages, particularily when installing packages 2014-09-19 10:13:51 most distributions do not have many such keys regardless of the number of devs involved in the project 2014-09-19 10:14:16 <^7heo_> So yeah, they are the apk keys for verification of the package source and integrity. 2014-09-19 10:14:42 <^7heo_> isn't it one key per dev? 2014-09-19 10:15:28 no 2014-09-19 10:16:02 im not sure what the policy is 2014-09-19 10:16:10 but i think arm also introduced a new key 2014-09-19 10:16:22 most distributions do the signing on the build servers rather than the individual packet maintainer doing the signing 2014-09-19 10:16:51 just read gitlog of alpine-keys and you should find out. 2014-09-19 10:17:14 it may even be wiser doing the building on a specific build server with a tightly controlled invironment to ensure a predictable result 2014-09-19 10:17:51 especially on a distro with many developers on the team 2014-09-19 10:21:36 <^7heo_> makes sense 2014-09-19 10:21:47 <^7heo_> (about the singing on the compilation server) 2014-09-19 10:22:29 <^7heo_> R0113: I think that building on a specific server should be irrelevant 2014-09-19 10:22:38 <^7heo_> what's relevant is the environment. 2014-09-19 10:27:14 definitely, but if you are not carefull with your housekeeping you will end up with packages that depends on wrong things, or packages that are built against libraries that do not exist in the dependency list in the package 2014-09-19 10:27:43 having a build server will make keeping a consistent build envirionment a non issue 2014-09-19 10:28:27 since I am not an alpine dev I do not know how they might organize the build workflow, I am just talking from my own experience 2014-09-19 10:29:24 as with any decent distribution you should be able to build anything by yourself in anyway you may particularily care 2014-09-19 10:30:14 hence having access to a specific build server should not be a technical recuirement, but it makes the craft of assembling a working distribution easier 2014-09-19 10:31:13 <^7heo_> personally 2014-09-19 10:31:16 ^7heo: but I have no inside knowledge here, just speculating 2014-09-19 10:31:51 <^7heo_> I would like build environments to be easily deployable using distcc over temporary docker deployments 2014-09-19 10:32:10 <^7heo_> so anyone could take the scripts, build them, and distribute them. 2014-09-19 10:34:34 lxc works nicely, dont see the need for docker. 2014-09-19 10:35:07 <^7heo_> I have very little experience with each 2014-09-19 10:35:09 <^7heo_> I know mostly Xen 2014-09-19 10:35:21 <^7heo_> and it's much much heavier 2014-09-19 10:35:45 lxc is light and works on any decent recent kernel. 2014-09-19 10:37:19 clandmeter: docker has benefits mostly for people that don't understand much of os installation etc. like, if you'd not be able to reliably manage / setup the OS and dependencies, then it's best to use a tool that does it 2014-09-19 10:37:48 "hmmm, i installed the container and port 7000 come online, must've worked" 2014-09-19 10:37:58 this is the use case from what i was shown ;) 2014-09-19 10:39:46 ^7heo: then even if you are using docker it is like having a build server even if you are not allocating any particular pieces of silicone for the job 2014-09-19 10:40:46 ^7heo: sorry misspelled your nick 2014-09-19 10:41:11 ^7heo_: sorry for spelling like a bicycling bonobo on acid 2014-09-19 10:46:06 <^7heo_> R0113: that's fine, I'm also on acids 2014-09-19 11:29:57 <^7heo_> ncopa: webkit does INDEED take a freaking long time to build... 2014-09-19 11:30:11 <^7heo_> maybe it'd be faster if I would use my machine rather than kvm 2014-09-19 11:30:42 ^7heo_: ccache. first thing to do :) 2014-09-19 11:39:46 <^7heo_> yeah 2014-09-19 11:39:55 <^7heo_> but I have a very very fast processor 2014-09-19 11:40:10 <^7heo_> so I think that using KVM might be a good thing, but not without the proper parameters 2014-09-19 11:40:18 <^7heo_> (by default, the machines are rather low on ram and all) 2014-09-19 12:40:33 <^7heo_> ncopa: any idea when webkitgtk+2.0 would be available? 2014-09-19 12:45:34 <^7heo_> https://bugs.alpinelinux.org/issues/3374 2014-09-19 12:46:08 dont have time today sorry 2014-09-19 12:46:19 it is not very hard 2014-09-19 12:46:24 just timeconsuming 2014-09-19 12:48:15 https://bugs.alpinelinux.org/issues/3374 2014-09-19 12:48:41 i think it might be due to usbkbd 2014-09-19 12:48:46 i disabled it recently 2014-09-19 12:48:52 fabled had similar issue i think 2014-09-19 12:49:10 try blacklist usbkbd module 2014-09-19 12:49:20 it won't help 2014-09-19 12:49:27 unless hard disk install 2014-09-19 12:49:38 try ssh in, rmmod usbkbd, and replug keyboard 2014-09-19 12:50:58 <^7heo_> ncopa: I'll try, thanks for the info. 2014-09-19 12:51:06 <^7heo_> ncopa: I wanted to build webkit in the week end. 2014-09-19 12:51:15 <^7heo_> problem is, i have no experience with abuild at all. 2014-09-19 12:51:28 <^7heo_> not yet. 2014-09-19 12:51:35 <^7heo_> it'll change during the week end I guess ;) 2014-09-19 12:54:06 ok, let me see what i can do - but i doubt i'll be able to fix webkitgtk2 before weekend 2014-09-19 12:54:31 <^7heo_> I tried to do something but here I can only do it in a KVM 2014-09-19 12:54:35 <^7heo_> and I have no time to configure my KVM machine 2014-09-19 12:54:48 <^7heo_> I have 20 tickets to solve in something like 30 minutes 2014-09-19 12:54:51 <^7heo_> v_v 2014-09-19 13:24:10 people always think ccache somehow won't be the most beneficial tuning there is because "something", but the linker caching alone is always faster than the fastest cpu. unless ~/.ccache is put on a floppy disk 2014-09-19 13:25:54 that must be a large floppy 2014-09-19 13:26:04 a floppy array 2014-09-19 13:28:52 clandmeter: they make good music ;) 2014-09-19 13:31:30 <^7heo_> huhu 2014-09-19 13:51:52 bummer, btrfs doesnt support swap files... 2014-09-19 14:47:32 hmm, wasn't there a nocow option for that? 2014-09-19 14:48:23 it was a planned feature few years ago when I was first poking it 2014-09-19 14:52:22 ah, seems it's more complicated due the linux internal api being tied to single dev stuff, with possibility of using the new apis merged with swap-on-nfs but that's not been done yet 2014-09-19 14:56:30 CcxCZ: in ltsp they offer something called NBDSWAP, network block device swap, I do not remember if this is a ltsp-specific protocol or something generic, but regardless of that, it works well, it is optimized for the swap task wich nfs is not 2014-09-19 14:56:30 setting up the nbd-server is very easy 2014-09-19 14:59:00 yeah, swap-on-n{fs,bd} share the same api. You can already use loopback devices with btrfs, but you're taking a performance hit 2014-09-19 15:02:45 Re: swap on btrfs....can't use it, even with nocow option....Can't you do another partition dedicated for swap? 2014-09-19 15:03:23 Or, do you even need swap? Is moreram an option? 2014-09-19 15:03:26 or just buy more ream :P 2014-09-19 15:03:35 lol :) 2014-09-19 15:03:36 hehe ninja'd 2014-09-19 15:04:16 its ooming on c++ build 2014-09-19 15:04:32 I always look at it like this: If you NEED to use swap, then you NEED to buy more ram ;) 2014-09-19 15:04:35 ahh, the webkit plague? 2014-09-19 15:04:44 yep 2014-09-19 15:04:49 ginjachris: only if it's a regular thing 2014-09-19 15:05:28 how much does it eat nowadays? disabling all the debugging flags helps 2014-09-19 15:05:32 aye, that was my meaning...long day, probably i'm not making much sense 2014-09-19 15:05:59 i got 16g on lxc host, but its not enough :| 2014-09-19 15:06:09 lol, I think C++ is your problem then 2014-09-19 15:06:41 I've never had more than 4 GB :( 2014-09-19 15:06:53 Yours is bigger than mine :S 2014-09-19 15:07:44 whoopsie. I wonder why nobody still made a linker that'd have it's working symbolds in DBM instead of memory 2014-09-19 15:07:52 assuming it crashes on linking 2014-09-19 15:08:16 yes 8 cores with ht (16 total) and 16GB should be enough i would guess, but it isnt. 2014-09-19 15:08:29 probably some other containers are using memory. 2014-09-19 15:08:50 tried to make swap device with losetup? 2014-09-19 15:08:57 not yet 2014-09-19 15:09:12 i read thats a workeround 2014-09-19 15:09:15 but im too lazy 2014-09-19 15:09:20 @clandmeter Are you also simultaneously calculating the orbit of all known universal objects to the nearest micrometer? That's an awesome system! 2014-09-19 15:09:58 i just dont like btrfs... 2014-09-19 15:10:36 would like to have zfs back, guess its not yet fixed for musl. 2014-09-19 15:10:59 I experimented with btrfs a while back, was not impressed with its crash resistance.....invalid superblock alot. No such problem on raid 2014-09-19 15:11:50 --no-keep-memory 2014-09-19 15:11:52 ld normally optimizes for speed over memory usage by caching the symbol tables of input files in memory. This option tells ld to instead optimize for memory usage, by rereading the symbol tables as necessary. This may be required if ld runs out of memory space while linking a large executable. 2014-09-19 15:12:04 I wonder if you could make the build use that ↑ 2014-09-19 15:15:22 trying it 2014-09-19 15:18:09 thats a gcc configure flag? 2014-09-19 15:23:20 ld option http://ftp.gnu.org/old-gnu/Manuals/ld-2.9.1/html_mono/ld.html 2014-09-19 15:23:38 eh, the page is pretty old, dunno if it still exists 2014-09-19 15:24:43 seems so, found it in current `info ld` 2014-09-19 15:25:18 so it should go into LDFLAGS I guess 2014-09-22 07:39:31 <^7heo_> moin 2014-09-22 11:24:32 <^7heo_> is it normal that the alpine iso boots with all interfaces "DOWN"? 