2017-06-01 11:19:41 Hello everyone :) 2017-06-01 11:20:56 I have a problem with the openjdk8-jre package that has been released in 3.6. However i do not manage to submit a bug report in the alpinelinux bug tracker. How can i contact the maintainer? 2017-06-01 11:25:02 the maintainer is fabled 2017-06-01 11:39:04 AGiardini[m], preferably file bug, or email to alpine-devel mailing list. on minor issues irc is good too, but there's tendency to forget things said here. 2017-06-01 14:06:28 fabled: I tried to file a bug but i didn't succeed, I don't manage to register on the bug platform since it's asking me for 'confirmation' and the mail `alpine@bugs.alpinelinux.org` errors as `undeliverable` 2017-06-01 14:26:41 AGiardini[m], as which user did you register? 2017-06-01 14:27:35 clandmeter: There is a mandatory field called `confirmation` and I have no idea what to put in there... is it some kind of invitation code? 2017-06-01 14:28:07 huh? 2017-06-01 14:28:12 screenshot? 2017-06-01 14:29:11 ACTION uploaded an image: wat.png (27KB)  2017-06-01 14:29:54 heh 2017-06-01 14:30:01 password confirmation 2017-06-01 14:31:04 ow 2017-06-01 14:31:08 i feel so stupid :D 2017-06-01 14:32:22 you shouldnt 2017-06-01 14:32:26 its not very clear 2017-06-01 14:32:45 although in my own language it is. 2017-06-01 14:33:12 <_ikke_> bevestiging? 2017-06-01 14:34:15 no it says: Bevestig wachtwoord 2017-06-01 14:34:20 <_ikke_> ah ok 2017-06-01 14:34:25 which makes it obvious 2017-06-01 15:42:36 what is the difference between linux-vanilla and linux-hardened? 2017-06-01 15:45:36 <_ikke_> linux-hardened has several patches applied which would improve security / harder to exploit vulnerabilities 2017-06-01 15:45:57 <_ikke_> vanilla is the original linux kernel 2017-06-01 15:46:19 how does -hardened compare to -grsec? 2017-06-01 15:46:49 <_ikke_> Xe: grsec has been renamed to hardened after grsec announced that they will not provide any patches anymore to the public 2017-06-01 15:46:57 ah 2017-06-01 15:47:23 <_ikke_> currently, it still uses grsec, but the question is for how long 2017-06-01 16:16:48 Which alpine do you recommend to run from ram? 2017-06-01 16:19:20 "Most common used packages included. Suitable for routers and servers. Runs from RAM." What i'm missing is a good howto.. 2017-06-01 16:22:46 helpImStuck: the descriptions are kinda off right now 2017-06-01 16:22:50 id just go with standard 2017-06-01 16:22:58 if you have networking 2017-06-01 16:23:06 okey 2017-06-01 16:23:52 Didn't we talk about locales and chasrsets the other week? 2017-06-01 16:24:56 helpImStuck: there is no locale but UTF-8 2017-06-01 16:26:31 Xe, I'm sorry if I can't express the right terminology. With locale i mean charset.. loadkeys 2017-06-01 16:28:07 I couldn't get it to stick.. always some keys that where wrong.. but i have now found out that the keyboard i'm using isn't to trust. 2017-06-01 16:29:47 helpImStuck: # setup-keyamp 2017-06-01 16:29:49 map* 2017-06-01 16:33:59 ye, my memory is good but short.. loadkeys is for arch.. 2017-06-01 16:53:01 hello @ all 2017-06-01 16:54:36 killall -9 claws-mail 2017-06-01 16:54:38 er 2017-06-01 16:54:42 I'm trying to get my keyboard setup reboot stable, any hint's? 2017-06-01 17:03:25 heya 2017-06-01 17:03:35 feuerteufel: what seems to be the issue? 2017-06-01 17:05:13 IOn mei PC is a german keyboard but Alpine starts always up with an US 2017-06-01 17:05:40 I always have to change it 2017-06-01 17:06:01 could you tell me what /etc/conf.d/keymaps says? 2017-06-01 17:07:22 same question to helpImStuck 2017-06-01 17:08:50 The first line say's --keymap="us"-- 2017-06-01 17:09:18 the last line is what i need to know 2017-06-01 17:10:08 fix_euro="NO" 2017-06-01 17:10:30 Zhat's the last line 2017-06-01 17:10:46 *That's 2017-06-01 17:11:05 hmm 2017-06-01 17:11:11 after you run setup-keymaps, what's the last line? 2017-06-01 17:13:07 I yust looked at mei other Maschine, but that is a VM ther is one more: 2017-06-01 17:13:10 Shiz, don't know.. but the keymap problem is no more.. it was my keyboard.. it is broken :) All keys and such works fine via ssh 2017-06-01 17:13:17 KEYMAP=/etc/keymap/de-latin1-nodeadkeys.bmap.gz 2017-06-01 17:13:49 helpImStuck: well, keymaps are supposed to only apply to your phyiscal keyboard, so if that doesn't work our keymaps are likely broken 2017-06-01 17:13:50 Is that the line you are missing? 2017-06-01 17:13:51 not your keyboard :) 2017-06-01 17:14:25 feuerteufel: right -- does that file exist? 2017-06-01 17:14:56 Yes, on both maschines 2017-06-01 17:15:19 So yust put that line in there? 2017-06-01 17:15:43 well, that's the hting 2017-06-01 17:15:49 setup-keymaps should put that line there for you 2017-06-01 17:16:27 I remember having problems even after i put the line there.. i always had to switch to tty2 to get the right keys.. or log in via ssh 2017-06-01 17:18:40 That's the reason I always forgot to ask, most of the time I use SSH 2017-06-01 17:34:34 alpine has been around since 2005 right? 2017-06-01 17:59:31 helpImStuck: yeah, 2005 2017-06-01 18:02:01 12 years of making and still problems with the keymap :$ 2017-06-01 18:02:09 but i still like alpine 2017-06-01 18:02:45 i think the keymap issues are only recent 2017-06-01 18:02:51 'pushing to get them fixed in 3.6 2017-06-01 18:02:58 :) 2017-06-01 18:03:05 sorry, 3.6.1* 2017-06-01 18:03:20 yes, next release, me get it 2017-06-01 18:03:46 doesn't help that i don't have a physica lbox handy to test on :P 2017-06-01 18:04:40 can't you virtualise it? 2017-06-01 18:06:36 i have turned off upnp on my modem.. but in tcpdump it still pings 224.0.0.1.. quality stuff my isp gave me 2017-06-01 18:07:20 multicast* 2017-06-01 18:09:32 it was.. is it safer to bridge the modem and use a router with tomato by shibby on? but that's off topic.. i'm going to config my lan now.. again. bbl 2017-06-01 18:11:32 can't even access my modems web interface.. and in the log file there was a ddos notice.. who would ddos a person.. well.. later 2017-06-01 18:27:44 A distribution that is more then 10 years old, so it must be perfect... 2017-06-01 19:48:50 hi 2017-06-01 19:55:56 hello 2017-06-01 19:56:40 hello Shiz 2017-06-01 20:13:36 \o/ 2017-06-01 20:15:43 \o/ 2017-06-01 20:15:44 congratz 2017-06-01 20:16:50 danke 2017-06-01 20:16:53 at least rpi will now boot again 2017-06-01 21:04:32 good night :) 2017-06-01 21:29:52 hi, on the wiki upgrade instructions (https://wiki.alpinelinux.org/wiki/Upgrading_Alpine#Update_remaining_packages_from_Web_repository) the setup-bootable command doesn't look right. It says to run: setup-bootable -u ERROR in latestalp.php: failed to read /var/www/dl-3.alpinelinux.org/htdocs/alpine/.latest.x86_64.txt /media/$LBU_MEDIA . What should the source be? 2017-06-01 21:30:30 ho lord this stuff is old 2017-06-01 21:33:04 yes it looks that way, but I can't find a newer article. How do I upgrade my installation running from RAM? 2017-06-01 21:34:22 budric[m]: http://dl-4.alpinelinux.org/alpine/v3.6/releases/x86_64/alpine-standard-3.6.0-x86_64.iso 2017-06-01 21:34:26 you'd want an url like this 2017-06-01 21:34:31 change arch/edition if relevant 2017-06-01 21:36:08 i see thanks 2017-06-01 21:37:37 does setup-bootable work with gz? The armhf doesn't have isos 2017-06-01 21:38:22 i dont think so 2017-06-01 21:42:13 is it possible to upgrade raspberry pi system? Or is it complete reinstall ? 2017-06-01 21:42:33 i'm honestly not entirely sure, do you use lbu right now? 2017-06-01 21:45:02 yes I use lbu 2017-06-01 21:46:49 then i think you can just overwrite the files on your sd or whatever you use to boot, but don't overwrite the apkovl 2017-06-01 21:46:51 and it should work 2017-06-01 21:46:58 but no guarantees since rpi is not my area, so keep backups :D 2017-06-01 21:48:58 by apkvol do you mean /media/mmcblk0p1/apks/* ? 2017-06-01 21:51:52 the file lbu writes to 2017-06-01 21:51:56 should end in apkovl.gz 2017-06-01 21:52:31 oh i see it 2017-06-01 21:52:47 yeah so dont delete that one 2017-06-01 21:52:57 but you can replace the other files by the new files form the release and it should work 2017-06-01 21:53:02 (make backups, again) 2017-06-01 22:09:07 Shiz: that actually worked. thanks 2017-06-01 22:09:17 :) 2017-06-01 22:09:44 you're just in time too, as 3.6.1 was released today that fixed the rpi kernel not booting... 2017-06-01 22:10:25 it wasn't booting in 3.6.0? 2017-06-01 22:10:50 apparently 2017-06-01 22:13:12 yeah that's lucky. I was trying to upgrade to 3.6.0, and the packages were upgraded, but I didn't realise I need these extra steps to upgrade the kernel 2017-06-01 22:13:50 Shiz is our hero 2017-06-01 22:15:47 Is it possible to run virt-manager on a remote host to manage alpine's xen distro? 2017-06-01 22:16:56 Every time i've tried i've faild because the system wouldn't let me do the changes needed. 2017-06-01 22:41:57 setup-apkcache to media/sdb/cache.. says read only filesystem.. it's a usb stick 2017-06-01 22:42:21 do I need 2 usb sticks for a ram-install? 2017-06-01 22:43:04 mount -o remount,rw /media/sdb, maybe? 2017-06-01 22:43:15 i'll try 2017-06-01 22:44:54 hm, did not help but dev/sdb is full when i typ df -h 2017-06-01 22:49:32 found another usbstick.. will use it instead 2017-06-01 23:16:34 hey hows the packager? 2017-06-01 23:18:49 hmm? 2017-06-01 23:56:06 is it a good idea to put usbcache in etc on a diskless install? 2017-06-02 00:08:05 mount: mounting /dev/usbdisk on /media/usb failed: Invalid argument 2017-06-02 00:08:05 failed to mount /media/usb 2017-06-02 00:14:32 setup-lbu > usb > mount: mounting /dev/usbdisk on /media/usb failed: Invalid argument.. > ls /dev/usbdisk -lah > lrwxrwxrwx 1 rootroot 9 Jun 2 02:07 /dev/usbdisk -> /dev/sdc1 2017-06-02 00:14:44 sdc1 is a ext4 formatted usb stick 2017-06-02 00:14:50 an* 2017-06-02 00:16:04 ah, fk it.. i'll do a disk install to the usb stick instead.. 2017-06-02 00:17:05 this takes too much time 2017-06-02 00:17:23 trying to create a ram-install 2017-06-02 00:20:03 noam chumpsky is an idiot 2017-06-02 00:20:13 openbsd here 2017-06-02 00:20:17 why move to alpine? 2017-06-02 00:27:12 helpImStuck: lsblk -f to se what drives you have 2017-06-02 00:27:54 azathoth99: maybe he wants a change? 2017-06-02 01:10:37 jushur, it took too long.. now i'm in the middle of getting lighttpd working.. I get 403 forbidden when i try to access the test.php with cgi enabled 2017-06-02 01:29:10 which dependencies does lighttpd with php5 need to run cgi? 2017-06-02 02:08:35 so I have this FAQ: http://wiki.alpinelinux.org/wiki/Alpine_Linux:FAQ#WARNING:_Ignoring_APKINDEX.xxxx.tar.gz but my /etc/apk/repositories correctly has just http://dl-cdn.alpinelinux.org/alpine/v3.6/main (and the community repo) in it 2017-06-02 02:40:41 helpImStuck: typically you want to add a -t argument to mount, as it can't always figure out the correct file system 2017-06-02 02:40:45 that's likely the invalid argument error 2017-06-02 02:40:51 azathoth99: what do you mean by 'how is the packager'? 2017-06-02 02:41:53 helpImStuck: it seems better to use fastcgi (php7-fpm) with lighttpd 2017-06-02 02:42:00 dotplus: what repo does it state before the message? 2017-06-02 02:42:27 e.g., what comes before $arch/APKINDEX.tar.gz in the error message? 2017-06-02 03:40:41 Shiz, which distro genre does alpine build on? 2017-06-02 03:40:57 what's a distro genre? 2017-06-02 03:41:09 like music genre but a distro 2017-06-02 03:41:35 ubuntu was born from debian.. manjaro from arch.. and alpine from? 2017-06-02 03:41:52 itself :P 2017-06-02 03:41:58 alpine doesn't derive from any distro 2017-06-02 03:42:20 it has influences from gentoo and arch but thats it 2017-06-02 03:42:41 it has drawn inspiration from LEAF about 12 years ago, but most of that is gone by now I think 2017-06-02 03:42:58 abuild is very loosely based on makepkg principles, and that's about where the similarities end 2017-06-02 03:44:29 I think the likeliness of abuild and makepkg are the biggest similarities you will find between Alpine and another distro 2017-06-02 03:44:54 i see :o 2017-06-02 03:44:59 in terms of what makes distros unique of course -- we all use Linux kernels and carry a POSIX userspace, of course 2017-06-02 03:45:10 yes 2017-06-02 03:45:40 Shiz: WARNING: Ignoring http://dl-cdn.alpinelinux.org/alpine/v3.5/main/x86_64/APKINDEX.tar.gz: temporary error (try again later) 2017-06-02 03:45:47 also 3.6 2017-06-02 03:46:09 dotplus: that's a different error than what the wiki says :) 2017-06-02 03:46:14 dotplus: is your DNS/internet working? 