2014-09-22 11:43:37 ^7heo_: you mean first boot? 2014-09-22 12:24:24 i tried alpine in virtualbox and i couldnt get mc to work properly 2014-09-22 12:24:49 strace shows that it cant find the linux terminfo 2014-09-22 12:25:17 so i tried to install ncurses-terminfo (i guess slang lib used by mc can use the same) 2014-09-22 12:25:43 but the linux terminfo is a broken symlink in the ncurses-terminfo package 2014-09-22 12:26:19 alpine-3.0.4-x86_64 2014-09-22 12:27:31 (the issue i'm trying to solve is that arrow and function keys dont work in mc) 2014-09-22 12:41:25 fyi if i fix the symlink (ln -s linux-vt linux) then mc works as expected 2014-09-22 12:42:44 nsz: hi 2014-09-22 12:42:51 nice to have you around 2014-09-22 12:43:28 so the 'linux' symlink is missing 2014-09-22 12:44:04 nsz: what happens if you apk add ncurses 2014-09-22 12:45:05 ah i think it is ncurses-base 2014-09-22 12:50:00 nsz: i think the terminfo you want is in 'ncurses-base', but i wonder if mc should depend on it or if slang should 2014-09-22 12:50:22 i suppose ncurses-terminfo should depend on ncurses-base 2014-09-22 12:51:53 ncurses-terminfo should depend on ncurses-base if that's where the linux symlink target is 2014-09-22 12:52:00 yeah 2014-09-22 12:52:09 thats what i figured too 2014-09-22 12:52:15 i think the terminfo lookup is generic slang logic 2014-09-22 12:52:30 it checks a lot of different places 2014-09-22 12:52:51 probably ncurses-base is enough for it 2014-09-22 12:53:16 is the terminfo required to make slang useful at all? 2014-09-22 12:53:29 i mean, should it be an optional dependency 2014-09-22 12:53:42 (which we dont really have support for) 2014-09-22 12:53:49 well the function keys and arrow keys are broken without it on the linux console 2014-09-22 12:53:53 as far as i can see 2014-09-22 12:55:41 ncurses-base seems to work fine here 2014-09-22 12:56:21 maybe it should be named as terminfo-base or something if it only has terminfo files 2014-09-22 13:02:32 seems like no other package depends on ncurses-base 2014-09-22 13:02:39 so should be fairly safe to rename it 2014-09-22 13:02:53 maybe ncurses-terminfo-base? 2014-09-22 13:03:39 i will not rename it in the stable branch 2014-09-22 13:06:43 nsz: feel free to hang around in #alpine-devel too ;) 2014-09-22 13:09:22 ok 2014-09-22 13:10:34 <^7heo_> clandmeter: yep. 2014-09-22 13:11:45 ^7heo_: no interface should be up afaik 2014-09-22 13:12:01 ^7heo_: yes, its normal that all interfaces are down 2014-09-22 13:12:36 <^7heo_> how is udhcpd supposed to grab any lease if no interface is up? 2014-09-22 13:12:48 you can quickly setup the basics with: setup-alpine -q 2014-09-22 13:13:23 ^7heo: i think you need to do ip link set dev eth0 up first 2014-09-22 13:13:24 hm 2014-09-22 13:13:28 <^7heo_> ip link set eth0 up; udhcp eth0 2014-09-22 13:13:34 yeah 2014-09-22 13:13:35 <^7heo_> not necessarily longer to write :P 2014-09-22 13:13:48 <^7heo_> actually 2014-09-22 13:13:53 <^7heo_> udh 2014-09-22 13:13:54 <^7heo_> :P 2014-09-22 13:16:10 <^7heo_> but in the end 2014-09-22 13:16:18 <^7heo_> is there a specific reason to have all interfaces down? 2014-09-22 13:16:26 <^7heo_> even if a cable is plugged? 2014-09-22 13:16:51 <^7heo_> 'cause it wouldn't be such a problem if the eth0 iface was up if a link is detected, and that udhcpc tries to get a lease, would it? 2014-09-22 14:25:15 Does anybody run a nfsv3 server on alpine? I have # ex. OPTS_RPC_MOUNTD="-p 32767" 2014-09-22 14:26:16 OPTS_RPC_MOUNTD="-N 4 -V 3 -N 2" ... in the /etc/conf.d/nfs and "get a daemon.notice rpc.mountd[32445]: authenticated mount request from ..." on the server and a "mount.nfs: No error information" on the client 2014-09-22 14:26:54 And nfsmount on grml tells me "error 521" (not unknown error) 2014-09-22 14:30:16 s/"get a /get a "/ 2014-09-22 14:37:37 oh 2014-09-22 14:37:55 do you need idmap? 2014-09-22 14:38:11 i think the problem might be lack of support for dnotify in kernel 2014-09-22 14:38:26 i have been thinking of it recently 2014-09-22 14:39:15 there was a thread about it: http://www.spinics.net/lists/linux-nfs/msg42003.html 2014-09-22 14:39:38 i would actually prefer to see nfs-utils add support for inotify than enable dnotify in the kernel 2014-09-22 14:43:01 <^7heo_> that'd be better yes. 2014-09-22 14:43:17 nfs3 should not need idmapd 2014-09-22 14:43:57 but my fileserver has turned into a diva with one of the last updates, nfs starts by restarting a number of nfs related things 2014-09-22 14:43:57 <^7heo_> also 2014-09-22 14:49:17 <^7heo_> is it normal that firefox doesn't work anymore? 2014-09-22 14:49:27 <^7heo_> when I install the package firefox, I get nothing displayed. 2014-09-22 14:49:36 <^7heo_> and ofc, since it's firefox, nothing is sent back to setderr 2014-09-22 14:49:51 <^7heo_> no log no nothing 2014-09-22 14:55:29 Is there another ff instance still running? 2014-09-22 14:57:07 I also don't think idmapd is needed for v3 2014-09-22 15:03:25 is it possible to update the Alpine system with a command? 2014-09-22 15:04:44 <^7heo_> like apk update && apk upgrade? 2014-09-22 15:06:30 yes, thanks 2014-09-22 15:24:00 <^7heo_> when I execute firefox 29, on a fresh install, I get no display 2014-09-22 15:24:05 <^7heo_> when I gdb it I get two warnings 2014-09-22 15:26:50 <^7heo_> Starting program: /usr/lib/firefox-29.0.1/firefox 2014-09-22 15:26:50 <^7heo_> warning: Cannot call inferior functions, Linux kernel PaX protection forbids return to non-executable pages! 2014-09-22 15:26:53 <^7heo_> warning: no loadable sections found in added symbol-file system-supplied DSO at 0x6b605734b000 2014-09-22 15:26:59 <^7heo_> basically, I never get a display 2014-09-22 15:31:58 <^7heo_> is firefox working with a new install for anyone? 2014-09-22 15:34:33 firefox on edge was working a few weeks ago, but i cant confirm it right now 2014-09-22 15:35:43 <^7heo_> pnutzh4x0r: I'm not on edge, I'm on 3.0 2014-09-22 15:35:53 <^7heo_> and it's not working at all. 2014-09-22 15:35:56 <^7heo_> But lemme check edge, yeah 2014-09-22 15:35:58 <^7heo_> good idea 2014-09-22 15:58:24 <^7heo_> edge has the same problem 2014-09-22 15:59:40 <^7heo_> I don't get it, firefox's working at home... 2014-09-22 16:50:23 <^7heo_> nah really, can't get firefox working... 2014-09-22 16:50:36 <^7heo_> should I build it myself or what? 2014-09-22 16:59:30 <^7heo_> does anyone have a rope? 2014-09-22 17:11:09 i remember fabled mention something about firefox issues this afternoon 2014-09-22 17:11:17 dunno if its related tho 2014-09-22 17:14:31 <^7heo_> it's not. 2014-09-22 17:14:41 <^7heo_> my problem's been a pebcak since the beginning 2014-09-22 17:14:51 <^7heo_> I'm going to do extra hours at work 2014-09-22 17:14:59 <^7heo_> 'cause I'm SO freaking ashamed 2014-09-22 17:16:25 <^7heo_> you know dwm? 2014-09-22 17:16:31 <^7heo_> it's having 9 workspaces right? 2014-09-22 17:16:40 <^7heo_> there's a rule to start firefox in the workspace 9, right? 2014-09-22 17:16:53 <^7heo_> I'm working on a kiosk setup, the bar is hidden. 2014-09-22 17:17:02 <^7heo_> I don't need to say more, I guess. 2014-09-22 17:17:56 ^7heo_: question, in a kiosk setup, how do you make firefox always open the page, even if it crashed 2014-09-22 17:18:07 i never found how to do that 2014-09-22 17:18:13 <^7heo_> darkfader: I guess you have to use a plugin for that, right? 2014-09-22 17:18:20 so if the client crashed you'd get profile recovery 2014-09-22 17:18:21 :( 2014-09-22 17:18:33 or not profile, session restore? 2014-09-22 17:18:36 <^7heo_> oh ofc you delete the profile every time 2014-09-22 17:18:39 it sucked anyway 2014-09-22 17:18:42 ah! 2014-09-22 17:18:47 <^7heo_> it's like 2014-09-22 17:18:51 <^7heo_> firefox url 2014-09-22 17:18:53 i guess that was the idiot thing i missed 2014-09-22 17:18:56 <^7heo_> in .xinitrc 2014-09-22 17:19:02 yeah and -kiosk 2014-09-22 17:19:08 <^7heo_> it's existing? 2014-09-22 17:19:11 but i think i didn't delete the profile 2014-09-22 17:19:13 <^7heo_> without extension? 2014-09-22 17:19:37 ^7heo_: idk, it was 2 years ago when i worked on that 2014-09-22 17:19:57 but so far i never met someone who used firefox in such a setup :) 2014-09-22 17:19:59 <^7heo_> okay, I have to try it. 2014-09-22 17:20:23 <^7heo_> I will as soon as I get that setup working again (deploying while I type - it's a deployment script) 2014-09-22 17:20:29 :) 2014-09-22 17:20:36 <^7heo_> and I somehow wish that dwm would be built into packages sometimes... 2014-09-22 17:20:44 <^7heo_> it doesn't make sense most of the time 2014-09-22 17:20:54 <^7heo_> but for my 194th deployment... 2014-09-22 17:21:07 <^7heo_> fetching build-base each time is a tad annoying 2014-09-22 17:22:32 <^7heo_> darkfader: was it -kiosk or --kiosk? 2014-09-22 17:27:49 doesn't show up in -h output 2014-09-22 17:27:54 i'll try to find it 2014-09-22 17:28:33 googles say it's only via addons 2014-09-22 17:28:45 idk. please disregard my memory, it must be wrong 2014-09-22 17:29:29 i don't have access to the sources from $back_then 2014-09-22 17:31:26 <^7heo_> darkfader: chrome has a kiosk mode tho 2014-09-22 17:31:31 <^7heo_> and many browsers do 2014-09-22 17:31:35 <^7heo_> but firefox sucks. 2014-09-22 17:31:44 <^7heo_> chrome's google toy 2014-09-22 17:31:53 <^7heo_> and all the other browsers crash/don't render 2014-09-22 17:31:57 <^7heo_> I love the web ™ 2014-09-22 17:36:32 <^7heo_> darkfader: I found that -new-window helps 2014-09-22 17:36:59 <^7heo_> even if it crashes. 2014-09-22 17:37:04 <^7heo_> so now I only need two things. 2014-09-22 17:37:11 <^7heo_> 1. find a way to fullscreen it. 2014-09-22 17:37:47 <^7heo_> 2. find a way to disable the f**ing warning when you connect with http://user:passwd@host/ 2014-09-22 17:46:02 wah i hate firefox again now. seems in 2008 it had a commandline option for it 2014-09-22 17:46:05 lol 2014-09-22 17:46:31 <^7heo_> yeah firefox sucks. 2014-09-22 17:46:34 <^7heo_> really. 2014-09-22 17:46:40 <^7heo_> but the web is made for it. 2014-09-22 17:47:01 the last comment in there is a bit interesting: https://support.mozilla.org/en-US/questions/966532 2014-09-22 17:47:10 (i was varying search terms in hope) 2014-09-22 17:47:26 of course that would need a profile 2014-09-22 17:47:35 i give up for today 2014-09-22 17:47:48 <^7heo_> darkfader: you're setting up a kiosk setup? 2014-09-22 17:48:13 firefox -height 1200 -width 1600 2014-09-22 17:48:15 ??? 2014-09-22 17:48:21 ^7heo_: no i did it in 2011 2014-09-22 17:48:26 i was just gonna help you :) 2014-09-22 17:48:47 but i gave up now -> going home 2014-09-22 17:48:48 <^7heo_> okay, many thanks :) 2014-09-22 23:47:15 could anyone help me building an apk package from source? source does compile using abuild but no apk is created. 2014-09-23 08:59:42 Good morning! How are all the humans that are responsible for the best OS ever? 2014-09-23 09:12:57 <^7heo_> 'can't answer, I'm not one of 'em. 2014-09-23 09:13:03 <^7heo_> But good morning to you. 2014-09-23 09:13:05 good morning :-) 2014-09-23 09:16:52 @^7heo, sure you are! You're here helping out :¬) 2014-09-23 09:17:04 and good morning :) 2014-09-23 09:32:52 <^7heo_> ginjachris: yeah depepends... 2014-09-23 09:39:20 I've been messing with run from ram (bootable usb) recently - only previously used sys installs to hdd - and I must say it's damn awesome, very well done! 'lbu' works like a dream :D 2014-09-23 09:59:46 :) 2014-09-23 10:15:53 Hi there, what is an apkovl file? 2014-09-23 10:23:25 fuzzy, apkovl is a tarball of configuration files that have been changed. it is used to restore running config during boot in run-from-tmpfs setups. 2014-09-23 14:01:47 <^7heo_> firefox's so bad... 2014-09-23 14:01:58 <^7heo_> broken software ftw... 2014-09-23 14:04:10 could be worse, could be IE :p 2014-09-23 15:06:44 ^7heo_: i'm working on chromium now, so you can help :) 2014-09-23 15:39:54 quick query on dropbear.....it doesn't appear to accept SCP connections, unless I also have 'openssh' package installed...for example from WinSCP, error = "Cannot execute SCP to start transfer. Please make sure that SCP is installed on the server and path to it is included in PATH. You may also try SFTP instead of SCP. Command failed with return code 127." 2014-09-23 15:40:45 I guess there must be some other function that openssh pulls in that's not compiled with dropbear by default? (<--Ich bin nicht ein Dev, so this could be way out) 2014-09-23 15:41:37 or maybe I should rephrase that, the SCP connection is accepted- this is just standard SSH at this point SFAIK - but you get the error if you try to do SCP file transfer 2014-09-23 15:43:42 Anyway, no biggy, just thought I'd mention it 2014-09-23 15:46:43 drobpear doesnt support scp/sftp 2014-09-23 15:46:50 http://en.wikipedia.org/wiki/Dropbear_(software) 2014-09-23 15:46:54 dropbear even 2014-09-23 15:47:18 it does support scp, for sure 2014-09-23 15:47:44 https://matt.ucc.asn.au/dropbear/CHANGES 2014-09-23 15:48:40 i think maybe it needs to be compiled with scp support, otherwise you need openssh package too? 2014-09-23 15:49:43 from the changes "Fix "make install" when scp is included since it has no manpage" (amongst others) 2014-09-23 15:50:01 it works, just would be better if I didn't need openssh package too 2014-09-23 15:52:59 anyway, I must depart Ladies(?) and Gentlemen. Have a great evening! 2014-09-23 15:53:57 dropbear has a scp client, but requires sftp-server from openssh on the server side 2014-09-23 15:54:04 at least from what i can read 2014-09-23 15:54:23 awwww that sucks :( thanks for setting me straight! 2014-09-23 15:54:53 I'll try and do more reading in the future! 2014-09-23 15:55:04 it's fine, it confused me too 2014-09-23 15:55:34 right, must dash! Things to see, people to do! 2014-09-23 15:55:36 Bye all 2014-09-24 08:56:47 Bonjour all! 2014-09-24 08:57:18 Salut! 2014-09-24 08:58:02 <^7heo_> moin. 2014-09-24 12:44:21 ¡hello! 2014-09-24 12:51:24 Got told off by an IT evangelist the other day. Apparently I had SYN'd 2014-09-24 12:51:37 (I can hear the groans from here) 2014-09-24 12:52:53 even if it took a while to grow, it still hurt. 2014-09-24 12:54:12 :) 2014-09-24 12:58:48 has the extlinux grub handler suddenly learned to respect the partitons and not use hd(0) when it should be hd(0,0) ? 2014-09-24 12:59:02 i just rolled out an update and the configs aren't broken any more 2014-09-24 12:59:35 or maybe there was no kernel update... :) i need a weekend for tinkering 2014-09-24 12:59:35 <^7heo_> darkfader: it can be a bug solved by another bug. 2014-09-24 12:59:38 <^7heo_> beware of those. 2014-09-24 13:01:54 ^7heo_: :) would be ok right now, anything that crashes me at least in an initrd is a good thing 2014-09-24 13:02:14 almost as good as updating and rebooting 2014-09-24 13:03:09 I'm sorry, but here's some more terrible IT jokes. Prepare yourselves! 2014-09-24 13:03:35 "I was once living very actively - playing football, tennis, participating into car races. Sometimes I would play poker and pool. But later somebody stole my PC and that was it..." 2014-09-24 13:03:56 "I would love to change the world, but they won’t give me the source code" 2014-09-24 13:04:26 ginjachris: did you see the guy that put his sequenced dna on github? 2014-09-24 13:04:38 go on.... 2014-09-24 13:04:44 and literally 15 minutes later someone sent him a pull rquest changing one chromosome 2014-09-24 13:04:51 and the comment on the patch was 2014-09-24 13:04:56 "enjoy your new life" 2014-09-24 13:05:09 lol 2014-09-24 13:06:25 just found this one online: "Sales representative of Samsung computer department tries to persuade an officer of a Russian city municipality to purchase computers: - Don't worry, computers will never substitute officers: no company has yet created a machine that would do nothing." 2014-09-24 13:06:42 i think I'll leave it at that 2014-09-24 13:08:14 <^7heo_> ginjachris: that last one was good :) 2014-09-24 13:31:47 alright here's one final one for the day: "A guy prepares to demonstrate voice control of a computer. - Silence, please, I will now demonstrate how easy voice control works. A man from the back of the room: - Format: C. Enter." 2014-09-24 13:46:14 Why is it that rivers flow upstream to downstream and wind blows from upwind to downwind, yet most of the people I speak to about network traffic flow state it goes from downstream to upstream? Am I mental or have they got it wrong? (i try to avoid the term and go with 'external' and 'internal' and so forth) 2014-09-24 14:11:20 <^7heo_> why has postscript been abandonned? 2014-09-24 14:20:41 ginjachris: maybe because network goes both directions? 2014-09-24 15:13:47 @@ncopa, agreed, but why must they defy all logic? I mean, it makes sense for traffic to go 'outbound' (towards the internet) or 'inbound' (from internet towards local network) but upstream and downstream? I think it's weird. 2014-09-24 15:14:26 Anyway, I wouldn't lose any sleep over it :¬) I'll just keep asking them to clarify to avoid simpletons like me getting confused! 2014-09-24 15:34:43 I'm afraid I must depart :( Thank you and good night! 2014-09-24 16:02:11 https://securityblog.redhat.com/2014/09/24/bash-specially-crafted-environment-variables-code-injection-attack/ 2014-09-24 16:02:16 uh... seems bad 2014-09-24 16:32:58 which init system does alpine linux use? 2014-09-24 16:33:21 openrc 2014-09-24 16:33:43 oh, nice 2014-09-24 16:42:35 nmeum: any particular reason to not go with systemd? i am curious because I don't really know the technical differences 2014-09-24 16:46:08 besides the politics, systemd doesnt support musl 2014-09-24 16:50:14 uselessd does :3 2014-09-24 16:56:47 pnutzh4x0r: interesting, i am not familiar with variences in standard c libraries 2014-09-24 16:56:56 what are the properties of musl that make it a necessity? 2014-09-24 17:04:56 it's very, very small, and doesn't have GNU bloat 2014-09-24 17:05:23 ahills: interesting, anyone get a chance to audit the code? 2014-09-24 17:05:33 for musl? 2014-09-24 17:05:35 yah 2014-09-24 17:05:40 yes, everyone has a chance, it's open source 2014-09-24 17:05:42 have at it 2014-09-24 17:06:12 ahills: i know everyone has the potential 2014-09-24 17:06:26 i mean has anyone done an actual formal or informal audit and documented their findings 2014-09-24 17:06:28 what sort of audit are you looking for? 2014-09-24 17:06:42 ansi compliance? 2014-09-24 17:06:44 no 2014-09-24 17:06:57 more for robustness, and security review 2014-09-24 17:08:21 http://www.etalabs.net/compare_libcs.html has a brief summary 2014-09-24 17:09:23 ahills: thanks 2014-09-24 17:09:42 systmkor: also check in #musl 2014-09-24 17:09:50 for sure 2014-09-24 17:10:45 ahills pnutzh4x0r to build alpine from source on a non alpine box, i am guessing build apk-tools first then build aports. 2014-09-24 17:11:12 I'm not an expert on Alpine, sorry 2014-09-24 17:12:05 ahills: no problem 2014-09-24 17:49:42 ahills: do you know how to use apk to build locally within a directory 2014-09-24 17:49:47 not to a system wide directory 2014-09-24 17:53:55 ahh found it i think 2014-09-24 18:51:00 ahills: okay so what would be the proper way of generating the necessary apk files for building files 2014-09-24 18:51:57 ahills: would chrooting and runnin setup-alpine a starting point? 2014-09-24 20:00:03 systmkor1: if you want to start building, you will need alpine-sdk 2014-09-24 20:00:11 m'kay 2014-09-24 20:00:19 yah checking out alpine-chroot at the moment 2014-09-24 20:00:28 will check out alpine-sdk 2014-09-24 20:00:51 its not like you have apk-tools and start building whatever you want. 2014-09-24 20:03:00 well yah 2014-09-24 20:03:12 so only alpine-sdk i am finding is under abuild 2014-09-24 20:03:17 which would require apk 2014-09-24 20:05:10 systmkor1: you can use a chroot, or could setup something virtual like lxc. 2014-09-24 20:05:27 true true 2014-09-24 20:05:27 apk can create the chroot for you if you like. 2014-09-24 20:05:38 well i am running on arch at the moment 2014-09-24 20:05:53 apline-chroot 2014-09-24 20:06:02 is breaking on me when i run it from arch 2014-09-24 20:06:10 i think i can fix it 2014-09-24 20:06:10 http://wiki.alpinelinux.org/wiki/Setting_up_Alpine_in_a_chroot 2014-09-24 20:08:01 yah i am using an existing script that does that 2014-09-24 20:08:20 clandmeter: http://git.alpinelinux.org/cgit/fab/alpine-chroot/ 2014-09-24 20:08:40 yes i thought so, when i saw you mention alpine-chroot 2014-09-24 20:08:49 its not working? 2014-09-24 20:09:22 yah not for me 2014-09-24 20:09:38 one sec lemme paste the stderr 2014-09-24 20:09:41 one sec 2014-09-24 20:11:52 clandmeter: https://gist.github.com/e2c4f3bf436ca2785391 2014-09-24 20:12:55 seems like on line #89 in the alpine-chroot script 2014-09-24 20:13:37 is using tar to unpack the .apk file 2014-09-24 20:14:56 lol 2014-09-24 20:15:08 the .apk file is an xml file of 404 not found 2014-09-24 20:18:33 clandmeter: I assume the URL used for curl to get the apk static tools needs to be updated 2014-09-24 23:41:15 hi 2014-09-24 23:42:45 I'm trying to install 3.0.4, it fails after formating the drive, complaints about unsatisfiable constraints, ".setup-apkrepos missing" 2014-09-25 00:12:06 hmm... it worked after a reboot. only thing I changed was using dhcp and not starting the install with nonworking internet access 2014-09-25 03:42:31 I'm sure you all are aware of the bash bug - http://ftp.gnu.org/gnu/bash/bash-4.3-patches/bash43-025 2014-09-25 06:39:09 zacts: our repos have been updated yesterday 2014-09-25 06:39:42 upgraded from 2.7 edge to 3.0.4 \o/ 2014-09-25 07:22:20 congrats kocka 2014-09-25 07:22:21 :) 2014-09-25 07:23:02 ok, make sure you guys have bash-4.3.25-r1 2014-09-25 07:23:07 the -r1 is important 2014-09-25 07:23:12 patch was not applied in -r0 2014-09-25 07:23:15 (my mistake) 2014-09-25 07:23:28 or make sure just not to install bash :) 2014-09-25 07:23:35 thats even better :) 2014-09-25 07:23:48 i am pretty glad we dont install it by default 2014-09-25 07:52:28 mksh \o/ 2014-09-25 07:54:06 why does windows always want to install a new device when you plug it into another usb port. :| 2014-09-25 07:58:23 hey all. I want to try this linux-module https://github.com/kierank/sxs-linux but don´t know what package i have to install to compile modules. 