2017-06-02 03:46:35 the wiki concerns the 'No such file or directory' case, which is the case when the repo doesn't exist 2017-06-02 03:46:43 <_ikke_> dotplus: works fine here 2017-06-02 03:46:44 temp error means apk is having issues reaching the repo 2017-06-02 03:49:18 Yes, it is a different warning/error. I get both. If it matters/helps, this building a docker container FROM alpine:(latest|3.5). So it sounds like the container can't reach the outside world. hmm. 2017-06-02 03:50:08 what does busybox have that can help me prove that? I guess I need to go read what's in busybox 2017-06-02 03:50:36 <_ikke_> ping? 2017-06-02 03:50:53 damongant: ping/nslookup/wget 2017-06-02 03:50:59 are probably your best friends 2017-06-02 03:51:09 you can also try if changing the mirror works, we'veh ad some issues with dl-cdn randomly lately 2017-06-02 03:51:27 so a sed -i 's/dl-cdn/cz/g' /etc/apk/repositories may prove to fix some stuff 2017-06-02 03:51:30 for instance 2017-06-02 03:51:51 there are always some mirrors not responding when looking for the fastest mirror 2017-06-02 03:54:00 and how can i make alpine a router, use split vpn for some users that seeds my tv-series, and a seedbox.. all in one. There's always something going bananas.. or i'll just have to keep trying and reading 2017-06-02 03:55:57 and why doesn't it work well with virt-manager? libvirt is the latest version, as well.. 2017-06-02 03:56:56 ok, checking busybox docs says I have standard iproute2 stuff. yeah, bad networking. as suspected, this is not alpine-linux's "fault". thanks for the hints. 2017-06-02 03:57:11 np :) 2017-06-02 03:57:27 helpImStuck: posting more specific errors would be helpful 2017-06-02 03:57:33 (although i'm going to try to catch some sleep now) 2017-06-02 03:57:47 <_ikke_> ACTION throws Shiz some sleep 2017-06-02 03:57:47 good thing this is just a toy project to learn something:) 2017-06-02 03:58:52 Shiz, it's not so much about the errors, i just can't seam to get it together, wish i could sleep too. but i'm doomed to play around with gnu/linux. 2017-06-02 03:59:48 dotplus: i can confirm that dl-cdn is having issues for me too 2017-06-02 03:59:52 id switch to another mirror 2017-06-02 04:00:02 "how can I make alpine a router"? it already is: net.ipv4.ip_forward = 1 2017-06-02 04:00:31 we're working on a better solution for dl-cdn that'll resolve these issues once and for all, hopefully 2017-06-02 04:02:49 cool, but I'm not even a datapoint on that atm. I'm not even able to get past the default gateway. something is borked with my docker networking it seems. 2017-06-02 04:03:20 But I'll troubleshoot that after some sleep:) 2017-06-02 04:03:55 aha 2017-06-02 04:03:57 good luck :) 2017-06-02 04:53:12 is alipine better than freebsd and openbsd? 2017-06-02 04:53:16 is so why? 2017-06-02 05:01:37 azathoth99: you use what you like/need/know. not what others say is good.. 2017-06-02 05:03:44 <_ikke_> and better is subjective 2017-06-02 05:03:53 <_ikke_> depending on the criteria 2017-06-02 05:11:02 it's a quite silly question to ask in a channel where people use alpine hang 2017-06-02 05:11:04 :p 2017-06-02 05:18:07 lol dont tell me what to use 2017-06-02 05:18:14 you are no authority 2017-06-02 05:18:20 nor in psition to give advice 2017-06-02 05:18:30 you presumptious wingnut 2017-06-02 05:18:37 is this a markov chain bot 2017-06-02 05:21:47 no its a ford-marigold lemma 2017-06-02 05:25:00 trolls are out early tonight i see 2017-06-02 05:25:58 shiz is right tho, for an objective answer on a subject it's usually not the best place to ask if you're in the den of one of the extremes 2017-06-02 05:27:32 okay, so it's not like this place is extreme in any meaning of the word, but... :) 2017-06-02 05:28:03 anti-gnu extremists! 2017-06-02 05:28:08 !11 2017-06-02 05:28:24 ACTION looks at his system-llvm branch 2017-06-02 05:28:26 ACTION starts sweating 2017-06-02 05:28:45 :^) 2017-06-02 05:30:11 all I have to say to the topic of alpine vs *bsd is, I switched away from freebsd about half a year ago, those boxes now run Alpine 2017-06-02 05:32:11 and the only real reasons to that were that a hard drive in my freebsd box finally went kaputt and I went for something I'm more familiar with; I don't tinker with freebsd that much anymore and from a sysadmin point of view that's a bad thing should everything requiring closer attention happen 2017-06-02 05:51:10 tcc! 2017-06-02 11:55:27 hello , i tried to run alpine 3.6 on raspberry pi 2 and with no success . I unpack tar to sdcard FAT32 partion , make bootable , put sdcard to rpi2 and nothinh happend , seems like kernel not booting . On 3.5 branch no problem , but some packages are outdated to me . Is this problem only in my rpi2 ? 2017-06-02 11:58:55 Guest18175, yes, 3.6.0 is broken on rpi2/3, it is fixed and pending 3.6.1 release 2017-06-02 11:59:34 oh, 3.6.1 was released yesterday, that should work 2017-06-02 12:54:10 azathoth99's question is warranted 2017-06-02 12:54:47 there are similarities between alpine and freebsd/openbsd that are not shared between alpine and mainstream linux 2017-06-02 12:55:05 this is part of the appeal for me 2017-06-02 13:57:21 i barely even felt like he was asking a question to be honest 2017-06-02 14:17:57 comparing alpine against freebsd and openbsd is valid, but the question was poorly asked and not qualified 2017-06-02 15:01:11 Finally, alpine on pi is up again. \o/ 2017-06-02 15:01:25 Need to rebuild it again. 2017-06-02 15:02:00 T_T 2017-06-02 15:02:18 ACTION is trying out ram mode now. 2017-06-02 15:04:32 algitbot: Who is your creator? 2017-06-02 15:04:39 Wow, what a bot. 2017-06-02 16:05:32 Shiz: agreed, especially his shit later suggests that :) 2017-06-02 16:05:58 still, i'd be happy to read any answer a comparable question would result in 2017-06-02 18:41:58 does anyone use Network Manager under Alpine linux? 2017-06-02 18:54:08 Hello o/ I have a bug for which I don’t know where to ask help for. 2017-06-02 18:54:46 I’m using alpine on my laptop (lenovo thinkpad x250), and I run some games that require glibc OS in a debian jessie chroot. 2017-06-02 18:54:47 heya, what's the matter 2017-06-02 18:55:26 And something like ~2 months ago, all games stopped in the chroot stopped working 2017-06-02 18:55:45 More precisely, all programs running opengl stopped working. 2017-06-02 18:56:06 They freeze once the first frame is displayed. 2017-06-02 18:56:17 well that seems problematic 2017-06-02 18:56:27 So my question is : should I ask for help here, or ask the debian community? 2017-06-02 18:56:33 anything in particular you did two months ago? 2017-06-02 18:56:44 Nope :/ 2017-06-02 18:56:52 I mean nothing but updating my system. 2017-06-02 18:57:08 the first think i'd think of is some kind of svere incompatibility between the game's shipped libraries and the X server/DRM/Mesa APIs that your alpine 'host' runs 2017-06-02 18:57:26 But a friend of mine with the exact same config (alpine on x250 + debian chroot) has not encountered this issue. 2017-06-02 18:57:43 i presume reinstalling the chroot did not fix it? 2017-06-02 18:57:46 or using their chroot 2017-06-02 18:57:58 Shiz: but even the mesa utils freeze. 2017-06-02 18:58:09 Even glxgears or glxinfo. 2017-06-02 18:58:09 alpine mesa utils? 2017-06-02 18:58:15 Nope, debian mesa utils. 2017-06-02 18:58:19 right 2017-06-02 18:58:24 well i mean 2017-06-02 18:58:30 I did not try re-installing another chroot though 2017-06-02 18:58:38 It might be worth it /o\ 2017-06-02 18:58:38 depending on the game it may probably link against the 'host' (debian chroot in this case) OS dynamic libraries 2017-06-02 18:58:44 so s/game's shipped libraries/debian chroot libraries/ 2017-06-02 18:58:50 in ym statement above 2017-06-02 18:58:51 Yep 2017-06-02 18:58:51 :P 2017-06-02 18:59:05 I hope it’s as simple as that, and some tweaks in debian may solve this. 2017-06-02 18:59:52 if alpine's glxgears etc works fine, i'd wager it's some kind of incompat between debian client libs and the alpine servers it interacts with 2017-06-02 19:00:05 or a bug in debian of course, but that seems a bit less unlikely 2017-06-02 19:01:44 I will give a try to a debian 9 chroot 2017-06-02 19:01:51 :) 2017-06-02 19:02:10 And if it’s still not working, will ask on #debian, and then maybe come back here :3 2017-06-02 19:04:27 sorry i can't give you anything more to go on, but i don't run debian myself 2017-06-02 19:05:45 No problem :) 2017-06-02 19:08:55 I'd also check to make sure your graphics driver hasn't changed 2017-06-02 19:10:20 qman__: how do I do that? 2017-06-02 19:12:24 lspci -k can do it, but I'm not sure if that'll work in alpine 2017-06-02 19:13:16 your xorg logs should show it too 2017-06-02 19:13:39 lspci -k does the trick :) 2017-06-02 19:13:56 But I don’t know which version I was running before the bug and if it was updated 2017-06-02 19:17:39 Hum, same problem in a freshly installed chroot x) 2017-06-02 19:20:40 I wonder if this bug happened after a debian upgrade 2017-06-02 19:21:30 But that would be strange that a x client application in debian would break compatibility with alpine edge’s xorg server. 2017-06-02 19:21:36 Am I right? 2017-06-02 19:23:57 somewhat strange, but it could happen, i'm not sure if the X protocol is intended to be stable 2017-06-02 19:24:23 X itself is mostly black magic to all but the devs 2017-06-02 19:24:38 running one distro inside another adds even more complications 2017-06-02 19:24:42 so it's really hard to say 2017-06-02 19:25:43 qman__: even to the devs it's black magic tbh 2017-06-02 19:25:51 IIRC opengl uses an extension 2017-06-02 19:28:22 is that a Caplain i see 2017-06-02 19:34:23 Hello... 2017-06-02 19:34:44 yaho 2017-06-02 19:35:14 Anyone here knows Adapter RAID 51645 (SAS) controller? 2017-06-02 19:35:43 Need to know if one can add multiple drives at once to a RAID5 array (using 'arcconf modify .... ' utility)? 2017-06-02 19:39:25 can't say i ever used it 2017-06-02 19:39:42 it also doesn't seem like we ship arcconf :p 2017-06-02 19:40:23 danci1973 i've got it working with my 71645 2017-06-02 19:40:36 just copied the relevant glibc files from a centos install since arcconf is binary only :/ 2017-06-02 19:42:15 scv: Nice. I have the 'arcconf' working and I added one drive, but reconfiguring is taking a loooong time so I started researching whether it's OK to add all new drives in one go - it's not explicitly documented (or at least I haven't found it). 2017-06-02 19:42:40 scv: i presume the libc6-compat package wasn't enough on its own? 2017-06-02 19:44:07 Shiz: yeah no dice 2017-06-02 19:44:20 didn't really poke at it further 2017-06-02 19:44:41 danci1973: i haven't reconfigured an array from inside the OS so not sure, sorry 2017-06-02 19:44:45 only use it to check array health/etc 2017-06-02 19:44:58 scv: Oh, I see... 2017-06-02 19:45:11 yeah understandable 2017-06-02 19:45:25 just checking if an easier solution was considered 2017-06-02 19:45:27 :p 2017-06-02 19:46:33 Shiz: trust me, copying bins from centos was the *last* resort :> 2017-06-02 19:46:38 :) 2017-06-02 19:47:59 Well, thanks anyway. 2017-06-02 19:48:03 Bye 2017-06-02 19:50:26 https://ghostbin.com/paste/8r2ok 2017-06-02 19:50:27 What does this mean? 2017-06-02 19:50:51 right, so 2017-06-02 19:50:59 around 3.4 (i think?) we switched to libressl 2017-06-02 19:51:07 but we also keep openssl around for the tiny amount of packages that need it 2017-06-02 19:51:22 but qt-dev needs libressl header files 2017-06-02 19:51:32 which conflict with openssl header files you explicitly installed earlier 2017-06-02 19:51:37 solution: apk del openssl-dev 2017-06-02 19:51:39 :p 2017-06-02 19:51:47 hmm. I might need openssl 2017-06-02 19:52:03 is there a reason you can't go with libressl? 2017-06-02 19:52:11 they offer an (almost) functionally identical API 2017-06-02 19:52:14 also, it's just the header files 2017-06-02 19:52:45 openssl-dev and openssl are diff I guess 2017-06-02 19:52:55 yep, -dev is just the header files (and static libraries) 2017-06-02 19:53:01 for if you want to compile stuff locally against openssl 2017-06-02 19:53:13 nope guess not 2017-06-02 19:53:16 libressl works fine 2017-06-02 19:53:18 :) 2017-06-02 19:53:28 now i'm just getting that mesa error 2017-06-02 19:53:43 same error minus the openssl/libressl stanzas? 