2014-09-25 07:59:39 mirac: hi, 2014-09-25 07:59:49 apk add build-base linux-grsec-dev 2014-09-25 08:01:02 looks relatively simple 2014-09-25 08:01:51 ncopa: hi, about patch in bash, i heared it doesn't work :/ 2014-09-25 08:02:21 k0r10n: it should be fixed in bash-4.3.25-r1 2014-09-25 08:02:22 the -r1 2014-09-25 08:02:34 but it looks like there are more issues in bash 2014-09-25 08:03:17 env X='() { (a)=>\' bash -c "echo echo vuln"; [[ "$(cat echo)" == "vuln" ]] && echo "still vulnerable :(" 2014-09-25 08:03:35 will write a file in $PWD 2014-09-25 08:06:31 ncopa: thx! 2014-09-25 09:03:38 G'day all! 2014-09-25 09:06:19 morning 2014-09-25 09:52:54 StarWarsFan|afk_: t looks like I can mount nfs on my desktop 2014-09-25 09:53:01 ncopa-desktop:~$ sudo mount 10.65.65.131:/nfs /mnt 2014-09-25 09:53:01 ncopa-desktop:~$ ls /mnt/ 2014-09-25 09:53:01 hello 2014-09-25 09:53:09 where nfs server is ubuntu 2014-09-25 09:53:49 i think nfs works 2014-09-25 10:36:33 ncopa: interesting 2014-09-25 11:55:47 ncopa, mailinglist said nfs v3 is broke with my patch 2014-09-25 12:02:53 i safw it 2014-09-25 12:02:56 also 2014-09-25 12:03:04 for idmap you need dnotify in kernel 2014-09-25 12:03:18 whcich we dont have enabled 2014-09-25 12:03:29 so we need either enable dnotify in kernel 2014-09-25 12:03:42 or add inotify support for nfs.utils 2014-09-25 18:16:02 ncopa: should I just go ahead and integrate the get_static_apk() function you referenced from the lxc containers 2014-09-25 18:17:12 ncopa: I knew it wasn't a great/longterm patch but I couldn't find anything describing a guideline for the layout of a alpine package mirror 2014-09-25 18:17:46 so that you could programatically always get the correct static apk without doing something hacky 2014-09-25 18:19:19 Who is working on the ARM builds of Alpine? 2014-09-25 19:56:55 okay, so I have a tiny request for two extra modules to be turned on in the vanilla and grsec kernels, what's the best place for that request? and how? a simple diff that patches the kernel config, or a full-blown git patch that does config+apkbuild? 2014-09-25 21:09:25 hey all so I asked the sytemd community to see if systemd works on musl and apparently someone has been messing around with that 2014-09-25 21:09:27 http://lists.freedesktop.org/archives/systemd-devel/2014-September/023177.html 2014-09-25 21:15:54 looks like there is some push back 2014-09-25 21:15:57 http://lists.freedesktop.org/archives/systemd-devel/2014-September/023212.html 2014-09-25 21:17:03 Obviously 2014-09-25 21:17:17 pnutzh4x0r: my thought would be that 2014-09-25 21:17:24 if it is minimal changes 2014-09-25 21:17:28 The systemd upstream have stated many times that they are not interested in portability patches 2014-09-25 21:17:34 possible of maintaining a shim between standard glib and musl version 2014-09-25 21:17:59 openfbtd: no reason not to keep pushing 2014-09-25 21:18:34 pnutzh4x0r and openfbtd I am trying to find where in the repository the Kconfig files are stored 2014-09-25 21:18:35 Interestingly enough, that guy's submission is not a lot 2014-09-25 21:18:45 to build the kernel from source correctly 2014-09-25 21:18:53 I am trying to get alpine linux working on ARM devices 2014-09-25 21:19:13 Dunno, I don't use Alpine currently 2014-09-25 21:19:19 ah 2014-09-25 21:19:44 seems like the systemd guys are basically saying turn musl into glibc 2014-09-25 21:19:50 Yep. 2014-09-25 21:20:04 systmkor: http://git.alpinelinux.org/cgit/aports/tree/main/linux-grsec 2014-09-25 21:20:22 ahhhh 2014-09-25 21:20:26 i think u are looking for the kernerlconfig? 2014-09-25 21:20:32 yah 2014-09-25 21:20:34 i guess 2014-09-25 21:20:34 ok 2014-09-25 21:21:04 well so my end goal is to be able to automate the process of updates that you guys make or are made to the kernel that you are using 2014-09-25 21:21:33 and then use bitbake from the yocto project as a means of building the kernel and such 2014-09-25 21:21:49 to whatever architecture I may need in the future 2014-09-25 21:22:01 if you guys know of a better method, I am all ears 2014-09-25 21:23:12 there's uselessd 2014-09-25 21:23:16 that works on musl 2014-09-25 21:23:34 and can do anything systemd does, just not the worst of the bullshit 2014-09-25 21:24:18 darkfader: cool, I will check that out 2014-09-25 21:24:57 darkfader: any thoughts on properly programatically building alpine for various architecutres? 2014-09-25 21:52:43 is there a recommended way of generating a dependency graph with apk packages 2014-09-25 22:25:58 systmkor: we have an image for rpi (arm) 2014-09-25 22:26:25 cool were at? and where can I repeat that process from source? 2014-09-25 22:26:42 clandmeter: thanks 2014-09-25 22:30:35 clandmeter: once I can get that working were would it be best for me to document on the wiki? 2014-09-25 22:38:54 http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz 2014-09-25 22:39:00 thats the latest image 2014-09-25 22:39:14 i think fabled wanted to creatae a new one today, 2014-09-25 22:39:20 maybe tomorrow 2014-09-25 22:39:31 our repo has arm 2014-09-25 22:39:59 bedtime 2014-09-25 23:18:23 btw what's the status of alpine's init system search? 2014-09-25 23:18:51 I remember that the project wasn't entirely satisfied with openrc 2014-09-25 23:21:41 I cant seem to get open-vm-tools installed. its listed in the online package search thing: http://alpinelinux.org/apk/main/x86_64/open-vm-tools 2014-09-25 23:23:25 this is on a fresh 3.0.4 install 2014-09-26 01:49:27 bye 2014-09-26 06:54:33 systmkor1: the alpine-chroot script is fabian_a's i dont have any opinion on it 2014-09-26 07:14:26 jzono1: i think i fixed open-vm-tools after v3.0 release, could you please file a bug and i'll cherry-pick the fixes for v3.0 2014-09-26 07:25:03 what's wrong with openrc? 2014-09-26 08:03:00 Morning all! 2014-09-26 08:07:11 ahills: almost nothing. and i'd rather fix openrc than replace it with some kind of nerd smelly ueberdaemon 2014-09-26 08:20:34 yay please keep opernrc, it's awesome! 2014-09-26 08:25:11 # apk update 2014-09-26 08:25:12 fetch http://dl-5.alpinelinux.org/alpine/v3.0/main/x86_64/APKINDEX.tar.gz 2014-09-26 08:25:12 WARNING: Ignoring /v3.0/main/x86_64/APKINDEX.tar.gz: No such file or directory 2014-09-26 08:25:18 is this warning normal? 2014-09-26 08:25:46 nsz: what do you have in /etc/apk/packages? 2014-09-26 08:25:54 err 2014-09-26 08:26:00 repositories 2014-09-26 08:26:41 ah there is this line: 2014-09-26 08:26:42 /v3.0/main 2014-09-26 08:26:59 that sounds not correct 2014-09-26 08:27:09 ok thanks 2014-09-26 08:27:10 we have a script to setup the repo's 2014-09-26 08:27:32 setup-apkrepos 2014-09-26 08:33:12 setup-apkrepos does not remove the /v3.0/main line 2014-09-26 08:33:20 i'll just remove it manually 2014-09-26 08:36:47 hello 2014-09-26 08:37:11 did anyone try to install any kind of caldav or carddav on alpine? 2014-09-26 09:15:02 hello, i have a firewall running alpine-linux which i didnt install myself, now iam a bit lost as to how to update bash 2014-09-26 09:15:24 i did a apk-update and an apk add --update bash 2014-09-26 09:15:44 but the version installes is 4.2.037-r0 2014-09-26 09:19:39 the version is alpinelinux 2.4 2014-09-26 09:55:51 Blashyrkh1: http://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2014-09-26 09:57:41 basically, v2.4 is out of support 2014-09-26 10:51:05 ncopa: is there a lua alternative for psutil python module? 2014-09-26 10:52:09 kocka: not that i know 2014-09-26 12:59:32 ahills, Didn't say anything was wrong with it, I just remember that it wasn't considered the final choice for alpine at some point 2014-09-26 13:13:17 you see the problem with openrc if you grep sleep */*.initd 2014-09-26 13:13:34 http://sprunge.us/bcgJ 2014-09-26 13:13:54 if you need to do that, something is wrong 2014-09-26 13:16:38 I've been pushing for equivalent to runit's "check" script in openrc 2014-09-26 13:17:17 that should eliminate most of those that actually have reason to be there 2014-09-26 13:17:32 which I don't think many have actually 2014-09-26 13:18:42 eg. the ones for stopping are completely silly, there's required functionality in s-s-d already 2014-09-26 13:19:12 ncopa, haha oh wow 2014-09-26 13:21:13 clearly many packagers don't know of ewaitfile :-( 2014-09-26 13:22:08 that's package QA issue rather than opernc issue 2014-09-26 13:22:51 Yeah, my service manager isn't much different from openrs, and the only sleeps I have are in the eudev and systemd-eudevd scripts 2014-09-26 13:23:01 And those are probably not necessary 2014-09-26 13:25:02 openrc* 2014-09-26 13:28:45 http://gentooexperimental.org/~patrick/weblog/archives/2013-10.html#e2013-10-12T05_54_04.txt and posts below it 2014-09-26 14:46:08 <^7heo> moin @ 2014-09-26 14:57:30 CcxCZ: re sleep, exactly 2014-09-26 15:05:44 ncopa: problem with complete elimination of sleep is that we can't really get notified of many things in a portable manner, we have to poll for eg. socket being listened on 2014-09-26 15:06:40 it will require some support from the daemon to do it right 2014-09-26 15:06:58 dailas had a nice idea on ewontfix 2014-09-26 15:07:12 readyfd 2014-09-26 15:07:18 I'd rather not make everyone change their daemons 2014-09-26 15:07:35 you cannot solve it properly other way really 2014-09-26 15:07:59 ok gotta go 2014-09-26 15:08:19 there are ways under linux where that could be implemented in a notification-based way though. seccomp should let you know when specified syscalls are called without much overhead for example 2014-09-26 15:08:38 <^7heo> any clue how to enable webgl in firefox 29, by any chance? 2014-09-26 15:09:17 CcxCZ: http://ewontfix.com/15/ 2014-09-26 15:09:29 gotta run 2014-09-26 15:09:30 see u 2014-09-26 15:09:41 bye :-) 2014-09-26 15:34:19 <^7heo> okay, what's the group for being able to use udev? 2014-09-26 15:35:45 <^7heo> ncopa: https://forums.grsecurity.net/viewtopic.php?f=3&t=4012 < I found that. Is there a way to make gl direct access work without sudo rights atm? 2014-09-26 15:52:15 i can't even take as big a breath as answer about sleep and systemd 2014-09-26 15:53:08 pretty close to buying a new car because dead battery. but then you buy a 80's lada with a jet engine tied to it 2014-09-26 15:53:33 or alternatively, a super modern ICBM but without any payload compartment 2014-09-26 16:55:43 ncopa: sorry for bashism and mistakes in aport :/ 2014-09-26 17:56:20 huh. open-vmware-tools and open-vmware-tools-grsec builds just fine, and works just fine when I build it myself from aports, why can't I apk install it? 2014-09-26 17:56:30 (from regular repos, I mean) 2014-09-26 18:44:38 ncopa: okay no problem, just curious, should I just email him directly? 2014-09-26 21:08:42 O.o 2014-09-26 21:09:00 so targetcli looks broken :/ 2014-09-26 21:14:04 it's missing a dependency, and the build process fails, at the same time abuild continues and puts together an incomplete package 2014-09-26 21:16:06 shouldn't the abuild fail and not create a package? 2014-09-26 21:27:34 <^7heo> k0r10n: please be careful about that, people don't want another Arschlinux. 2014-09-26 21:31:04 does zfs work in alpine yet? 2014-09-26 21:35:33 hm i just tried to mount an ext3 partition running from alpine livecd and it failed unless i explicitly specified the fs type 2014-09-26 21:36:25 it seems busybox mount checks /proc/filesystems and tries the names there without 'nodev' 2014-09-26 21:37:18 and there are only squashfs, vfat and iso9660 there 2014-09-26 22:36:50 ext3 should be without nodev too .... 2014-09-26 22:37:22 If you mean the case no-module-loaded-before, its because it would have to implement super magic checking otherwise. 2014-09-26 22:43:37 i just wanted to use the alpine install cd as a livecd (or rather liveusb) and mount an existing disk 2014-09-26 22:43:51 and in that case ext3 is not in the /proc/filesystems list 2014-09-26 22:44:02 nor ext2 2014-09-26 22:44:45 nsz: Yes, you'll have to modules=ext3 as kernel methinks. 2014-09-26 22:44:49 *kernel param 2014-09-26 22:45:07 (or just -t ext3 to poke the autoloader) 2014-09-26 22:45:42 mount -t ext3 works 2014-09-26 22:46:30 ok so ext3 is in a kernel module 2014-09-26 22:59:52 well, yeah 2014-09-26 22:59:59 if its not in /proc/filesystem its not loaded 2014-09-26 23:00:31 and mount syscall wants exact fs name. the easy way is to simply iterate all whats in /proc/filesystems 2014-09-26 23:01:04 but if you wanted it to work without -t when no mod is loaded, you'd have to magic(1) the superblock 2014-09-27 02:10:36 http://judecnelson.blogspot.fi/2014/09/systemd-biggest-fallacies.html 2014-09-27 02:11:07 echoes a lot of my personal thoughts 2014-09-27 02:15:49 but also memory management, thread management, thread synchronization, file descriptor management, low-level IPC, architecture-specific memory models 2014-09-27 02:16:23 minor nitpick, rather unfortunate statement wrt systemd as theres no such thing as arch specific memory models on unix, but anyway ... 2014-09-27 03:01:28 "You cannot run journald without systemd, and cannot run systemd without journald. " 2014-09-27 03:01:30 that's wrong 2014-09-27 03:53:56 lol yeah it's wrong, I had systemd running without journald because journald kept crashing 2014-09-27 03:54:08 but it only logged its output to journald, so I couldn't figure out why 2014-09-27 07:22:57 ahills: ah, that's pretty hilarious 2014-09-27 16:08:20 yeah, it was really funny after I was done migrating my mission-critical systems away from systemd 2014-09-27 17:18:36 (arch?) 2014-09-27 18:32:13 (yes) 2014-09-27 18:32:31 (now they're all alpine) 2014-09-27 18:35:12 \o/ 2014-09-27 22:47:19 hi 2014-09-27 22:48:15 I just installed an alpine. ran setup-alpine then change the repo to include edge/main and edge/testing in order to try openvswtich and qemu... 2014-09-27 22:48:41 after lbu vi and reboot, openvswitch id giging an error on boot. 2014-09-27 22:48:54 *s/id/is/ 2014-09-27 22:50:36 and I cannot add_br then. 2014-09-27 22:52:30 Are somewhere some tips about openvswitch with alpine? I cannot find any at the time of writing. I'm taking a lot of notes and will release my experience if not already published somewhere but If I can avoid to reinvent the wheel, it would be appreciated... 2014-09-27 23:01:50 Jean-Scotch: hi 2014-09-27 23:02:27 what is the content of your repo file? 2014-09-27 23:02:40 did you mix stable and edge repo's? 2014-09-27 23:03:55 no. 3 lines: 2014-09-27 23:03:55 /media/usb/apks 2014-09-27 23:03:55 http://dl-5.alpinelinux.org/alpine/edge/main 2014-09-27 23:03:55 http://dl-5.alpinelinux.org/alpine/edge/testing 2014-09-27 23:04:23 I did made an apk upgrade then an apk upgrade before doing anything else 2014-09-27 23:04:34 ok so now you are running edge. 2014-09-27 23:04:47 the setup is from almpinelinux-mini from an usb key 2014-09-27 23:04:50 i guess most pkgs got updated when you switched to edge. 2014-09-27 23:05:04 my goal is to prepare a master key for al batch of server 2014-09-27 23:05:25 they will be used to launch qemu from CEPH RDB 2014-09-27 23:05:44 so the key doesnt boot now? 2014-09-27 23:05:50 It was from alpine -mini that's it 14 packages to start from ;) 2014-09-27 23:06:08 it is booting but giving an errot on the console about openvswitch 2014-09-27 23:06:31 ok what does it say? 2014-09-27 23:06:56 the error is unfortunately not ont the screen anymore and the end of the boot process so I'm trying to grab it while booting on my laptop ... 2014-09-27 23:07:04 i must admit, i never tried ovs. 2014-09-27 23:07:15 can you restart ovs and see if you get the error? 2014-09-27 23:07:18 (I'm wrting this from my desktop) 2014-09-27 23:09:00 when (re)starting all 3 init scripts by hand, I have no erros... (/etc/init.d/script restart I mean) 2014-09-27 23:09:44 do you see anything in dmesg? 2014-09-27 23:09:47 anything strange 2014-09-27 23:09:58 i remember ovs needs a kernel module 2014-09-27 23:10:10 which is now mainline if i remember correctly 2014-09-27 23:10:35 you are right. it was not loaded while booting but was when restarting scripts by hand 2014-09-27 23:10:58 i know a few ppl were trying ovs before. 2014-09-27 23:11:19 some send patches to fix things. 2014-09-27 23:11:24 not sure what the current status is. 2014-09-27 23:11:47 if its the module, then we probably need to fix init.d scripts of ovs. 2014-09-27 23:12:00 or you can add the module to /etc/modules 2014-09-27 23:12:15 I will try that now 2014-09-27 23:14:05 again, it spells that openvswitch postinstall returned with exit status code 1 2014-09-27 23:14:41 then its probably the script thats faulty 2014-09-27 23:15:22 ok. I will try to have a look at the install scripts... 2014-09-27 23:15:24 :( 2014-09-27 23:15:40 ah 2014-09-27 23:15:43 you are running from ram 2014-09-27 23:15:46 I want to find a fix. So I will devote some time on this topic 2014-09-27 23:15:58 yes, I want to run it from RAM 2014-09-27 23:16:04 when its started 2014-09-27 23:16:15 can you lbu_commit and reboot 2014-09-27 23:16:26 i think the db is not restored 2014-09-27 23:17:05 i guess you have setup lbu.conf to store apkovl on usb? 2014-09-27 23:17:23 seems reasonable. because when I did setup the DB by hand, I found the module in lsmod but not previous of that 2014-09-27 23:17:28 yes 2014-09-27 23:17:34 lbu to usb 2014-09-27 23:17:42 the db is in etc 2014-09-27 23:17:49 so lbu should back it up 2014-09-27 23:18:02 and restore it on boot 2014-09-27 23:18:11 ok. should do the lbu ci after setting up everything 2014-09-27 23:18:52 seems normal wy og doing thing. still have to get used to alpine (I'comming from debian world) 2014-09-27 23:18:56 yes, thats desirable on tmpfs install :) 2014-09-27 23:21:11 I guess it will be tomorrow assignment. it's becoming late over here. thanks for the pointers. I guess those will help 2014-09-27 23:21:29 yes me too 2014-09-27 23:21:45 time for bed. 2014-09-27 23:22:21 morning 2014-09-27 23:22:35 morning 2014-09-27 23:22:39 lol good night 2014-09-27 23:22:45 well 1:22am here 2014-09-27 23:22:46 :D 2014-09-27 23:22:51 evidently morning 2014-09-28 03:48:53 clandmeter: could you send that rpi image again? 2014-09-28 08:31:51 Hi, I would like to use Alpine as my desktop distribution, yet I wonder if it's a good idea? 2014-09-28 08:32:15 I do 2014-09-28 08:32:17 it's great 2014-09-28 08:32:27 but your needs are likely different than mine 2014-09-28 08:32:44 well I mostly use open source software 2014-09-28 08:32:50 except matlab 2014-09-28 08:33:35 and I fond that there are not ghc (haskell compiler) and Octave (like matlab) packages 2014-09-28 08:33:44 which are really big to compile 2014-09-28 08:33:50 what worries me 2014-09-28 08:33:51 no, but building your own APKs is pretty easy 2014-09-28 08:35:17 I saw APKs, they are much like PKGBUILDS on Archlinux which I use now 2014-09-28 08:35:44 good, so you already have some experience 2014-09-28 08:36:18 but still compiling ghc for example would take much more than compiling linux kernel 2014-09-28 08:36:50 and considering my relative old Core2 CPU 2014-09-28 08:37:32 and one thing more: Is Alpine planning to switch to SystemD? 2014-09-28 08:37:51 if they do, I'll be hunting for another distro 2014-09-28 08:37:58 me too 2014-09-28 08:38:10 that's why I'm laving Arch 2014-09-28 08:38:19 that would be pretty hard to do, considering systemd doesn't even compile on alpine 2014-09-28 08:38:19 systemd causing critical failures is the reason I left... 2014-09-28 08:38:33 good 2014-09-28 08:39:37 asmir: if you craft an APKBUILD I'll build ghc for you for x86_64 2014-09-28 08:40:23 http://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package and http://wiki.alpinelinux.org/wiki/APKBUILD_Reference 2014-09-28 08:40:39 that's great 2014-09-28 08:40:48 ghc is in stable already 2014-09-28 08:40:53 oh 2014-09-28 08:40:55 nevermind then :) 2014-09-28 08:41:13 even better 2014-09-28 08:41:23 I can take care for Octave 2014-09-28 08:41:25 apk search ghc returns no results 2014-09-28 08:41:32 or 2014-09-28 08:41:33 hmmm 2014-09-28 08:41:50 I do remember "just" installing it when I had it on my 2.7 box 2014-09-28 08:42:26 or am I remembering wrong? 