2017-06-02 19:54:13 yep 2017-06-02 19:54:22 hmm 2017-06-02 19:54:25 try apk upgrade first? 2017-06-02 19:54:28 or apk upgrade -a 2017-06-02 19:54:50 Last time I did that it broke a lot of things 2017-06-02 19:54:57 that uh 2017-06-02 19:55:01 seems unfortunate :p 2017-06-02 19:55:06 what's your /etc/apk/repositories look like? 2017-06-02 19:55:38 https://ghostbin.com/paste/zbroa 2017-06-02 19:56:01 yeah, that's what i figured 2017-06-02 19:56:05 mixing 3.5 and edge is a really bad idea 2017-06-02 19:56:12 exactly it will break things like that 2017-06-02 19:57:02 if you want to go full edge, remove the 3.5 repos 2017-06-02 19:57:34 otherwise, it's wise to explicitly tag the edge repos so apk won't try to install from it unless you tell it to 2017-06-02 19:57:46 e.g. @edge http://nl.alpinelinux.org/alpine/edge/main 2017-06-02 19:57:53 and # apk add somepkg@edge 2017-06-02 19:57:56 can still break, but less chance to 2017-06-02 19:58:41 ah qt-dev installed now 2017-06-02 20:09:56 The qt-dev package is 4.8.6 but qt5-qtbase is 5.6.1 2017-06-02 20:13:39 you probably want qt5-qtbase-dev then 2017-06-02 20:13:41 :P 2017-06-02 21:27:39 "This is a bugfix release of the v3.6 musl based branch..." there's an alpine not based on musl? 2017-06-02 21:28:04 nyet 2017-06-02 21:39:05 shodan45: I suspect it's ncopa's habit to write like that, because in the old days (AL 2.x?) there was non-musl version of AL. but there are not that many non-glibc distros, so refreshing info that AL is musl-based brings no harm 2017-06-02 21:41:00 Shiz: Qt Creator 3.0.1 2017-06-02 21:41:00 Based on Qt 4.8.7 (GCC 4.9.2, 64 bit) 2017-06-02 21:42:04 That's after `apk del qt5-qtbase qt-dev qt-webkit qt-creator && apk add qt5-qtbase qt5-qtbase-dev qt-creator` 2017-06-02 21:46:49 i think that's just the version of qt qt creator is compiled against, not necessarily it's target? 2017-06-02 21:46:55 might be wrong though, i don't know much about qt 2017-06-03 13:09:19 Hello :) I was wondering if anyone has played with Alpine on a Cubietruck ? 2017-06-03 13:16:59 dont *think* so 2017-06-03 13:17:07 but if it can load a linux kernel... 2017-06-03 13:20:25 It can :) but I think it needs some specific patches... Running Armbian atm, I would like to try Alpine on it 2017-06-03 13:22:35 apparently it supports mainline 2017-06-03 13:22:42 https://linux-sunxi.org/Cubieboard#Upstream.2FMainline_kernel 2017-06-03 13:22:48 it just needs a dtb for the cubieboard 2017-06-03 13:23:21 which is upstream 2017-06-03 13:23:41 http://elixir.free-electrons.com/linux/latest/source/arch/arm/boot/dts/sun4i-a10-cubieboard.dts 2017-06-03 13:26:03 nice :) I'll give it a try 2017-06-03 13:26:25 i dunno how dtbs work exactly, but it should be feasible with some tweaking 2017-06-03 13:26:28 (uninformed opinion) 2017-06-03 15:12:44 hi everybody 2017-06-03 15:13:11 how can install pptp client on alpine? 2017-06-03 15:13:41 the ppp package, i presume 2017-06-03 22:54:55 there is a lot ppp packages: ppp-passprompt-2.4.7-r5 ppp-pppoe-2.4.7-r5 ppp-2.4.7-r5 ppp-doc-2.4.7-r5 ppp-minconn-2.4.7-r5 ppp-l2tp-2.4.7-r5 ppp-passwordfd-2.4.7-r5 ppp-daemon-2.4.7-r5 ppp-atm-2.4.7-r5 ppp-dev-2.4.7-r5 rp-pppoe-3.12-r0 ppp-winbind-2.4.7-r5 rp-pppoe-doc-3.12-r0 acf-ppp-0.5.0-r2 ppp-radius-2.4.7-r5 2017-06-03 22:55:33 ppp != pptp 2017-06-04 01:07:43 having a weird issue here 2017-06-04 01:08:18 trying to install virt alpine linux on a vserver of mine, but when I do setup-alpine i get Failed to add partition: Invalig argument 2017-06-04 01:10:03 hmm, that seems problematic 2017-06-04 01:10:10 does your vdisk get detected properly? 2017-06-04 09:52:13 <_spOOn_> hey guys, generally, how long once a package has been accepted into the aports git repo until we should expect to see it in the package repos? 2017-06-04 09:55:30 immediately 2017-06-04 09:55:36 if you're on edge, that is 2017-06-04 09:55:54 or well, ~5-10 mins depending on the package in question 2017-06-04 09:55:58 for the builders to build and push it 2017-06-04 09:56:16 you PR'd zim, right? 2017-06-04 09:57:24 <_spOOn_> yup 2017-06-04 09:57:52 lemme take a lookey 2017-06-04 09:58:13 <_spOOn_> ta 2017-06-04 09:58:20 <_spOOn_> edge 3.6.1? 2017-06-04 09:59:14 3.6.1 is the latest release, edge is the latest... edge :P 2017-06-04 09:59:33 edge is essentially the git head of aports 2017-06-04 09:59:48 buut 2017-06-04 10:00:01 it seems the testing builders have been strangely not building anything 2017-06-04 10:00:08 <_spOOn_> :) 2017-06-04 10:00:16 <_spOOn_> ACTION feels a little less crazy 2017-06-04 10:00:31 i'll ping infra on the matter, thanks for indirectly bringing it to attention 2017-06-04 10:00:37 <_spOOn_> nps 2017-06-04 10:00:52 <_spOOn_> thanks for looksing into it for me :) 2017-06-04 10:01:22 so as soon as THAT gets fixed, it should be in the edge/testing repo :P 2017-06-04 10:01:54 <_spOOn_> thanks again Shiz 2017-06-04 10:02:08 np 2017-06-04 10:02:11 thanks for contributing :) 2017-06-04 12:38:40 <_spOOn_> hey guys, I'm working on packaging up virtualbox for alpine, and have hit a little snag with libxml2-dev 2017-06-04 12:39:44 <_spOOn_> I've got the libxml2-dev alpine package installed, but when I'm testing with #include , Im being told that xmlversion.h doesn't exist... 2017-06-04 12:39:49 <_spOOn_> any suggestions? 2017-06-04 13:25:00 _spOOn_: it's in the package allright, so you probably need to refer to the path where it exists 2017-06-04 13:36:51 <_spOOn_> TBB: have tried setting LD_LIBRARY_PATH, but no joy as yet... 2017-06-04 13:39:39 <_spOOn_> CPLUS_INCLUDE_PATH seems to have done the job :) 2017-06-04 13:39:41 <_spOOn_> thanks TBB 2017-06-04 13:40:23 np 2017-06-04 14:48:35 hello, are there any docs RE installing on ZFS? I haven't been able to locate any 2017-06-04 14:50:52 I searched the wiki and looked through https://wiki.alpinelinux.org/wiki/Setting_up_disks_manually and others, but only another forum question asking for docs/guides on ZFS 2017-06-04 16:06:47 anyone home? 2017-06-04 16:11:26 Sorry, I'm on holiday. 2017-06-04 16:13:28 ok, enjoy 2017-06-04 16:14:37 ACTION looks 2017-06-04 16:14:50 ACTION exits east 2017-06-04 18:51:56 clandmeter: daaaym, it -worked- 2017-06-04 20:34:29 trying alpine linux for the first time today on my hp 8460p; interested in it being systemd-free like devuan but incorporating the grsecurity patches 2017-06-04 20:37:45 excellent hardware detection and set up of my wireless card (debian can't), also very cool that installer has capability to find fast mirror - other distros have that capability after install but as far as I know not in the installer 2017-06-04 20:45:04 set up my internal ssd as sys, "Initializing partitions on /dev/sda...Failed to add partition: Invalid argument" 2017-06-04 20:53:16 this this is what happens by default if there's an existing partitioning scheme. the error message should be more descriptive or the installer should allow installer to overwrite existing partitioning scheme. where can i file a bug report? 2017-06-04 20:54:52 was reported on the forums in january https://forum.alpinelinux.org/forum/installation/setup-alpine-failed-add-partition-invalid-argument 2017-06-04 20:57:23 it would also be nice if there was an option to install lvm on luks too; got to root on my installed system! 2017-06-04 21:03:50 matlock - see https://it-offshore.co.uk/linux/alpine-linux/25-alpine-linux-luks-encrypted-installations 2017-06-04 21:04:03 BitL0G1c thanks! 2017-06-04 21:04:31 ;-) 2017-06-04 21:21:39 getting wired and wireless networking set up now, it's a shame the config didn't come over from the install 2017-06-04 21:23:16 BitL0G1c: nice 2017-06-04 21:24:03 I have to say the alpine linux documentation is top notch, much better than just googling and getting outdated ubuntu answers 2017-06-05 00:30:20 hello, I have a tool which wants `useradd`, should I tell the tool how to use `adduser`, or is there a package to add for `useradd`? 2017-06-05 00:32:59 preferably the first, but there's the shadow package that provides useradd and the like 2017-06-05 00:33:11 the issue is that it can also change some login semantics as it instead uses PAM 2017-06-05 00:33:52 hmm, so what are the potential issues if I use `useradd` from the shadow package? 2017-06-05 00:37:10 it replaces the entire busybox login stack with PAM, which means that in corner cases login semantics may differ where the PAM policies differ from standard busybox stuff 2017-06-05 00:37:37 e.g.: can the super su without requiring a password? by default this was no due some corner case pam policy difference before 2017-06-05 00:38:38 ok, that is interesting. is busybox's login stack better than PAM's, in your or the project's general opinion? 2017-06-05 00:39:53 personally, and i think this goes for many alpine core devs, i'm not a big fan of PAM :P 2017-06-05 00:40:40 IDK a whole lot here, but it seems similar to the situation with openssl 2017-06-05 00:41:05 how do you mean? 2017-06-05 00:41:10 OK, I'll poke the tool's upstream devs and see if I can convince them to care about Alpine's use of busybox and `adduser` 2017-06-05 00:41:48 PAM is a bit overly complex, prolific, probably buggier than many realize, etc.. similar to openssl 2017-06-05 00:42:04 yeah 2017-06-05 00:42:21 adduser is pretty standard for busybox systems, so if upstream provides support for it, more than alpine will profit from it 2017-06-05 00:42:26 think various embedded systems :) 2017-06-05 00:42:37 good point, thanks! 2017-06-05 01:39:00 iirc adduser was a perl wrapper around useradd, why busybox decided to do adduser instead of useradd is confusing 2017-06-05 01:40:37 is there something that needs to be messed with in order to use unshare w/ user namespace on grsec kernels? 2017-06-05 03:12:14 hey guys I would like to try alpine 2017-06-05 03:12:20 bare metal 2017-06-05 03:12:48 onto partition sda5 2017-06-05 03:13:30 the installer does not allow for specific partion, only allows for entire hdd 2017-06-05 03:13:58 how do I go about this without formatting the whole drive 2017-06-05 03:14:55 so grsec kernel seems to block user_ns 2017-06-05 03:15:12 switching to vanilla allowed unshare -U to work 2017-06-05 04:49:17 I think you need to set a sysctl flag on grsec to enable userns. 2017-06-05 04:49:30 I remember running into that before. 2017-06-05 17:53:38 hi, i have a small question. i have an lxc container with alpine. I configure the network interface with lxc. Is there a way to tell alpine no to try to reconfigure it and still consider it done ? 2017-06-05 17:55:00 Lord, what you mean? 2017-06-05 17:55:07 " no to try to reconfigure it and still consider it done ?" 2017-06-05 17:55:16 Setting a static ip address? 2017-06-05 17:55:45 just set inet eth0 manual in /etc/network/interfaces 2017-06-05 17:55:48 that should work 2017-06-05 17:56:50 hoo yes manual was it 2017-06-05 17:56:53 thanks a lot 2017-06-05 17:57:55 fcolista : if you set it to static (instead of manual) but don't provide any address, openrc complains that network isn't satistfied 2017-06-05 17:58:04 so nginx and other services can't start 2017-06-05 17:58:17 Lord, yes..but you didn't specified that 2017-06-05 18:00:08 now he did :p 2017-06-05 18:06:04 i raised a bug about this a while ago 2017-06-05 18:06:32 need net breaks lxc containers that set ips in the lxc config 2017-06-05 18:10:55 (you should be dropping net_admin in containers) 2017-06-05 18:11:23 it seems like inet eth0 manual is a sensible default config for minirootfs almost 2017-06-05 18:11:25 :p 2017-06-05 18:14:48 also having networking to the runlevels by default in lxc containers is not good for this same reason 2017-06-05 18:15:46 although this only delays the container starting a little 2017-06-05 18:16:39 Lord - you probably should: rc-update del networking inside the container too 2017-06-05 18:17:07 setting the ip's in the lxc config starts networking 2017-06-05 18:18:41 good idea ;-) 2017-06-05 18:19:55 enabling logging in /etc/rc.conf is helpful too 2017-06-05 18:24:14 Lord - these capabilities can be safely dropped too https://hastebin.com/zuwebubabi.go - if you drop net_admin you will need to configure iptables in a script in lxc.hook.pre-mount 2017-06-05 18:29:50 configure iptables to do what ? 