2014-09-28 08:42:38 http://wiki.alpinelinux.org/wiki/Porting_GHC_to_Alpine 2014-09-28 08:42:52 last updated may 2013 2014-09-28 08:43:16 I may try this myself tomorrow 2014-09-28 08:44:27 hmm yeah, I must have forgotten about haskell about the same time I left arch for alpine, I guess 2014-09-28 08:44:59 personally I'd probably keep a chroot of another distro around if I were to use alpine as a desktop distro 2014-09-28 08:45:08 is there anything you miss from arch? 2014-09-28 08:45:19 occasionally a package here and there 2014-09-28 08:45:31 especially with python, since python's package manager is so poor 2014-09-28 08:46:43 I see, well I don't use python and most of my programs are lightweight 2014-09-28 08:46:52 I don't even use a DE 2014-09-28 08:47:18 so I guess compiling them and even maintaing with ccache would make no problem 2014-09-28 08:47:42 using the suckless.org tools in alpine is easy 2014-09-28 08:48:01 great, I use some 2014-09-28 08:50:46 and x86 is i386? 2014-09-28 09:33:49 K, I've made a bootable usb and a backup 2014-09-28 09:33:54 thanks for your help 2014-09-28 09:36:00 ahills: http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz 2014-09-28 17:52:39 clandmeter: thanks! 2014-09-29 02:40:02 ahem 2014-09-29 02:40:12 alpine's initrd ignores init=? 2014-09-29 02:40:19 That's shitty 2014-09-29 02:43:48 In any case. uselessd builds and works nicely on alpine 2014-09-29 02:43:48 Well 2014-09-29 02:43:50 It boots 2014-09-29 02:56:41 bleh :-p 2014-09-29 02:56:46 how... useless :) 2014-09-29 02:57:16 Yeah, it is in the context of alpine 2014-09-29 02:57:40 It's just that I wanted to test it on musl and Alpine is one of the easiest ways to do that 2014-09-29 04:21:26 do you know if alpine will migrate to systemd too? 2014-09-29 04:23:14 phanton: not anytime soon 2014-09-29 04:24:30 God is god 2014-09-29 04:24:36 good news 2014-09-29 04:24:39 tks mate 2014-09-29 04:25:16 np 2014-09-29 04:25:42 we have so many distros today that use it 2014-09-29 04:25:55 sorry for my vision, but I can see few sold their souls 2014-09-29 04:26:11 slackware, gentoo and alpine 2014-09-29 04:26:24 and of course, the bsd* ones 2014-09-29 04:28:25 pnutzh4x0r: tks anyway 2014-09-29 04:28:37 voidlinux and crux too 2014-09-29 04:30:57 ohh yeah, just tried to say, betwenn 400 distros, 390 did this mistake 2014-09-29 05:00:53 Some distros didn't actually nail systemd to the core of their distro 2014-09-29 05:01:13 It's just the default is all 2014-09-29 05:01:42 There are distros like debian, where EVERYTHING assumes that the core packages exist 2014-09-29 05:02:01 But there is also, for example, Arch, that gives you the tools to easily get rid of systemd 2014-09-29 05:02:07 Or most core packages really 2014-09-29 05:02:31 this is really good, thanks for share this 2014-09-29 05:02:36 I have an Arch install with suckless tools instead of gnu ones. Had to fix a few things, but it boots and works 2014-09-29 05:02:52 Needless to say, it doesn't have systemd either 2014-09-29 05:03:05 this is the better to do 2014-09-29 05:03:32 I don't see any advantage in systemd, and I think 90% fell the same 2014-09-29 05:03:45 I've actually found alpine harder to tinker with than Arch so far :D 2014-09-29 05:03:53 openrc is a bit tricky to remove 2014-09-29 05:04:08 yeah, I've used to use arch too 2014-09-29 05:04:13 also, archbang 2014-09-29 05:04:33 but force everybody use systemd is a mistake 2014-09-29 05:04:33 I see advantages of systemd. Lots of them. I just think they are not worth the downsides 2014-09-29 05:04:44 Arch doesn't force you. 2014-09-29 05:04:52 I can't see by your side, to me it is a trojan 2014-09-29 05:04:59 worth than windows nt 2014-09-29 05:05:32 also, can't see any respect for the linux community, by it developer 2014-09-29 05:05:39 Proper service management is a huge plus. I just don't like how they did it. 2014-09-29 05:05:42 he don't give a sh¨&¨(* 2014-09-29 05:05:56 dbus everything! 2014-09-29 05:05:58 Why? 2014-09-29 05:06:31 But anyway, if you want to use something else, there are plenty of options currently 2014-09-29 05:06:35 Including Alpine. 2014-09-29 05:06:53 I'm currently using alpine as a base for my docker vms 2014-09-29 05:06:57 And loving it 2014-09-29 05:07:12 The road to hell is paved with good intentions 2014-09-29 05:07:38 incredible that we install alpine in 3 or 5 minutes 2014-09-29 05:07:59 I actually install Arch in roughly the same ammount of time these days :) 2014-09-29 05:08:23 cool 2014-09-29 05:08:25 Although currently I have five basic VMs that I just clone 2014-09-29 05:08:42 No reason to reinstall the systems that are running on metal 2014-09-29 05:08:52 I loved arch too, to me was like a slack evolution 2014-09-29 05:09:15 yeah, arch is very easy, and alpine follows the same philosophy 2014-09-29 05:09:37 It's the only distro that allows me to build what I want relatively easily with rolling release model and binary packages. And it's glibc-based 2014-09-29 05:09:50 what I more line in arch linux 2014-09-29 05:09:53 is that tool 2014-09-29 05:10:04 network tool that even in shell 2014-09-29 05:10:18 I can choose one of the my essid and use it 2014-09-29 05:10:23 As much as I like minimalism and proper software, the lack of nvidia drivers in Alpine makes it useless to me on my home desktop :< 2014-09-29 05:10:41 sorry, but I like this tool, I love stay on shell, I just need a wm like openbox to use 2014-09-29 05:10:42 i.e. the proprietary module doesn't build. 2014-09-29 05:11:21 Heh. I've written a simple wifi discovery tool to use with wpa_supplicant 2014-09-29 05:11:34 nvidia ... right now I'm reading some arguments about 2014-09-29 05:11:38 You only need a scanner that will show you the networks basically 2014-09-29 05:11:45 in shell? 2014-09-29 05:11:49 yep 2014-09-29 05:12:04 I should probably release the damn thing 2014-09-29 05:12:05 cool 2014-09-29 05:12:15 would be good man 2014-09-29 05:12:55 Oh, and by the way, systemdless Arch: https://fleshless.org/pages/spark.html 2014-09-29 05:13:18 I shamelessly shill the thing everywhere :D 2014-09-29 05:13:38 Wait, shill is not the proper word 2014-09-29 05:14:17 no problem, I'm checking it right now 2014-09-29 05:14:34 this systemd is really a pleague 2014-09-29 05:14:46 wanna stay in everything and in everywhere 2014-09-29 05:15:04 this is your tools and blog? 2014-09-29 05:15:45 Yep. 2014-09-29 05:15:55 Well. Not all of the mentioned things are mine 2014-09-29 05:16:11 Some are from the suckless project 2014-09-29 05:16:21 Others are from other people 2014-09-29 05:17:37 spark-rc seems a good idea 2014-09-29 05:18:00 very nice, I remembered one person, ignorant guru 2014-09-29 05:18:33 I like creative people, and competent, of course, like that guy 2014-09-29 05:18:41 If you don't care about things that don't build and run on glibc, Alpine is awesome. Also there's morpheus 2014-09-29 05:18:52 http://morpheus.2f30.org/ 2014-09-29 05:19:13 But morpheus is pretty obscure 2014-09-29 05:19:30 how? 2014-09-29 05:19:41 Prepare your anus for broken scripts that you take from other places, because it doesn't use gnu coreutils and util-linux 2014-09-29 05:20:01 That breaks some stuff. 2014-09-29 05:20:03 seems they have experience in plan9 systems 2014-09-29 05:20:13 hahaha 2014-09-29 05:21:42 I don't actually like package management of morpheus. It's neat and all, but meh. local package cache and direct updates are enough for me. 2014-09-29 05:24:33 pacman, apk and ports are enough to me too 2014-09-29 05:24:42 the best ones 2014-09-29 05:32:32 hmm, rpc.statd: failed to create RPC listeners, exiting 2014-09-29 05:32:39 does NFS work out of the box for anyone? 2014-09-29 08:30:25 Hi, I have a question regarding webktigtk 2014-09-29 08:30:41 since thre says that webkitgtk2 exists 2014-09-29 08:30:51 but I can't find it in the main repo 2014-09-29 08:30:55 :/ 2014-09-29 08:32:16 its not in main 2014-09-29 08:32:20 its in testing 2014-09-29 08:32:29 i added it recently 2014-09-29 08:32:38 edge/testing should be it 2014-09-29 08:32:46 oh, ok 2014-09-29 08:32:49 I'll see 2014-09-29 08:32:58 if you confirm it works as expected, i can move it to main. 2014-09-29 08:33:30 Ok, I'll do 2014-09-29 08:36:59 seems like it confilcts with webkitgtk in /usr/share/gtk-doc/html/webkitgtk 2014-09-29 08:37:06 hm 2014-09-29 08:37:14 i think it should be in main 2014-09-29 08:38:09 clandmeter: i edited the main/webkitgtk to build both for gtk2 and gtk3 2014-09-29 08:39:39 ncopa: did you delete my apkbuidl? 2014-09-29 08:39:46 no 2014-09-29 08:39:55 you didnt know i already made it? 2014-09-29 08:40:04 no i missed it sorry :-/ 2014-09-29 08:40:35 it doesnt conflict? 2014-09-29 08:40:47 or it just overwrites each other 2014-09-29 08:41:05 which? 2014-09-29 08:41:24 what i did was build both gtk2 and gtk3 from same apkbuild 2014-09-29 08:41:29 so that should not conflict 2014-09-29 08:41:40 but it has the same name as the one i made probably 2014-09-29 08:41:44 yes 2014-09-29 08:41:54 but i guess yours is in main 2014-09-29 08:41:57 mine in testing 2014-09-29 08:42:05 yes, because webkitgtk was there already 2014-09-29 08:42:13 you added untested code to main 2014-09-29 08:42:20 ;-) 2014-09-29 08:42:40 and now i get spanked... :-( 2014-09-29 08:42:42 :) 2014-09-29 08:42:54 just delete mine and its fine. 2014-09-29 08:43:05 ok, sorry about that 2014-09-29 08:43:13 np 2014-09-29 08:43:21 i think its better to build both from same apkbuild as it means less maintenance 2014-09-29 09:09:05 some script to check bash for vulnerabilities https://github.com/hannob/bashcheck 2014-09-29 09:17:03 what is the default shell on alpine? 