2017-06-05 18:38:18 Lord - as well as having an iptables firwall on the host - you can also have an additional firewall on the container 2017-06-05 18:38:47 haaa ok. no i don't need one on the guest 2017-06-05 18:39:00 the host will drop everything before it reaches the guest 2017-06-05 18:40:07 yes - but not things out from the guest 2017-06-05 18:53:42 the guest will just be a web reverse proxy 2017-06-05 18:54:01 it will be readonly with no way to access it 2017-06-05 18:54:07 only nginx 2017-06-05 19:05:39 Lord - hiawatha is a very good reverse proxy - it filters out more attacks by default 2017-06-05 19:12:04 i won't have attacks, it's just static content 2017-06-05 21:19:08 how can i set my network interface to get an ip6 address automatically? 2017-06-05 21:23:35 Keridos: inet6 dhcp, i think? 2017-06-05 21:23:39 it also depends on your provider 2017-06-05 21:23:48 sorry 2017-06-05 21:23:56 iface inet6 dhcp 2017-06-05 21:23:58 rather 2017-06-05 21:24:09 but it depends how your provider/upstream does ipv6 2017-06-05 21:24:23 for stuff like prefix delegation you need a more specialized implementations like dhcpcd in my experience 2017-06-05 21:30:24 apparently my provider just gives me a 64 subnet 2017-06-05 21:30:30 good 2017-06-05 21:30:35 i just set an ip address statically and it seems to work 2017-06-05 21:30:48 even left out the default gateway 2017-06-06 07:00:08 hi, i'm trying to compile open lighting architecture on alpine, but it's complaining about a missing libuuid. i've installed libuuid, but the configure script fails checking for base_uuid and ossp_uuid 2017-06-06 07:16:48 so, uuid.h is not included in that package. is there some sort of archlinux's pkgfile equivalent for alpine? 2017-06-06 07:17:35 <_ikke_> http://pkgs.alpinelinux.org lets you search for package contents 2017-06-06 07:18:11 i have to select a package for that first... 2017-06-06 07:18:24 and if i dont know which package contains that file, it doesnt help at all 2017-06-06 07:18:25 furthermore, .h's are not necessarily in the main package, they can be in the -dev subpackage 2017-06-06 07:18:26 but google did 2017-06-06 07:18:34 util-linux-dev contains is 2017-06-06 07:18:39 there is no libuuid-dev package 2017-06-06 07:18:41 there's both package name and package contents search 2017-06-06 07:18:45 which was the first thing that i've looked for 2017-06-06 07:43:52 Good day, noble Alpine people, I came in peace and with a question in my pocket. 2017-06-06 07:44:28 "take me to your leader" ? ;) 2017-06-06 07:44:50 How can I install certbot on Alpine https://certbot.eff.org/ 2017-06-06 07:45:01 That's my question. 2017-06-06 07:45:37 Never mind, found it in the package repository. 2017-06-06 07:46:08 https://pkgs.alpinelinux.org/packages?name=certbot*&branch=&repo=&arch=&maintainer= this site is often a good start for finding the packages you might need 2017-06-06 07:46:40 describing your problem is half the solution, except when it works itself out by magic. =) 2017-06-06 11:16:50 Why I can't install php7-gmp package? apk add php7-gmp tells missing 2017-06-06 11:17:05 ERROR: unsatisfiable constraints 2017-06-06 11:23:33 <_ikke_> pihpah: did you add the correct repo? 2017-06-06 11:24:43 <_ikke_> You need the community repo for 3.5, 3.6 or edge 2017-06-06 11:34:33 echo '@edge-community http://nl.alpinelinux.org/alpine/edge/community' >> /etc/apk/repositories 2017-06-06 11:34:50 and then apk -U add php7-geoip@edge-community 2017-06-06 11:34:53 Still nothing 2017-06-06 11:38:08 <_ikke_> there is no package called php7-geoip 2017-06-06 11:39:52 Yeah, I see 2017-06-06 11:40:00 Weird 2017-06-06 11:40:21 pecl install geoip also fails to compile it, looks like I am in trouble 2017-06-06 12:44:35 hmm, interesting. one of our internal apps that builds fine no longer installs build time deps and fails in the current stable... 2017-06-06 12:47:08 fails right after analyzing dependencies, doesn't install them with -r but manually installing them results in a successful build 2017-06-06 13:56:18 anyone here know how i can do something like the bash <<< to assign a string to standard input in ash? 2017-06-06 13:57:46 <_ikke_> heredoc is what you are looking for, but not sure if ash supports that 2017-06-06 13:58:53 <_ikke_> https://github.com/mozilla-b2g/busybox/blob/master/shell/ash_test/ash-heredoc/heredoc.tests 2017-06-06 13:58:59 ash supports heredoc just fine 2017-06-06 13:59:02 it's standard posix 2017-06-06 13:59:12 Keridos: cmd < $str or whatever 2017-06-06 13:59:18 EOF 2017-06-06 13:59:30 <_ikke_> right, with two <<, not 3 2017-06-06 14:00:21 ok thanks 2017-06-06 14:00:30 meh certbot docker image has a segfault -.- 2017-06-06 14:04:09 I guess most sh implementations support heredocs, not sure but I think I even saw the requirement in some POSIX compliance document 2017-06-06 14:04:56 yeah it's posix 2017-06-06 14:06:11 http://pubs.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html 2017-06-06 14:06:13 2.7 2017-06-06 14:15:52 apparently the official docker certbot image segfaults on alpine linux o.o 2017-06-06 14:16:22 cannot use certbot via docker, it just segfaults for me whenever i try to call the certbot command 2017-06-06 14:17:52 anyone here having success with certbot in a docker container? 2017-06-06 14:18:17 immunity:~# certbot --version 2017-06-06 14:18:19 certbot 0.14.0 2017-06-06 14:18:22 works on my normal alpine install at least 2017-06-06 14:18:53 i am running it via docker 2017-06-06 16:48:29 Anyone know if the Generic Arm spin will run on a ODROID-XU4? 2017-06-06 16:51:43 I have no idea what's going on, but when i wget a certain file from within my alpine linux container, then wget just stops with no errors. I can wget that specific file from a normal ubuntu server, so it's not the remote being messed up (I think) 2017-06-06 16:52:09 I have also tried downloading another file from my own webserver, which worked fine from within alpine, so is there any way i can debug? 2017-06-06 16:53:32 alec_, userland should work, kernel/bootloader may need enabling config 2017-06-06 16:53:36 Normally i would enable -v, but that flag is apparently not supported 2017-06-06 19:14:40 I want to use the qt-creator package, but it's in the "edge" branch of testing. Should I just do an install using all edge repos? 2017-06-06 19:14:55 will using testing repos work if I want to install that? 2017-06-06 19:42:09 luxio: change to edge/main, edge/community and add edge/testing 2017-06-06 19:42:16 then test the qt-creator package 2017-06-06 19:42:25 if it works, then let us know so we can move it to community 2017-06-06 19:45:18 ok 2017-06-06 20:50:30 hm is gliderlabs/docker-alpine not the base image for docker containers with alpine? 2017-06-06 20:50:37 cant seem to pull it 2017-06-06 21:04:03 ncopa, I'm on edge now and audio isn't working 2017-06-06 21:07:14 something about gstreamer 2017-06-06 21:11:04 brb 2017-06-06 22:08:16 if i want a docker container to have access to the groups and users of the main system, what would I do? 2017-06-06 23:00:16 hm in the offial dovecot builds for alpine linux, why is the passdb driver pam not compiled in? 2017-06-06 23:08:09 because we don't ship pam by default/wish to encourage pam 2017-06-06 23:08:32 how can I get sasl auth work with postfix then? 2017-06-06 23:08:36 cyrus sasl? 2017-06-07 19:54:11 Howdy, anyone else running kvm guests without polkit? Having a pita to get it to run on alpine with a bunch of strange permission errors (even as root). Any suggestions welcome at this point :) 2017-06-07 20:01:10 bulldooozer: sharing the errors would help... 2017-06-07 20:35:43 tdtrask: password reset email fixed 2017-06-07 20:50:58 nangel: thanks 2017-06-07 20:51:11 ncopa had previously sent me a reset (somehow) 2017-06-07 21:41:20 fcolista: I keep missing you, I see you wrote the mkimage wiki page. Have you seen any examples of building a PXE image? 2017-06-07 21:43:12 On a bouncer now so if you drop me a reply I will see it :) 2017-06-08 04:40:27 Hi folks. Does anyone have any insight into what'll happen around GRSecurity in Alpine now that it's paywalled? 2017-06-08 04:42:08 <_ikke_> auraltension: It's still a bit of a status quo 2017-06-08 04:56:54 auraltension: predictions: the #linux-hardened bunch will continue to spend the next few months completely failing to understand the implications of the patches they are isolating, meanwhile alpine will evaluate the situation when the next LTS release series happens 2017-06-08 04:59:12 auraltension: we have yet to see any reason to engage the linux-hardened people, and overall we're pretty ok with 4.9 LTS right now 2017-06-08 07:57:35 we have 4.9 until january 2019, and minipli's spending time to backport the last patch 2017-06-08 07:57:57 (or is it forward-port?) 2017-06-08 07:58:24 we have our own forward-port anyway 2017-06-08 07:59:25 <_ikke_> Do we want to stay on 4.9 for 2.5 years? 2017-06-08 08:01:48 hope not 2017-06-08 08:07:16 it's 1.5 years 2017-06-08 08:07:22 but yeah 2017-06-08 08:10:31 <_ikke_> right 2017-06-08 08:16:07 well, we're definitely supporting 4.9 for the next almost-two-years 2017-06-08 08:16:11 because of the 3.6 release cycle 2017-06-08 09:20:52 Hello... I have a bit of a problem with time in system logs - while the clock is set to my local timezone, time in log files is 2 hours behine (UTC)... 2017-06-08 09:20:57 Is there a way to fix that? 2017-06-08 09:23:35 Shiz, you have a customer ;-) 2017-06-08 09:23:53 <_ikke_> /etc/conf.d/syslog contains 'SYSLOGD_OPTS="-Z"' 2017-06-08 09:24:15 depends which version of alpine afaik 2017-06-08 09:25:55 This is 3.5.2 and syslogd doesn't accept '-Z'... 2017-06-08 09:26:22 None of the syslogd options seem to be time related... 2017-06-08 09:28:01 danci1973, its not supported on 3.5 2017-06-08 09:28:08 i think is edge and 3.6 that has it. 2017-06-08 09:28:18 Shiz added it recently 2017-06-08 09:28:27 I see... 2017-06-08 09:28:30 Hm... 2017-06-08 09:29:00 How risky would an upgrade from 3.5.2 to 3.6 be on a critical and remote router? :) 2017-06-08 09:29:40 i would say acceptable risk which you have to decide yourself of course. 2017-06-08 09:30:57 but a real critical device without direct access... thats always a risk. 2017-06-08 09:31:45 I guess I'll make me a 'clone' as a VM and try it there first... 2017-06-08 09:31:58 maybe there are other sysloggers that support time offset. 2017-06-08 09:34:44 i think we could possibly also backport it to 3.5 2017-06-08 09:35:08 not sure how ncopa or Shiz feel about that. 2017-06-08 22:06:12 Hi all! Quick question, there doesn't appear to be any support in AL for puppet, chef, salt, or any other configuration management tool, yes? My various searches didn't come up with anything, and I didn't see them in the packages list.. 2017-06-08 22:17:02 thabersack: there's at least partial support in ansible 2017-06-08 22:17:19 and rudder stalled for now:/ 2017-06-08 22:17:26 oh 2017-06-08 22:17:52 and cfengine has AL support since 5 years or so 2017-06-08 22:18:15 darkfader: thanks for the reploy 2017-06-08 22:18:16 reply 2017-06-08 22:18:35 still a bit thin on the range of supported tools for sure 2017-06-08 22:18:56 I've also been looking into salt ssh, which *may* let me enforce config settings from another system, no agent required.. https://docs.saltstack.com/en/latest/topics/ssh/ 2017-06-08 22:19:06 yeah 2017-06-08 22:19:34 gonna had to try and see how well it works 2017-06-08 22:20:03 definitely. I'll write a blog post of my findings (Which are hopefully good!) 2017-06-08 22:32:45 wait, AL does support salt! https://pkgs.alpinelinux.org/package/v3.6/community/x86_64/salt 2017-06-08 22:33:30 better link https://pkgs.alpinelinux.org/packages?name=salt*&branch=v3.6&repo=&arch=x86_64&maintainer= 2017-06-09 01:21:13 kanini: thanks for the response re grsec. 2017-06-09 10:53:40 hi! im currently trying to build kodi on a raspberry pi 3 running alpine. but i get an error "error: invalid use of incomplete type 'class CEGLNativeType'" .. not sure whether its a kodi bug or something wrong with the egl libs installed. someone experienced with similar problems? 2017-06-09 12:19:08 Hello! 2017-06-09 12:23:27 hello 2017-06-09 12:23:37 I've installed Alpine on my Raspberry and I have only one question so far - how can I make screen not go sleep every like 5 minutes? 2017-06-09 12:32:01 gregman i think xset is your friend 2017-06-09 12:32:15 (assuming you run xorg) 2017-06-09 13:00:23 gregman_ what desktop environment and protocl (X11/Wayland)? 2017-06-09 13:03:17 hi 2017-06-09 13:03:43 hi 2017-06-09 13:22:57 IAMB3NW: just clean install/terminal only 2017-06-09 13:28:56 gregman_So you're just using TTY1 for example and want to stop the screen sleeping? 2017-06-09 13:31:26 нуы! 2017-06-09 13:31:28 yes! 2017-06-09 13:32:01 https://unix.stackexchange.com/a/32043 2017-06-09 13:32:20 Thank you IAMB3NW ! 2017-06-09 13:32:22 It's set by the kernel, so it's not actually distro specific 2017-06-09 13:32:32 It's set by the kernel, so it's not actually distro specific 2017-06-09 13:32:41 yeah, I tries xset and stuff - didn't work 2017-06-09 13:32:42 You're welcome :) 2017-06-09 13:32:50 *tried 2017-06-09 13:33:20 how about this? https://unix.stackexchange.com/a/23636 2017-06-09 13:34:04 It seems that setterm doesn't exist 2017-06-09 13:34:21 it's in util-linux 2017-06-09 13:34:58 wow thanks! didn't know that 2017-06-09 13:35:19 https://pkgs.alpinelinux.org/contents?file=setterm&path=&name=&branch=v3.6&repo=&arch=x86_64 this site is your friend 2017-06-09 13:36:30 yeahs pkg search is amazing 2017-06-09 13:39:45 evening. never had this issue before: man page for nmap is in german... help? 2017-06-09 13:46:05 schmakk, have you set your locale? 2017-06-09 13:46:31 if i recall correctly, i used the virtual image, which appears to now have locale stuff 2017-06-09 13:47:44 what does env | grep -e LANG -e LC say 2017-06-09 13:48:00 nothing 2017-06-09 13:48:27 Hmm, man uses the "LANG" env by the looks of it 2017-06-09 13:48:39 does using the locale flag fix the issue? 2017-06-09 13:48:49 --locale= 2017-06-09 13:51:10 trying a few, not sure what format it accepts 2017-06-09 13:51:36 LANG=en_US.utf8 man nmap 2017-06-09 13:54:11 tried setting $LANG and calling man nmap --locale=en_US.utf8 - still german 2017-06-09 13:55:25 docs are just from apk add nmap-doc - could there be something package related? 2017-06-09 13:55:35 schmakk ohh there's a simple fix for this! 2017-06-09 13:55:41 Learn German :) 2017-06-09 13:55:45 oh you 2017-06-09 13:56:28 ill just google translate every single word till i know german at a techical level 2017-06-09 13:57:21 at least the man pages will be a lot shorter 2017-06-09 13:57:38 German has a word for every sentence it's possible to make 2017-06-09 14:00:46 would it be a decent assumption that i should have gone with the standard install instead of the virtual? I really just wanted a shell in a vm for various work stuff 2017-06-09 14:01:51 schmakk: show me what you try to find out 2017-06-09 14:02:27 right now i'm curious as to why the nmap man page is in german instead of english :) 2017-06-09 14:02:54 i figured i'd learn a bit more linux while figuring that out 2017-06-09 14:09:24 heh, wonder how normal this is... dumping a 500 GB disk with dd to another disk, both USB3; it's taken almost 18 hours now 2017-06-09 14:10:21 being bzip2'd on the fly is in the mix too but surely it can't take this long, even if it's mostly busybox doing the job 2017-06-09 14:10:54 bzip2 is really long 2017-06-09 14:10:56 bzip2 is slow and not as efficient 2017-06-09 14:11:48 look at lzop if you really require compression for this kind of size 2017-06-09 14:12:08 or maybe snappy but I do not know if that's easily packaged in alpine :) 2017-06-09 14:12:58 for modern CPUs gzip is ok-ish (with its default -6), not that much CPU-intensive. if I have a few more spare cycles, I go with xz -1 2017-06-09 14:13:05 I also could've gone with gzip as compression factor is not such a big deal, or even xz as it multithreads nicely and the system in question has 16 cores 2017-06-09 14:14:02 przemoc: still pretty slow compared to usb3, usually :) 2017-06-09 14:14:30 TBB: do hdparm -t /dev/sdX on both devices one by one and later simultaneously. you may also have slow devices or crappy controller 2017-06-09 14:16:01 at this moment it seems the package is growing so slowly that the process has reached the part of disk that's empty or something 2017-06-09 14:16:35 there can be 0s on unused space 2017-06-09 14:16:41 yup 2017-06-09 14:17:14 ah, I thought that you finished that dump, but you're still doing it? 2017-06-09 14:18:08 so better to not run hdparm to not slow it down even further ;) 2017-06-09 14:18:55 yeh, still running but as I finally got a sign of this actually ending soon I'm okay with waiting for a bit longer 2017-06-09 14:20:38 aaand it's finished 2017-06-09 14:20:53 Segmentation fault 2017-06-09 14:22:57 About a lifetime ago in 3.1, there was an nbtscan package in testing on armhf. Is there any chance someone could bring this to x86_64, please? 2017-06-09 14:24:34 file package request on bugs.alpinelinux.org 2017-06-09 14:26:08 Cheers 2017-06-09 14:27:56 Also, reference https://bugs.alpinelinux.org/issues/5011 ... I'm here! 2017-06-09 14:28:26 clandmeter: I just read your message on the above bug 2017-06-09 15:31:40 Has anyone got any experience with the mkimage scripts? 2017-06-09 18:16:09 I try to attach a bluetooth loudspeaker to a RPi3 but I cannot get bluetooth working. The wiki says I need to attach the serial port. But 'hciattach' command seems to be deprecated now. What is the correct way to access the serial port now? 2017-06-09 18:30:09 hey all, is there a kernel howto for alpine? https://wiki.alpinelinux.org/wiki/Custom_Kernel doesnt seem to have much content 2017-06-09 18:31:31 also, does anyone know why `CONFIG_MLX5_CORE_EN=y` for aarch64 but unset for x86_64? https://git.alpinelinux.org/cgit/aports/tree/main/linux-vanilla/config-vanilla.aarch64?h=3.5-stable and https://git.alpinelinux.org/cgit/aports/tree/main/linux-vanilla/config-vanilla.x86_64?h=3.5-stable 2017-06-09 18:31:43 consider why you require a custom kernel 2017-06-09 18:31:46 as you probably don't need one 2017-06-09 18:32:00 if you're experiencing a bug, it is better to just fix the bug. 2017-06-09 18:32:54 as for CONFIG_MLX5_CORE_EN unset on x86_64, we were unaware of that hardware being used on x86_64. 2017-06-09 18:33:19 kaniini: it seems we need CONFIG_MLX5_CORE_EN=y in order for our mellanox NICs to show up. dmesg/lspci report fine but we have no entries being created under /sys/class/net/ 2017-06-09 18:33:35 right 2017-06-09 18:34:22 I'll submit a bug report in the meantime but hoping to fix this up myself as soon as I can 2017-06-09 19:03:44 bug report is in 2017-06-09 19:04:42 in the meantime when I try to compile, I'm seeing " No rule to make target `arch/x86/entry/syscalls/syscall_32.tbl', needed by `arch/x86/entry/syscalls/../../include/generated/asm/syscalls_32.h' " I'm probably doing something dumb or missing a dep 2017-06-09 19:04:46 or both :) 2017-06-09 19:16:37 i've resorted to trying to build from alpine-sdk with aports 2017-06-09 19:59:49 Guys, anyone can control network via nm-applet without getting authorization errors? 2017-06-09 23:20:16 Small bug in the awesomewm package, it's built against/requires dbus to start, but the package doesn't list it as a dependency. 2017-06-10 01:43:17 How do I find which -doc package a given manpage is in? For example, I want the manpage for "interfaces(5)", but it's not in anything that seemed obvious to me after going over "apk search doc" 2017-06-10 01:44:00 I also tried "apk search interfaces.5.gz" 2017-06-10 01:55:02 I also tried figuring out which package /etc/network/interfaces belongs to, but both "apk info -W" and "apk info -P" failed 2017-06-10 01:58:33 I also tried this golang tool (github.com/jessfraz/apk-file), but it shows only shorewall packages 2017-06-10 02:09:10 TML: have you tried: apk info --who-owns /usr/share/man5/interfaces.5.gz 2017-06-10 02:09:30 could not find owner package 2017-06-10 02:35:04 TML: it seems that /etc/network/interfaces isn't "owned" by any package, however it is read by the openrc networking init.d script 2017-06-10 03:06:30 TML: The networking init.d script calls the 'ifup' BusyBox command which expects '/etc/network/interfaces'. So unless BusyBox provides the 'interfaces(5)' manpage, I don't think there is a package that provides it. 2017-06-10 03:07:47 run /sbin/setup-interfaces 2017-06-10 05:31:14 BitL0G1c: Thanks for the script helper, but I'm actually looking for documentation on the format and the available options, so I can make my YAML generate correct files :) 2017-06-10 05:31:44 Don't really want to sit there and toggle every possible option and compare output, unless there really is not a better method 2017-06-10 05:42:40 but maybe the source code for /sbin/setup-interfaces will be more helpful than the busybox source 2017-06-10 07:41:11 TML: there's no man pages for interfaces that i know of, but it mostly follows debian format 2017-06-10 08:23:35 Shiz: yeah, the format seemed similar, but some of the options our YAML generates for Debian don't seem to parse correctly by Alpine. That's what sent me looking. 2017-06-10 14:30:17 Also, reference https://bugs.alpinelinux.org/issues/5011 ... I'm here! 2017-06-10 14:30:26 Oops, disregard 2017-06-10 14:30:31 Though I am still here 2017-06-10 16:20:38 hi. Short question: what is the alpine package for the "script" command? (http://man7.org/linux/man-pages/man1/script.1.html) - its pretty hard to google that or find it, because of obvious reasons. ^^ 2017-06-10 16:22:09 <_ikke_> wmchris: https://pkgs.alpinelinux.org/contents?file=script&path=&name=&branch=&repo=&arch= 2017-06-10 16:22:48 util-linux. Thank you. Didn't know this webpage 2017-06-10 16:23:09 Very handy, indeed. 2017-06-10 16:23:13 <_ikke_> wmchris: That's why I pointed to it 2017-06-10 16:23:22 <_ikke_> :) 2017-06-10 17:23:57 hi, somebody here has experience with embebed systems? 2017-06-10 17:25:11 and if somebody try UpCore device with alpine 2017-06-10 19:28:07 Classsic: i've got two up-boards 2017-06-10 19:28:20 they're not embedded hw, just fun toys for tinkering 2017-06-10 19:28:28 no matter what website says :) 2017-06-10 19:28:40 i think they'll run alpine just fine 2017-06-10 19:29:03 had the stock debian thing on them, install was very slow, after that perf was good 2017-06-10 19:29:12 they had some stability issues with the emmc driver 2017-06-10 19:29:19 that could be the biggest issue for you 2017-06-10 20:39:02 thanks darkfader, that' s sad, because this board look perfect for my project 2017-06-10 22:27:59 Classsic: try it, i didn't mean to say it can't be ok 2017-06-11 14:15:25 hello! I am testing ZFS on alpine. I `apk add zfs`, and then I see `zfs -h` tells me to load the kernel module... so I attempt to do that with: `/sbin/modprobe zfs`, and I get: `modprobe: FATAL: Module zfs not found in directory /lib/modules/4.9.30-0-virthardened` 2017-06-11 14:17:32 did you recently update your kernel without rebooting? 2017-06-11 14:18:32 that's possible. `apk -L info zfs` does not list any kernel modules in `/lib/modules` 2017-06-11 14:18:42 I guess I should try rebooting? 2017-06-11 14:18:44 the kernel module is not in that package 2017-06-11 14:19:00 `zfs-hardened-4.9.30-r0` ? 2017-06-11 14:19:20 https://pkgs.alpinelinux.org/contents?file=zfs.ko&path=&name=&branch=v3.6&repo=&arch=x86_64 2017-06-11 14:19:45 so indeed it's not a kernel mismatch thing but just a missing package 2017-06-11 14:20:06 ah, yes! thanmks 2017-06-11 14:36:47 clandmeter: zfs-{hardned,vanilla} have wrong package spec in install_if 2017-06-11 14:37:35 clandmeter: s/-$_kver/=$_kver/ 2017-06-11 14:41:59 hi, can one switch back and forth between the standard and vanilla versions post-install? 2017-06-11 14:42:41 sure, you can change the kernel package and install/remove whatever 2017-06-11 14:47:02 Bun: thanks 2017-06-11 16:28:43 what is the difference between hardened and virthardened? can I use ZFS with virthardened? 2017-06-11 16:37:58 virthardened is probably aimed at VMs 2017-06-11 16:40:05 yea, I targeted virthardened during install, but I also see hardened. I wonder if that came from installing `zfs-hardened`. When I search APK for `zfs`, I see packages for vanilla and hardened, but not virthardened. Is this OK, or is zfs not available on virthardened? 