2014-09-29 09:17:09 ash? 2014-09-29 09:24:37 kocka: busybox I think 2014-09-29 09:57:43 busybox ash 2014-09-29 10:24:32 haha. i just figured 2014-09-29 10:24:48 7 years from now we'll have the same kind of vulns in systemd 2014-09-29 10:24:57 and *that* will be rough 2014-09-29 10:24:59 well, not us 2014-09-29 10:25:06 but systems installed by other people :) 2014-09-29 10:50:40 >not us 2014-09-29 10:50:57 Unless something really major (or everything) depends on systemd 2014-09-29 10:51:01 In a few years :D 2014-09-29 10:51:09 (I don't actually think that will happen) 2014-09-29 10:51:37 The only thing systemd provides that makes sense to depend on is logind 2014-09-29 10:51:49 But not much actually needs that API 2014-09-29 10:52:06 Well and socket activation 2014-09-29 10:52:17 But hard-depending on socket activation seems silly 2014-09-29 10:52:35 why do i need logind on either my home desktop/server or my laptop? 2014-09-29 10:53:29 Makes sense to depend for devs who want to do stuff it makes easy to do 2014-09-29 10:53:32 Not for you or me 2014-09-29 10:53:48 logind gives proper multiseat 2014-09-29 10:54:19 It's just that the only place multiseat makes sense in is DM/DE's 2014-09-29 10:54:36 multiseat sounds rather niche to me. there are specialised kiosk/terminal/whatever distros, let them have it 2014-09-29 10:54:44 Yeah, same here 2014-09-29 10:54:53 I don't get why multiseat is so glorified 2014-09-29 10:55:05 Hyped as a killer feature when it's not needed for most 2014-09-29 10:55:08 was reading somewhere of tyeing socket library like nanomsg+runit..etc+kdbus into more like able init system 2014-09-29 10:55:13 most use cases* 2014-09-29 10:55:49 I don't even get why GNOME hard-depends on that API. Or is planning to. 2014-09-29 10:55:59 Can't you do that as an optional feature? 2014-09-29 10:56:33 I don't care about GNOME or multiseat, but is it really that hard to separate multiseat stuff into an option? 2014-09-29 10:58:35 option to socket-activation (I guess that requires services/daemon to be patched) is service availability testing, like querying port 80 for httpd, this feature I think is in "monit" 2014-09-29 10:59:02 And, again, can be made to be completely optional 2014-09-29 10:59:17 I've not seen a daemon yet that hard-depended on it 2014-09-29 10:59:40 so needed would be nice binder for 'service dependency' 2014-09-29 11:00:00 And even then as I use Arch, I just have the systemd libs installed, but not systemd itself 2014-09-29 11:00:07 And the code just doesn't get used :) 2014-09-29 11:00:34 I'm too lazy to support a bunch of packages to build them without the code 2014-09-29 11:00:43 If it stays that way, it's ok. 2014-09-29 11:03:22 But if stuff starts hard-depending on that, the ecosystem will split 2014-09-29 11:03:33 That may actually be good btw. 2014-09-29 12:41:01 I don't get why multiseat is so glorified 2014-09-29 12:41:01 Hyped as a killer feature when it's not needed for most 2014-09-29 12:41:18 I feel like many think that it's required to get Linux adoption in the desktop 2014-09-29 12:41:39 both at home (family members supposedly need separate accounts) and in the business (shared machines) 2014-09-29 12:41:54 so I guess this one goes into the FUD category 2014-09-29 13:12:41 wm4: Weird, I had the impression that both kde a gnome supported this with some minor tweaks. 2014-09-29 13:12:57 something something Compositing WM iirc 2014-09-29 13:14:55 As far as I understand, multiseat isn't the same as the fast user switching thing. Multiseat actually means having multiple keyboards and monitors plugged into the same box, doesn't it? 2014-09-29 13:15:14 yes 2014-09-29 13:15:17 Which is kind of silly these days. 2014-09-29 13:15:28 but according to reddit this is VERY IMPORTANT 2014-09-29 13:15:41 and since reddit reflects the general stupidity of people... 2014-09-29 13:15:43 teiresias: Oh, I remember that. 2014-09-29 13:16:17 teiresias: We kinda tried to do that in the early 2000s to have single PC as a fancy terminal server. When PCs used to be expensive n stuff. 2014-09-29 13:17:26 katlogic: In the early 2000s I was in college, and we had a server with a bunch of thin clients. 2014-09-29 13:17:55 Yeah, this was poor mans version of that. 2014-09-29 13:18:15 Since VGA CRT + PS2 keyboard was still much cheaper than full blown x terminal. 2014-09-29 13:18:27 but basically, it wasnt terribly difficult to get running either 2014-09-29 13:19:00 it was simply multiple x server instances on same box, each using different gfx + ps2 keyboards hacked on lpt port 2014-09-29 13:20:46 wm4: Btw, in case of reddit it almost reeks of sock campaign. 2014-09-29 13:21:10 sock campaign? 2014-09-29 13:21:18 Reddit is basically glorified wikipedia. 2014-09-29 13:21:37 Including the "fight" there to influence "public" opinion. 2014-09-29 15:51:16 regarding webkitgtk*, is javascript disabled 2014-09-29 15:51:34 because I can't get it working on dwb or midori 2014-09-29 15:55:13 asmir: http://git.alpinelinux.org/cgit/aports/tree/main/webkitgtk/APKBUILD 2014-09-29 15:56:02 I've checked it, but still I don't know why doesn't it work on 2 different browser 2014-09-29 15:56:14 browsers* 2014-09-29 16:04:08 ash == busybox 2014-09-29 16:04:11 so yes 2014-09-29 16:04:30 oh, interesting, my client stopped scrolling 2014-09-29 16:04:32 many hours ago 2014-09-29 16:45:25 How do you add a user to a specific group without gpasswd? Do I have to edit /etc/group manualy? 2014-09-29 16:49:39 Ok, I figured it. :$ 2014-09-29 18:50:23 hey all I am looking for paxtest, is there not a repository for the source code 2014-09-29 18:50:40 i can only find two old versions of it at https://pax.grsecurity.net/ 2014-09-29 19:05:29 well I downloaded 0.9.7 version but I apparently a.out.h isn't installed 2014-09-29 19:05:36 I am not sure what package would add this to the system 2014-09-29 19:06:32 i have installed gcc, libgcc, libgc++, libc-dev, libc-utils, musl, 2014-09-29 19:06:41 and the other musl packages 2014-09-29 19:07:07 but none of them contain a.out.h, I search for glibc figuring it would be under there but as far as i can tell that isn't main, 2014-09-29 19:07:15 I will go add and checkout testing 2014-09-29 19:09:25 ? 2014-09-29 19:09:38 a.out.h? a.out is an output i guess.. 2014-09-29 19:09:46 what command are you using for compiling? 2014-09-29 19:11:31 anyway you can download the latest version from here: http://grsecurity.net/~spender/ 2014-09-29 19:12:32 kocka: no as in its reports while I run make 2014-09-29 19:12:41 that #include doesn't exist 2014-09-29 19:13:02 you should not run make... 2014-09-29 19:13:54 kocka: what should I run? 2014-09-29 19:14:34 hm sorry 2014-09-29 19:15:05 under the readme it says to run make generic 2014-09-29 19:15:22 kocka: I will download the current version and see what happens 2014-09-29 19:15:42 yep im wrong 2014-09-29 19:15:55 try make linux with current (0.9.11) 2014-09-29 19:17:51 yah did make linux same problem 2014-09-29 19:19:00 from chpax-0.7/aout.c line num 7 2014-09-29 19:20:03 which includes chpax.h and chpax.h includes on line num 15 in chpax-0.7/chpax.h 2014-09-29 19:20:09 brb 2014-09-29 19:23:31 try to install alpine-sdk 2014-09-29 19:23:57 its a meta-package, maybe installing the missing package for you 2014-09-29 19:24:55 m'kay will do 2014-09-29 19:25:34 installing 2014-09-29 19:26:00 kocka: is there a reason grsec doesn't have a git,hg,cvs,svn or w/e repo for source code? 2014-09-29 19:26:32 nope same problem with alpine-sdk installed 2014-09-29 19:27:24 it has, but private :( 2014-09-29 19:27:33 WAT 2014-09-29 19:27:38 TEh 2014-09-29 19:27:40 FAK 2014-09-29 19:27:46 chpax-0.7/chpax.h:7:19: fatal error: stdio.h: No such file or directory 2014-09-29 19:27:51 #include 2014-09-29 19:28:10 yah i got that but when I installed the list of packages, i listed above 2014-09-29 19:28:13 that went away 2014-09-29 19:28:52 hmm maybe paxtest not compatible with musl 2014-09-29 19:29:45 kocka: should I just run make with the ignore flag and just ignore that recipe? 2014-09-29 19:32:23 seems like just about every recipe is failing running make without the #include file 2014-09-29 19:37:49 kocka: thanks for your help so far 2014-09-29 19:47:43 kocka: also when I run gradm to turn on fullsystem learning mode to mess around with grsec it says /etc/grsec/pw file doesn't eist 2014-09-29 19:49:03 ahh i think I found it 2014-09-29 19:49:09 *what I did wrong 2014-09-29 19:54:43 still can't figure why I can't use js with webkit browsers 2014-09-29 19:55:03 does anyone here use them and has a working js? 2014-09-29 20:19:45 asmir: i haven't tried yet but should be able shortly 2014-09-29 20:20:29 systmkor2: I don't get it? 2014-09-29 20:21:08 systmkor2: oh, just ignore ^^^ 2014-09-29 20:21:54 asmir: m'kay 2014-09-29 20:22:12 systmkor2: I guess it's easiest just to install midori which is in main 2014-09-29 20:23:03 m'kay 2014-09-29 20:23:10 asmir: what browser are you trying with? 2014-09-29 20:23:34 systmkor2: I tried with dwb, surf and midori 2014-09-29 20:24:03 systmkor2: none works, firefox does though :( 2014-09-29 20:38:44 asmir, try in settings to see if there is "Allow all sites to run JavaScript (recommended)" option and it is enabled 2014-09-29 20:40:06 gtg 2014-09-29 20:43:29 oh, sry I didn't see the message 2014-09-29 20:43:59 and yes javascript is enabled on all sites and all browsers I tried (dwb, surf, midori) 2014-09-29 20:45:01 I just hope it's not some webkit-musl bug since apple doesnt really fix things that don't affect safari 2014-09-29 20:45:19 and I can't browser without dwb :( 2014-09-29 20:45:27 browse* 2014-09-30 11:53:25 Hi, I would like badly to use alpine to launc qemu-kvm vms connected trough openvswitch. unfortunately, openvswitch is not yet ready fot alpine (it segfault at boot time and later on when tried manually). How can I help to fix that? I'm new to Alpine but not to Linux ;) I already compiled and installed OVs on other platforms but with "standard" kernel and libc... I have no experience with musl yet. 2014-09-30 11:58:09 does it segfault on alpine with the vanilla kernel? 