2017-06-11 16:45:08 i followed instructions from https://wiki.alpinelinux.org/wiki/PXE_boot 2017-06-11 16:45:19 but booting stops at "Mounting boot media" 2017-06-12 02:04:21 I'm trying to build a custom alpine-extended iso but I don't see where the APKs for Docker are 2017-06-12 02:04:41 oh wait, wrong directory, found it 2017-06-12 04:47:24 question about the flatpak package - does that allow running random flatpaks built on other distros, or only musl-linked flatpaks? 2017-06-12 04:47:34 if the former, that's pretty neat 2017-06-12 05:32:45 hi, how can I get a service to start at boot, considering that I don’t have an OpenRC script but only an old-school sysvinit script? 2017-06-12 05:34:08 rewrite it 2017-06-12 05:34:42 tidux: …thanks 2017-06-12 05:34:48 any way I can run it without rewriting it? 2017-06-12 05:34:55 it’s a rather complex one 2017-06-12 05:36:56 also, documentation to write openrc scripts is awful… way worse than systemd :) 2017-06-12 08:01:27 Hi all. I recently updated postfix (3.2.0), and now TLS connections don't work with an error of no shared ciphers. I'm using ECC certs/keys. The same certs/keys worked previously, and also work fine with nginx, prosody, etc. Anyone else having any issues with latest postfix TLS? 2017-06-12 08:02:09 AFAIK I didn't have any changes in the configs from the previous version of postfix either. 2017-06-12 08:03:54 Also, if it matters, I get the error even when connecting from localhost using openssl s_client, so it should be the same library that postfix is itself built against (and thus *should* support the same ciphers). 2017-06-12 08:30:55 flux_control: check so both service and client uses same version of the protocol by config. 2017-06-12 08:31:25 flux_control: and restart service maybe. 2017-06-12 08:44:17 I've tried restarting the service, and as I've tried connecting directly with openssl s_client I also know that the protocol version is the same. 2017-06-12 08:45:08 Also, the output from s_client shows that I get TLS 1.2 as the protocol even though the handshake failed due to the cipher problem. 2017-06-12 08:49:18 does smtpd_tls_loglevel = 2 show anything useful? I assume openssl ciphers -v DEFAULT | grep ECDSA shows stuff 2017-06-12 08:49:21 This is the error in my mail log: 2017-06-12 08:49:23 warning: TLS library problem: error:140270C1:SSL routines:ACCEPT_SR_CLNT_HELLO_C:no shared cipher:ssl_srvr.c:1024: 2017-06-12 08:49:26 2017-06-12 08:50:25 Yes, my openssl (libressl, actually) supports ECDSA. And the other services (nginx, dovecot, etc.) work just fine with no TLS errors. 2017-06-12 08:52:28 have you tried explicitely configuring the supported cipher list, at least to see if that works as expected? 2017-06-12 08:53:45 Not yet. From what I've read, that shouldn't be necessary as recent postfix is configured to support elliptic curves out of the box, but I'll give it a try. 2017-06-12 08:53:59 indeed, and I don't see any alpine specific patch to change the ciphers 2017-06-12 08:54:07 Bun: I take it you haven't had any problems with postfix + TLS recently? 2017-06-12 08:54:16 nope, but then I also don't use ECDSA 2017-06-12 08:54:20 ah 2017-06-12 08:54:22 RSA? 2017-06-12 08:54:27 yeah 2017-06-12 08:55:08 It was working fine for me for many months, and then I recently did an apk upgrade and suddenly I'm having problems. 2017-06-12 08:55:59 It's hard to tell if that's something from the newer postfix, or something in how the alpine postfix got built/linked against libressl (or, possible, something even in libressl itself?). 2017-06-12 08:56:18 I have doubts that it's a libressl problem, though, as other stuff works fine with it. 2017-06-12 08:56:45 seems very strange indeed 2017-06-12 09:00:45 The base alpine docker image (and perhaps the normal host install?) contains a postgres user by default. Is there a way to delete til user? 2017-06-12 09:03:30 flux_control: I just got an ECDSA cert, let me see 2017-06-12 09:05:16 well, it definitely does not work on postfix. 2017-06-12 09:06:07 oh nevermind, it does. 2017-06-12 09:06:26 Bun: If it matters, when I connect to my other ports/services (i.e., openssl s_client -connect -starttls xmpp localhost:xmpp-client), I get no errors and it shows my cipher as ECDHE-ECDSA-CHACHA20-POLY1305. 2017-06-12 09:06:46 I used openssl s_client -connect wan.awoo.io:587 -starttls smtp 2017-06-12 09:06:51 If I run openssl ciphers -v ECDSA, that cipher is definitely in the output 2017-06-12 09:07:05 Bun: what cipher did it show? 2017-06-12 09:07:17 ECDHE-ECDSA-AES256-GCM-SHA384 2017-06-12 09:07:34 I have no TLS options specified beyond key/cert obviously and smtpd_tls_security_level=encrypt 2017-06-12 09:07:37 Hmm... I wonder if there's something blocking chacha due to a recent change? 2017-06-12 09:08:07 I am looking for old donated hardware (not more than 4-6yrs), anything that can be useful for building AL, if someone comes across or can mediate the source would be helpful. I am located in India (I can bear some import cost if its reasonable). email: vkrishn@insteps.net, thanks 2017-06-12 09:09:05 I configured ECDHE-ECDSA-CHACHA20-POLY1305 on nginx, that worked at least 2017-06-12 09:10:12 smtpd_tls_mandatory_ciphers = ECDHE-ECDSA-CHACHA20-POLY1305 does not work 2017-06-12 09:10:28 oh that's not how the syntax works, woops 2017-06-12 09:10:55 Yeah ;) 2017-06-12 09:11:16 Under nginx you got chacha as the cipher in use via openssl s_client? 2017-06-12 09:11:40 Or just that nginx didn't error out on restart when you listed it in the config file? 2017-06-12 09:11:52 I can connect fine if I specify -cipher ECDHE-ECDSA-CHACHA20-POLY1305 to postfix though 2017-06-12 09:12:12 Great, so it's somehow just me... 2017-06-12 09:12:18 I specified it in the nginx conf and s_client negotiated it for me 2017-06-12 09:12:28 I tried commenting out all my settings except for the 3 you mentioned and I had no change 2017-06-12 09:12:35 Oh, what do you have in master.conf? 2017-06-12 09:12:40 Maybe there's something there... 2017-06-12 09:12:52 http://sprunge.us/iFMd 2017-06-12 09:14:50 I don't see anything there that's radically different 2017-06-12 09:15:17 I have wrappermode=no for submission, but I think that's the default anyway. 2017-06-12 09:15:57 Either way it shouldn't matter as that would only affect the need for starttls or direct s_client (neither of which work for me anyway). 2017-06-12 09:16:46 I can try to connect to you if you want 2017-06-12 09:17:48 Also, that would only matter for the submission port. For pure port 25 you (and I also) have nothing in master.conf beyond just enabling the port. 2017-06-12 09:18:15 Bun: You can try, but I don't think it would matter. The server is mail.bitnoise.net. 2017-06-12 09:18:39 wan.awoo.io? 2017-06-12 09:18:48 yep 2017-06-12 09:19:14 Yeah, same "no shared ciphers" line in my log. :( 2017-06-12 09:20:22 welp, very strange 2017-06-12 09:20:35 is this on 3.6 or edge? 2017-06-12 09:21:42 edge 2017-06-12 09:22:22 I believe the version for postfix is the same in both currently though, yes? 2017-06-12 09:23:04 https://pkgs.alpinelinux.org/packages?name=postfix&branch=&repo=&arch=x86_64&maintainer= 2017-06-12 09:23:20 (so no) 2017-06-12 09:23:54 Ah 2017-06-12 09:24:07 For some reason I thought latest stable was also at 3.2.0 2017-06-12 09:24:31 My bad 2017-06-12 09:24:31 you can try a downgrade with package pinning I guess 2017-06-12 09:24:37 Yeah 2017-06-12 09:24:51 Your postfix is 3.1.4 then? 2017-06-12 09:25:06 yes, I'm on v3.6 2017-06-12 09:25:21 Right 2017-06-12 09:27:50 Now I realize, I don't remember how to downgrade a pkg in alpine. :) 2017-06-12 09:27:51 hehe 2017-06-12 09:28:43 Ah, I forgot the suffix on it, nvm. Downgraded. 2017-06-12 09:29:54 Yup, works 2017-06-12 09:30:05 So there's definitely something funny with the 3.2.0 version then. 2017-06-12 09:30:11 Bun: Thanks a ton! :-D 2017-06-12 09:30:19 :) 2017-06-12 09:30:39 what's your opinion on ECC in the mail world though? 2017-06-12 09:30:58 I feel like most people don't even support TLS at all, so ECC is going to be a pain as well 2017-06-12 09:31:53 I'm only using it for incoming client connections, and I have fairly few of those so I can push that onto my users. 2017-06-12 09:32:08 From what I've had, I've had no problems with it. 2017-06-12 09:32:29 guess that's true 2017-06-12 09:32:52 If you have a large site with all kinds of clients (especially non-standard ones like Outlook tends to be), that might be more of a nightmare. 2017-06-12 09:33:52 I think it doesn't really make sense to do server-server TLS anyway, since smtp is inherently insecure as a protocol. It was never meant to do anything more than plain-text transfers. 2017-06-12 09:34:24 For incoming clients, having TLS->SMTP_AUTH just helps to cut down on the number of abusers really, rather than securing anything. 2017-06-12 09:34:25 it's pretty much impossible to enforce anyway 2017-06-12 09:34:29 Yup 2017-06-12 09:35:08 I wanted to make a tool to learn domains that do TLS and then remember them to force them.. but then my bank randomly sends me mails with no TLS support, stuff like that 2017-06-12 09:35:13 great fun! 2017-06-12 09:35:37 I have postfix->dovecot(ldap) for incoming auth, and since users are submitting LDAP passwords there it makes sense to encrypt that as much as possible. 2017-06-12 09:36:15 Yeah, dealing with external sites basically requires that you forego any TLS altogether. Take it when you can, but those are lucky cases. 2017-06-12 09:36:57 If you want security/encryption offsite, either encrypt the mails with s/mime or gpg, or use something other than SMTP entirely for the messages. 2017-06-12 09:37:45 I think I'll just not send any messages at all, much easier to set up 2017-06-12 09:39:02 haha 2017-06-12 09:39:09 Indeed, if you can get away with it :) 2017-06-12 09:42:28 Now I just need to figure out a decent webmail to use... 2017-06-12 09:42:38 Morning. Thanks to ncopa and khanku for getting nbtscan built and into the repos so quickly. The slight problem is that it segfaults when I run it under docker. Anyone got any tips on where/how to start debugging it, please? 2017-06-12 09:42:39 lol, good luck 2017-06-12 09:42:42 Roundcube is... well... it's not great. ;) 2017-06-12 09:43:11 Eh, correction, s/Roundcube/PHP/ 2017-06-12 09:43:12 :) 2017-06-12 10:08:25 Kruge: i think we need to build it with debugging symbols, generate a core dump and finally use the coredump to generate a backtrace 2017-06-12 10:11:45 Kruge: i added a nbtscan-dbg package with debugging symbols 2017-06-12 10:12:15 Kruge: would be nice if you can report a bug on bugs.alpinelinux.org and tell how to reproduce it 2017-06-12 10:20:50 Will do 2017-06-12 10:20:55 Cheers 2017-06-12 10:35:21 Hmm. Where would the core file be written? 2017-06-12 10:38:42 nm, got it. Needed ulimit -c unlimited beforehand 2017-06-12 10:49:27 yeah 2017-06-12 11:07:13 Hello! I would like to install Alpine to a disk with ZFS. I am running tests on a VM. I have booted the "extended" release ISO, `modprobe zfs`, and `apk add zfs`. I have setup the disk and zpool. It is mounted. 2017-06-12 11:07:57 looking at http://wiki.alpinelinux.org/wiki/Alpine_setup_scripts and http://wiki.alpinelinux.org/wiki/Setting_up_disks_manually, it sounds like setup-disk should be able to recognize that my disk has been setup. 2017-06-12 11:08:40 eg, I thought if I provided `-m sys`, and had my zfs pool mounted at `/mnt`, I'd setup alpine on that disk. 2017-06-12 11:09:04 but setup-disk only seems to want to format/etc my disk 2017-06-12 11:10:37 oic, you have to pass the mount path.. so `setup-disk -m sys /mnt` will tell it to install alpine to /mnt (and leave the disk alone). 2017-06-12 11:10:58 however, setup-disk complains: "zfs is not supported..." 2017-06-12 11:11:40 that seems... crazy.. if you put "zfs alpine" into a search engine, you get lots of results for "3.5 ships with root zfs support"... 2017-06-12 11:45:16 ncopa: Do I need to do something special to use the dbg version? I installed nbtscan-1.5.0-r1 and the corresponding dbg package, ran and crashed the program, got the core file and looked at it in gdb. bt shows a thread, but says there is no symbol table info available 2017-06-12 11:46:03 I ran: (gdb) thread apply all bt full 2017-06-12 11:47:02 Bearing in mind that I've got this far largely by mashing my face against the keyboard until it worked, and I have no idea what I'm doing when it comes to debugging 2017-06-12 11:49:17 Did some more mashing. Cancel my last. 2017-06-12 12:12:22 alpe: the initramfs will find and be able to boot on zfs root. the install scripts are not yet modified to handle it 2017-06-12 12:12:45 also the /boot partition needs to be ext4 or xfs or something that bootloader understands 2017-06-12 12:12:51 or fat if its uefi 2017-06-12 12:13:27 i think `setup-disk /mnt` should get you close to a working system 2017-06-12 12:13:58 actually 2017-06-12 12:14:10 i think it should work 2017-06-12 12:14:22 you just need a separate /boot partition mounted 2017-06-12 12:14:43 so, for example, sda1, 1G, for /boot, ext4 2017-06-12 12:14:49 and sda2 for zfs 2017-06-12 12:14:55 for root 2017-06-12 12:15:18 then create the zpool and mount it on /mnt 2017-06-12 12:15:24 and mount /dev/sda on /mnt/boot 2017-06-12 12:15:31 and from there run setup-disk /mnt 2017-06-12 12:17:33 my immediate thoughts on that were that a sort of a profile based installer would probably serve the userbase better, but early init related things would also require some work, something like what dracut is or so 2017-06-12 13:26:50 ncopa: I had to modify `setup-disk` to allow for ZFS 2017-06-12 13:27:24 I hadn't figured out the stuff for /boot, so your info helps, thanks for that 2017-06-12 14:00:29 the more I do Alpine installs the more I like it 2017-06-12 14:04:06 and an additional benefit of doing a lot of things with Alpine has been that my shell scripting has really taken a leap forward 2017-06-12 14:04:17 yay 2017-06-12 14:04:21 I feel similarly 2017-06-12 14:04:54 it's really refreshing to use alpine after a few years with systemd/etc 2017-06-12 14:07:07 yeh, I don't share whatever the vision systemd makers have 2017-06-12 14:08:18 there is a vision? 2017-06-12 14:08:56 alpe: what changes did you have to do in setup-disk? 2017-06-12 14:09:50 ncopa: add `zfs` to the list of supported FS. there is a check, and `zfs` needs to be in the list 2017-06-12 14:10:34 ncopa: https://git.alpinelinux.org/cgit/alpine-conf/tree/setup-disk.in#n147 2017-06-12 14:10:50 id advice to run a main drive for /boot and cache/swap and put zfs on the full sata data drive 2017-06-12 14:12:12 jushur: this is for VM, so I would like to keep it to one disk.. but I think your recomendation makes a lot of sense for a physical system 2017-06-12 14:13:02 alpe: that was the only needed change? 2017-06-12 14:13:56 so far, yea, though haven't seen this boot. My edit to setup-disk got me through the core of the installer with an answer file. 2017-06-12 14:14:02 like my systems all have a small ssd for /boot /boot-backups /isos and so on. and a array of drives for the "system install" wich i also keep a copy of the ssd content on. and hash checks the ssd content against on bootup 2017-06-12 14:14:14 ncopa: I'm doing this over again clean, setting up /boot like you recommended 2017-06-12 14:15:52 ok, i dont think syslinux supports zfs so i think the error message is correct 2017-06-12 14:17:09 yea, I will retry without the edit, but with /boot on ext 2017-06-12 14:17:58 I'll ask here aswell (sorry for the duplicate for devel people): 2017-06-12 14:18:22 wrt to a previous conversation regarding firejail and its unreliability, I don't think other such tools exist in the available packages. I could spend a bit of time making an APKBUILD for an alternative, but which one? opinions welcome 2017-06-12 14:26:19 ncopa: is a zfs-virthardened package needed to run zfs on the virt release? 2017-06-12 14:53:09 TBB: when was that conversation? so I can look it up in the logs 2017-06-12 14:56:16 it was on -devel sometime last week 2017-06-12 15:11:10 ncopa: Do you want the full core dump, or the backtrace? 2017-06-12 15:12:11 (for nbtscan) 2017-06-12 16:42:54 alpe: yes its needed 2017-06-12 16:43:03 Kruge: barcktrace 2017-06-12 17:06:40 why is the default syslog.conf throwing errors in 3.6? 2017-06-12 17:06:57 it fails on something about mail or news logs, which don't exist 2017-06-12 17:08:35 ACTION installs syslog-ng anyways 2017-06-12 18:50:04 i am considering to apply to give a talk at dockercon eu in copenhagen about alpine 2017-06-12 18:50:42 if you are going to dockercon, what would you like to hear about? 2017-06-12 18:50:52 or what woudl be a good alpine talk at dockercon? 2017-06-12 18:51:58 ncopa: show how alpine gets so small (so they can relate), a bit about musl, but also what makes up the OS alpinelinux 2017-06-12 18:52:01 so boot to ram, etc 2017-06-12 18:52:15 even aconf for a slide or two 2017-06-12 18:52:39 a kind of a a "it's amazing for you because it IS amazing" 2017-06-12 18:53:43 aconf? 2017-06-12 18:53:44 (but i'm only asnwering as to the "what would be a good alpine talk at dockercon" - i'm not going, so please correct me guys) 2017-06-12 18:53:54 the lua web interface thing 2017-06-12 18:53:59 aha 2017-06-12 18:54:12 i don't really use it but it's a good bridge to beginners 2017-06-12 18:54:21 probably does not make sense in docker context 2017-06-12 18:54:25 no 2017-06-12 18:54:30 only as a "what else is there" thing 2017-06-12 18:54:37 ok 2017-06-12 18:54:54 many are not at all aware how many other parts are in alpine 2017-06-12 18:55:33 i was thinking: why alpine matters, when to use and when to not use alpine 2017-06-12 18:55:46 compare with debian/ubuntu/centos 2017-06-12 18:55:52 eg size stats 2017-06-12 18:56:10 and maybe some realtime demo (to show the speed) 2017-06-12 18:56:23 package install demo for sure 2017-06-12 18:56:35 i always do that if someone asks me about alpine 2017-06-12 18:56:42 downgrade maybe too? 2017-06-12 18:56:51 i 2017-06-12 18:57:14 i've been very happy when i found I could back out of an upgrade once with no more than apk-static 2017-06-12 18:57:26 (3.3 back to 3.2 or so) 2017-06-12 18:57:42 :) 2017-06-12 18:58:05 yeah, thats actually pretty cool, with busybox.static + apk-static you can do pretty much everything 2017-06-12 18:58:15 but from what you wrote... i think your version can take a few more colours 2017-06-12 18:58:18 yes! 2017-06-12 18:58:21 mention that 2017-06-12 18:59:00 you could go as far as saying it turned out an extremely good fit because it's always tried to work for appliance-ish use cases 2017-06-12 18:59:10 it's not random that these things work well in alpine 2017-06-12 18:59:16 yes 2017-06-12 18:59:20 thats a part of the history 2017-06-12 18:59:26 designed to run from ram 2017-06-12 18:59:37 which is kind of a disposable system 2017-06-12 18:59:48 which is very similar to what docker does 2017-06-12 19:00:06 yup. but if you tell that to people you cant' assume they have a background in industrial/embedded stuff 2017-06-12 19:00:14 so really mention "this is why" 2017-06-12 19:01:33 *nod* 2017-06-12 19:01:43 and if you go into image sizes, also go into deployment time differences 2017-06-12 19:02:51 bye & good luck with the slides 2017-06-12 19:03:01 thanks for hte ideas! 2017-06-12 19:10:45 if we had a/b upgrades in apk it would be even more fappish ;f 2017-06-12 19:46:57 oO 2017-06-12 19:49:30 A/B upgrades and some sort of "glibc-run" helper for those few programs that can't be rebuilt on musl would make Alpine actually pretty sweet as a desktop 2017-06-12 19:49:52 sort of like how "optirun" forces a different libgl 2017-06-12 20:08:13 ncopa: what would go into making that happen? maybe I can help. 2017-06-12 20:46:35 nbtscan bug and backtrace added to the end of the package request issue: https://bugs.alpinelinux.org/issues/7408 2017-06-12 20:53:30 scadu: a/b upgrades is when you atomically upgrade a symlink to /usr ? 2017-06-12 20:53:47 darkfader: it's like what ChromeOS does iirc 2017-06-12 20:53:51 alternating root partitions 2017-06-12 20:54:29 one can do that for live (no reboot, no downtime) application upgrades 2017-06-12 20:54:35 but, tbh 2017-06-12 20:54:39 that's for good applications 2017-06-12 20:55:13 what would improve if the OS has features like that what one runs day-to-day 2017-06-12 20:55:19 do you think it'd help/ 2017-06-12 20:55:20 ? 2017-06-12 21:34:41 '' 2017-06-12 22:15:07 Hi folks, I've gone over /etc/abuild.conf, but I can't find a variable to declare that would allow me to specify the workdir where artifacts are built. I would like to keep aports in my homedir on disk, and then have all my packages be built on a ramdrive, is this possible? 2017-06-12 22:16:14 I could maybe have aports mounted in a readonly filesystem, and then put a ramdrive overlay fs on top, but maybe there is a simple variable i can set? 2017-06-12 22:16:50 mount --bind 2017-06-12 22:16:54 or overlayfs 2017-06-12 22:17:55 Mhmm yeah overlayfs would work 2017-06-12 22:18:51 still though, I'd have thought the WRKDIR used by the build system to be a configurable variable 2017-06-12 22:19:16 like most other port based build systems 2017-06-12 22:20:45 Thanks though tidux, I'd forgotten about mount --bind :) 2017-06-12 22:22:03 Actually, maybe I can just hack in a new variable myself 2017-06-12 22:22:25 Im sure there's others that would appreciate functionality like that 2017-06-12 22:24:13 Hello folks 2017-06-12 22:24:30 ello. 2017-06-12 22:26:17 Anyone got a few cycles to help me puzzle out why localmount is not finding my logical volumes? I "fixed" the issue by creating another script in init.d to do a mount -a, but I'm guessing I missed a step during setup. =) 2017-06-12 22:27:33 I would but I only have 35 cycles le 2017-06-12 22:27:46 =D 2017-06-12 22:29:07 :p 2017-06-12 22:29:59 I'm guessing it's some kind of timing issue - another machine of mine that has LVM over RAID works properly although its array is only four drives versus the troublesome one's eight. 2017-06-12 22:30:40 I have zero experience with lvm, sorry 2017-06-12 22:31:03 *nod* It's cool. 2017-06-12 22:33:24 If its a timing issue though, try putting a sleep in the init file maybe before it does the mount, to verify? 2017-06-12 22:34:32 I'll give that a go. 2017-06-12 22:36:33 Does it report "Some local filesystem failed to mount" on the console during boot? 2017-06-12 22:37:26 Yeah, that's the message 2017-06-12 22:37:28 Aha! 2017-06-12 22:37:40 xfs isn't in the list of filesystems to mount. 2017-06-12 22:38:17 Hehe awk 2017-06-12 22:39:38 Hmm, weird list of filesystems to automount: afs, ceph, cifs, coda, davfs, fuse, fuse.sshfs, etc. 2017-06-12 22:40:12 But none of the usual suspects like ext4, btrfs 2017-06-12 22:42:17 There's two lists though 2017-06-12 22:43:24 Those sound like they would be in the extra_net_fs_list as opposed to net_fs_list 2017-06-12 22:44:04 I dont know where either of those are declared though, not in the localmount init file 2017-06-12 22:44:23 Yeah. 2017-06-12 22:47:43 A 10-second sleep didn't cure it, trying 30. 2017-06-12 22:48:00 Then I'll try adding xfs directly to the list of things to mount and see if that helps. 2017-06-12 22:48:04 hehe, its def not a timing issue then 2017-06-12 22:48:18 yes i would just prepend xfs to types 2017-06-12 22:51:42 Okay, that's weird. That list must be some sort of exclusion - when I added it, it didn't try to mount it and so didn't warn me about the devices not existing. 2017-06-12 22:52:19 maybe appending it as a kernel flag 2017-06-12 22:57:39 did you append it to types or to the *net_fs_list? 2017-06-12 22:57:49 types 2017-06-12 22:57:57 the net_fs_lists are indeed an exclusion list if uname == Linux 2017-06-12 22:58:07 by the looks of it 2017-06-12 22:58:13 Gotcha 2017-06-12 22:58:52 I haved booted the extended ISO in a VM, then used fdisk to create 2 partitions on the virtual disk.. which fdisk shows as `/dev/sda1` and `/dev/sda2`, however, `mkfs.ext3` and such only see `/dev/sda`. Is this some GPT weirdness I'm confused about? 2017-06-12 22:59:30 alpe: is there a reason you're not using the install scripts? 2017-06-12 22:59:30 Ohh what whas it? 2017-06-12 22:59:55 Xe: yea, zfs 2017-06-12 23:00:09 alpe: that'd be a good reason lol 2017-06-12 23:00:24 what does `dmesg` say? 2017-06-12 23:00:30 I got the zfs part working, but now focusing on providing /boot 2017-06-12 23:00:33 elegast: I am le dumb, basically. I was sure that I'd added LVM to the boot runlevel. Nope. T'was in default. So LVM activated after mounts. *facepalm* 2017-06-12 23:00:49 usually when there's some issue finding device nodes you'll get a complaint in the kernel ring buffer 2017-06-12 23:01:19 haha, I considered that for a moment, but i was like nahh, surely lvm would have been enabled 2017-06-12 23:01:38 Problem solved though! 