2014-09-30 11:58:42 That was my next planned move. to use the vanilla kernel as most errors are related to grsec in dmesg 2014-09-30 12:05:35 i think its openvswitch with musl combo 2014-09-30 12:19:32 I think you are right :/ I will test vanilla this afternoon 2014-09-30 12:24:01 Jean-Scotch: i think there is an openvswitch-dbg package too 2014-09-30 12:24:11 try generate a core dump and a bactrace 2014-09-30 12:27:01 is there an -edge iso for run-from-ram? 2014-09-30 12:27:38 (because of upgrade in future) without need to do setap-bootalbe and then switch to edge again 2014-09-30 12:28:49 setup-alpine 2014-09-30 12:28:49 echo < /media/usb/apks 2014-09-30 12:28:49 http://dl-5.alpinelinux.org/alpine/edge/main 2014-09-30 12:28:49 http://dl-5.alpinelinux.org/alpine/edge/testing 2014-09-30 12:28:49 EOF > /etc/apk/repositories 2014-09-30 12:28:49 apk update 2014-09-30 12:28:50 apk upgrade 2014-09-30 12:30:59 crow: we sometimes do edge snapshots 2014-09-30 12:31:16 edge snapshots isos 2014-09-30 12:31:25 i thinkwe should do that more often really 2014-09-30 12:31:31 you can also easily spin them yourself 2014-09-30 12:31:46 i also have an idea on updating kernel on run-from-ram 2014-09-30 12:31:52 although asking ncopa is much easier *grin* 2014-09-30 12:32:24 a script that apk fetch kernel to temp dir and generates modloop and initramfs 2014-09-30 12:32:35 a script to regenerate /boot in other words 2014-09-30 12:33:03 ncopa can you trigger one new edge snapshot iso? 2014-09-30 12:33:11 ok 2014-09-30 12:33:41 i would like to run-from-ram an edge and dont see reason why keep stalbe arround redo setup-bootable and switch repo 2014-09-30 12:33:49 yeah 2014-09-30 12:33:55 clandmeter to generate them myself? 2014-09-30 12:34:28 crow: git clone git://git.alpinelinux.org/alpine-iso && cd alpine-iso && make iso 2014-09-30 12:34:43 damn you are too fast :) 2014-09-30 12:34:50 great thanks, so setup-bootable with an edge snapshot iso will bring it to latest released snapshot or? 2014-09-30 12:34:51 ncopa, arch makes em once a month. That's probably a reasonable frequency for edge isos? 2014-09-30 12:34:55 make PROFILE=*** iso 2014-09-30 12:35:12 if you like another iso then default. 2014-09-30 12:35:20 openfbtd: once a month sounds good, i was thinking more like, whenever kernel is updated :) 2014-09-30 12:35:33 Hmm. Yeah, that would work too. 2014-09-30 12:35:43 the problme is that its not automated 2014-09-30 12:35:45 +1 for whenever kernel is updated 2014-09-30 12:35:56 and i forget... 2014-09-30 12:35:57 maybe have them named after kernel if you do that? 2014-09-30 12:35:58 clandmeter thank you :) 2014-09-30 12:36:18 ncopa, well are you subscribed to kernel updates? :) 2014-09-30 12:36:38 i do update the kernel... 2014-09-30 12:36:54 i am actually doing it as we speak 2014-09-30 12:37:06 Well, then tying regenerating isos to that would probably be perfect for you. 2014-09-30 12:37:50 hi all 2014-09-30 12:38:42 any one known a way to a unattended install of alpine ? 2014-09-30 12:40:10 castorinop: there is an undocumented way to do it: setup-alpine -c unattended.conf 2014-09-30 12:40:14 edit unattended.conf 2014-09-30 12:40:23 and then: setup-alpine -f unattended.conf 2014-09-30 12:41:12 Is it undocumented because it's not tested very well? 2014-09-30 12:41:17 Or someone just forgot? :) 2014-09-30 12:41:39 i think its undocumented because i am not sure this is the way we want do it in the longterm 2014-09-30 12:41:50 ah 2014-09-30 12:41:52 i was hoping come up with something better, but have not been able to 2014-09-30 12:43:48 ncopa so setup-bootable with an edge snapshot iso run-from-ram sd card will bring it to latest released edge snapshot in future right? 2014-09-30 12:45:40 ncopa: im working with theforeman, this use puppet. And i writting some modules to include apk-tools, and update anothers for include alpine. 2014-09-30 12:47:25 theforeman include Provisioning, this deploy instances of bare-metal or VMs 2014-09-30 12:49:28 castorinop: could you say pxe + http or similar instead of the specific tool 2014-09-30 12:49:29 i wish generate the scripts to add alpine 2014-09-30 12:49:43 better chance to help 2014-09-30 12:50:49 yes, darkfader. 2014-09-30 12:51:24 crow: setup-bootable -u should upgrade the kernel on your usb. 2014-09-30 12:51:45 clandmeter: yes, but he needs an updated iso 2014-09-30 12:52:40 i dont really understand his question 2014-09-30 12:52:54 i know he asked for an updated iso 2014-09-30 12:53:17 but he also talks about upgrading his sd card 2014-09-30 13:00:17 with the vanilla kernel, I do not have a segfault for openvswitch at startup. It still fire some post-install error at boot time but the message scrolled to fast to be able to copy it by hand 2014-09-30 13:00:32 Jean-Scotch: try apk fix 2014-09-30 13:00:37 to see the msg 2014-09-30 13:00:38 interesting 2014-09-30 13:00:52 interesting that it does not segfault with vanilla kernel 2014-09-30 13:01:45 ok. /etc/openvswitch/.conf.db.~locl~: cannot lock file because it is already locked by pid 1884 2014-09-30 13:02:35 ovsdb-tool: /etc/openvswitch/conf.db: failed to lock lockfile 5Resource temporarily unavailable) 2014-09-30 13:03:02 ERROR: openvswitch-2.3.0-r3.post-upgrade: script exited with error 1 2014-09-30 13:03:26 those are the 3 relevant lines 2014-09-30 13:03:59 where can I find this script to read it's code? 2014-09-30 13:05:46 Jean-Scotch: http://git.alpinelinux.org/cgit/aports/tree/testing/openvswitch 2014-09-30 13:05:56 thanks 2014-09-30 13:08:57 maybe both scripts get called on tmpfs install? 2014-09-30 13:09:11 ovs-vswitchd is in fact still crashing in the background (ps ax shows "ovs-vswitchd: monitoring pid 2118 (4 crashes: pid 2014 di..." 2014-09-30 13:09:50 and giving repeated segfault (viez from dmesg) 2014-09-30 13:10:22 ncopa, would it be difficult to have web interface for generating iso for a given kernel ? 2014-09-30 13:10:44 iso would be generated for 24hrs and then notified to user then deleted 2014-09-30 13:11:16 user could also select packages, maybe limit the total iso to say upto 300mb 2014-09-30 13:12:19 if the iso's are generated on earch new kernel, then there is no reason to create a new one. 2014-09-30 13:12:58 except if you really want specific iso's 2014-09-30 13:13:16 then the pkgs choice makes it different 2014-09-30 13:13:50 start hacking :) 2014-09-30 13:14:33 vkrishn: not difficult, just timeconsuming 2014-09-30 13:14:39 problem is getting to know the git triggers , ones that effect after each new kernel intro 2014-09-30 13:15:38 those code are still incubating somewhere in ncopa's laptop ;) 2014-09-30 13:25:59 I guess the post-upgrade script of openvswitch gives an error because urandom is not started yet... 2014-09-30 13:28:35 and apk fix is working gracefully if I stop ovsdb-server manualy before running it 2014-09-30 13:29:52 ncopa: since i just remembered, you had asked someothing about edge upgrade docs some time back. maybe this helps: 2014-09-30 13:29:55 http://hastebin.com/reyinarova.vhdl 2014-09-30 13:30:02 its' the silly ansible playbook i used for updating 2014-09-30 13:30:19 i don't remember much more but i thought i better put it online before i forget again 2014-09-30 13:41:47 clandmeter well once i am on edge iso snapshot, setab-bootable -u will pull also if avaiable new edge snapshot iso. not the "stable" one.. 2014-09-30 13:48:16 oh i check the script, i thought it would generate a initramfs and modloop from repo and put it on disk, but it seems it copies it from iso. 2014-09-30 13:49:59 clandmeter yea iso need to be mounted first to be able to update 2014-09-30 13:50:15 ncopa sorry to ask but iso build done? 2014-09-30 13:52:20 which type you want? i can build one for you if you like. 2014-09-30 13:53:18 edge x86-64 2014-09-30 13:53:33 thank you as i dont have build env jet.. 2014-09-30 14:02:05 crow: pm 2014-09-30 14:31:25 http://nl.alpinelinux.org/alpine/edge/releases/x86_64/ 2014-09-30 14:31:30 crow: ^^^ 2014-09-30 14:42:15 ncopa thanks 2014-09-30 14:53:15 vkrishn: im thinking of changing fs to zfs on nl-infra server 2014-09-30 14:53:28 not sure when ill make the change though. 2014-09-30 14:53:40 clandmeter: we should replace the vserver host with lxc too 2014-09-30 14:53:47 it was actually not that hard 2014-09-30 14:54:09 that should be easier yes 2014-09-30 14:54:34 does the current kernel support lxc? 2014-09-30 14:54:36 i just rsynced /vserver/ to remote:/var/lib/lxc//rootfs 2014-09-30 14:55:07 and created an /etc/network/interfaces 2014-09-30 14:55:21 the current vserver kernel? 2014-09-30 14:55:22 i dont know 2014-09-30 14:55:32 i could run both vserver and lxc :) 2014-09-30 14:55:37 zfs now works on AL :-)) 2014-09-30 14:55:40 nice 2014-09-30 14:55:43 migrate 2014-09-30 14:55:54 and then just change back to grsec kernel after finished 2014-09-30 14:56:19 clandmeter: sounds like a good plan 2014-09-30 14:56:26 im not 100% it supports lxc though 2014-09-30 14:56:31 :) 2014-09-30 14:56:42 I want to change the network also 2014-09-30 14:56:44 but it wouldnt surprise me if it does 2014-09-30 14:56:47 ok 2014-09-30 14:56:50 I like the way tis not on my other server 2014-09-30 14:57:13 i just create a 172 subnet and manage the rules from our main firewall. 2014-09-30 14:57:35 and bridge? 2014-09-30 14:58:07 we could maybe even run the lxc directly on the iron, instead of inside a kvm box 2014-09-30 14:58:15 would save resources 2014-09-30 14:58:41 yes of course 2014-09-30 14:58:50 oh yes you are right 2014-09-30 14:59:09 we have 2 layers 2014-09-30 14:59:10 can you boot from a ZFS partition? or you have to create an ext2/3/4 to place the kernel and initramfs? 2014-09-30 14:59:32 royger: i didnt try initramfs 2014-09-30 14:59:54 then you compile zfs support inside of the kernel? 2014-09-30 15:00:09 royger: why? don't you want to have an encrypted boot? if so, that doesn't have to be zfs, while your rootfs might be. 2014-09-30 15:00:22 grub2 has zfs.mod 2014-09-30 15:00:32 meh s/encrypted boot/encrypted root/ 2014-09-30 15:00:58 meh i'm confused and return to what i've been doing before spilling more stupidity. 2014-09-30 15:01:00 im running it from tmpfs, so my boot is always on something else then zfs. 2014-09-30 15:01:01 have not worked on zfs though 2014-09-30 15:02:42 ok, you can also have encrypted root with ZFS 2014-09-30 15:52:20 ncopa installed that one iso to (inside an vm), wonder why the repository points to main and not to edge or that is not how it works :) 2014-09-30 15:52:36 ncopa for me it looks like current snapshot 2014-09-30 15:53:22 i was thinking repository will point to edge like here http://wiki.alpinelinux.org/wiki/Edge 2014-09-30 19:03:45 clandmeter: why switch from vserver? afaik lxc is still not really PaaS ready - not without significant restricting in selinux/apparmor/someother