2017-06-12 23:01:49 Oh, it was ... just far too late ... :P 2017-06-12 23:02:04 gods, I KNOW I checked that ... dangit. 2017-06-12 23:02:26 Thank you for letting me talk through it though =D 2017-06-12 23:02:40 Xe: I don't see anything about that in dmesg, and the sda looks ok there 2017-06-12 23:02:41 Hehe no worries, and you're welcome 2017-06-12 23:03:02 alpe: weird, what does `ls /dev/sd*` say? 2017-06-12 23:03:26 Xe: /dev/sda 2017-06-12 23:03:30 reboot? 2017-06-12 23:03:45 oh, maybe.. I did update fdisk 2017-06-12 23:04:36 that should have updated the device nodes though, yolo reboot and check? 2017-06-12 23:05:00 nope, still only `/dev/sda` after reboot 2017-06-12 23:05:25 `fdisk -l /dev/sda` shows me both partitions as expected 2017-06-12 23:05:32 fdisk -l does still report both partitions? 2017-06-12 23:05:39 yep 2017-06-12 23:06:00 do you get the same behavior using parted? 2017-06-12 23:06:21 great question... 2017-06-12 23:07:18 idk if parted lets you do ZFS partitions, but if it doesn't maybe make it a simple EXT4 volume and then recreate as zfs? 2017-06-12 23:08:30 parted doesn't see the same as fdisk. OK, I'll see if I can use parted (or sgdisk) to create the partitions and see what happens... 2017-06-12 23:30:45 ZFS really really wants multiple whole disks 2017-06-12 23:31:07 the best way to do Alpine with ZFS is diskless root 2017-06-12 23:31:43 the extended iso is a good starting point 2017-06-13 00:06:34 I don't understand alpine's default password setup. Everything is expired by default, and the expired handler doesn't do anything? 2017-06-13 03:29:18 https://ptpb.pw/TQVb what a lovely package manager 2017-06-13 03:37:22 indeed 2017-06-13 03:52:28 why would you --initdb on your current root? 2017-06-13 04:05:26 https://git.alpinelinux.org/cgit/aports/log/community/php7?h=3.6-stable 2017-06-13 04:05:33 How do I rollback to php 7.0.18 2017-06-13 04:35:46 <_ikke_> alexw: You'd need to get php7 from the 3.5/community repo 2017-06-13 04:35:58 Yeah but then php7-mongo is missing 2017-06-13 04:36:02 <_ikke_> though, that's 3.0.16 2017-06-13 04:36:09 :( seems silly that they updated it with no notice 2017-06-13 04:36:19 Stuffed up all my docker bases as now it's seeding from 7. 2017-06-13 04:36:48 <_ikke_> Well, edge is always updated to the latest version 2017-06-13 04:37:26 <_ikke_> 3.6 is 7.1.5 2017-06-13 04:41:29 the issue is that edge is now tracking php7-mongo 2017-06-13 04:41:33 which is for 7.1.5 2017-06-13 04:41:39 so I'd have to build mongo myself 2017-06-13 04:42:06 Ahh I found this https://php-alpine.codecasts.rocks/ 2017-06-13 04:42:45 <_ikke_> Yeah, you cannot rely on edge to keep a stable version 2017-06-13 04:43:54 <_ikke_> I don't see a php7-mongo at all in the official repos 2017-06-13 04:44:09 That code casts mirror is working well 2017-06-13 04:44:13 I think it's a fork anyway 2017-06-13 04:44:35 Makes more sense as it's how it works in the ubuntu community also - a seperate repo to track all the different versions 2017-06-13 05:18:50 xentec: well, because --initdb isn't --please-delete-all-installed-packages-and-their-contents 2017-06-13 05:20:33 buhman: it pretty much is. an "initialized database" has no data. 2017-06-13 05:23:35 however, it's probably for the better to put it behind a --force flag in the case apk detects an existing world file 2017-06-13 05:24:26 kaniini? ^ 2017-06-13 05:37:03 xentec, --initdb is currently "create if not already existing" 2017-06-13 06:23:37 buhman: ouch 2017-06-13 06:24:30 buhman: that's definitely a bug. i will look at it tomorrow if fabled doesn't get to it first 2017-06-13 06:30:52 buhman: i cannot reproduce it 2017-06-13 06:32:27 the only way i can reproduce it is to delent/rename /etc/apk/world first 2017-06-13 06:32:31 then it will happen 2017-06-13 06:33:31 (if someone wants to try i recommend to first install busybox-static and apk-tools-static and make a copy ov /bin/busybox.static and /sbin/apk.static) 2017-06-13 06:35:08 (and make a backup of /etc/apk/keys) 2017-06-13 06:35:27 buhman: are you sure that your /etc/apk/world was intact? 2017-06-13 06:36:51 what is nice though is that i was able to recover the system with only 3 files: busybox.static apk.static and /etc/apk/world 2017-06-13 07:34:14 ncopa: I don't think it was 2017-06-13 07:34:23 ncopa: that's why I was trying to --initdb anyway 2017-06-13 07:37:15 the /etc/apk/world tells apk what is supposed to be installed 2017-06-13 07:37:30 you can manually edit it and next apk upgrade will adjust 2017-06-13 07:37:41 well, unless apk doesn't exist 2017-06-13 07:38:32 if you are trying to rescue a system, then busybox-static and apk-tools-static is your best friends 2017-06-13 07:39:55 so, the idea is I'd re-add that to the root filesystem from some working system? 2017-06-13 07:40:23 yes 2017-06-13 07:40:53 fi you copy the /etc/apk/world from a working system, you will end up with the exact same packages 2017-06-13 07:41:12 if you want to clone the system (inclusive config) 2017-06-13 07:41:24 then you can generate an "apkovl" 2017-06-13 07:41:28 lbu package 2017-06-13 07:41:46 that is, on a working system 2017-06-13 07:42:38 that will run apk audit, which will compare the files from apk db with the files on the running system (the checksum), generate a file list of files that are modified compared to the file in the apk 2017-06-13 07:42:50 and then generate a tarball form this file list 2017-06-13 07:43:11 this tarball you can copy to another system 2017-06-13 07:43:28 it will also include the /etc/apk/world 2017-06-13 07:44:10 that sounds pretty cool actually; is there documentation for that/ 2017-06-13 07:44:29 I don't actually need to rescue anything; it was a toy root filesystem luckily 2017-06-13 07:44:37 https://wiki.alpinelinux.org/wiki/Alpine_local_backup 2017-06-13 07:44:47 that is how we do our run-from-ram setups 2017-06-13 07:44:59 alpine will install the packages at boot time to tmpfs 2017-06-13 07:45:03 root is tmpfs 2017-06-13 07:45:30 then you can "lbu commit" which will create an apkovl and save it to usb stick 2017-06-13 07:45:40 at boot, alpine will look for apkovl files 2017-06-13 07:45:50 and if found, it will use it when setting up the tmpfs root 2017-06-13 07:45:56 and restore the system the way it was 2017-06-13 07:46:40 it will not backup data (eg stuff in /var) 2017-06-13 13:07:01 https://janweitz.de/article/creating-a-zfs-zroot-raid-10-on-ubuntu-16.04/ shows parted with ZFS support, but when I use `mkpart zfs ...` with the parted in alpine 3.6.1, I get "invalid token: zfs... File system type? [ext2]" 2017-06-13 13:48:12 vim -- "https://janweitz.de/article/creating-a-zfs-zroot-raid-10-on-ubuntu-16.04/" 2017-06-13 13:48:58 sorry 2017-06-13 15:09:23 ZFS really needs to use its own tools 2017-06-13 15:09:28 and ideally its own whole disks 2017-06-13 15:13:55 tidux: sure, but you first need to setup GPT and the partitions (you need to keep /boot separate) 2017-06-13 15:16:38 no you don't 2017-06-13 15:16:44 you need entire separate physical media for ZFS 2017-06-13 15:16:54 unless you're doing something utterly stupid like ZFS on a laptop 2017-06-13 15:17:21 I'm putting ZFS on a VM, so it's boarderline stupid 2017-06-13 15:17:32 still give it a separate virtual disk 2017-06-13 15:17:42 one for /, /boot, swap, one for ZFS 2017-06-13 15:18:11 my goal is to put / on ZFS 2017-06-13 15:18:17 (minus /boot) 2017-06-13 15:19:53 then make /boot a separate, tiny disk 2017-06-13 15:20:38 OK, yes, that could make sense 2017-06-13 15:21:12 still, it seems like alpine's parted is older, or isn't built with zfs support 2017-06-13 15:24:57 Hello 2017-06-13 15:25:49 I try to install Alpine Linux in a Virtual Box and have a question regarding the drivers that X requests for its screen. 2017-06-13 15:26:07 I understand that they are in a virtualbox-guest package, but there are several and I am not sure which one to use. 2017-06-13 15:26:39 Is it: virtualbox-guest-modules-hardened, virtualbox-guest-modules-virthardened or virtualbox-guest-additions? And what are the others for? 2017-06-13 15:27:14 -hardened and -virthardened refer to different kernels you can install and run 2017-06-13 15:27:35 so -modules-virthardened probably have the kernel modules built for the virthardened kernel 2017-06-13 15:27:51 you probably want -virthardened, but that depends on the ISO you booted 2017-06-13 15:28:13 if you use vagrant, there is an existing vagrant box you could use with alpine 3.6.1 2017-06-13 15:28:31 I downloaded the "ISO prepared for Virtual machines". 2017-06-13 15:28:47 that's -virthardened 2017-06-13 15:29:52 Ah. Great. Thank you! 2017-06-13 16:04:34 Hello, we're trying to install mysql-client built on mysql and not on MariaDB, because right now one of the options of mysqldump built on MariaDB is unavailable yet required for us 2017-06-13 16:04:57 Sadly we haven't found a way to do this, do you have any idea? 2017-06-13 17:30:30 ThomasRuiz: unfortunally we dont have mysql package, only mariadb 2017-06-13 17:32:00 looks like the last alpine release with mysql was alpine 3.1 2017-06-13 17:39:35 MariaDB's binary is still called mysql 2017-06-13 17:39:46 so there should still be a mysqldump 2017-06-13 17:41:23 <_ikke_> tidux: Problem is aparently that it is not feature complete 2017-06-13 17:41:34 ah 2017-06-13 17:48:16 erm 2017-06-13 17:48:18 https://mariadb.com/kb/en/mariadb/mysqldump/ 2017-06-13 17:48:46 https://pkgs.alpinelinux.org/contents?file=mysqldump&path=&name=&branch=&repo=&arch= 2017-06-13 22:15:07 tidux, out of curiosity what features are missing? 2017-06-13 22:20:39 I have no idea, ask _ikke_ 2017-06-14 03:51:27 <_ikke_> tidux: I was reacting on "one of the options of mysqldump built on MariaDB is unavailable" 2017-06-14 13:16:13 About the mariadb question yesterday, the missing feature from mysqldump is the --set-gtid-purged=OFF, which is required when dumping a google cloud sql database 2017-06-14 13:36:20 Hi folks, I'm running alpine as a low footprint desktop on a small netbook, which works surprisingly well for a distribution that's still with one foot in the router/embedded side of things. But, I'm a sucker for eye candy, and spotted Plymouth in the aports tree, has anyone gotten that to work? 2017-06-14 13:37:10 Most docs on plymouth are based on systemd integration, and to be honest, I've never actually took a look good at *how* it actually does its thing. So I'm not sure what things to look for to get this working 2017-06-14 13:38:27 im not sure we have plymouth support but i think we had some hack using busybox tools 2017-06-14 13:39:27 Well, if there's plymouth support I'd be surprised :) 2017-06-14 13:39:36 How is busy box involved though? 2017-06-14 13:39:44 brb 2017-06-14 13:39:52 sure take your time! 2017-06-14 13:43:46 So far, I've got extlinux booting with vesamenu in native resolution, kernel mode setting is done as early as possible, and no mode switching between bootloader and kernel init anymore, I've built and installed the plymouth package, and I *think* I've included the binaries correctly in the ramdrive image 2017-06-14 13:47:00 My guess is, the only thing left to do for me is to get openrc to poke at plymouthd at the right time, but I'm not sure if plymouth is simply unconfigured, doesn't work all together, or simply isn't triggered to start yet. It does some flickering and what looks like modeswitching just before the first openrc message, which I think it didn't do before I updated the ramdrive image 2017-06-14 13:49:36 Gentoo has an openrc-plugin, that might be something worth checking out 2017-06-14 13:49:40 There's also no mode switching happening anymore during the transition from text-mode boot to starting X 2017-06-14 13:50:26 so I think, this little system should be able to do plymouth just fine, and hmm, I was going through arch resources, but I hadn't considered gentoo yet 2017-06-14 14:04:56 Lets say I wanted to inspect the initramd image from a running system to make sure it has everything included in the right places, before it gets replaced by the root filesystem mounts, would the "keep ramdrive" related kernel flags and enabling single user mode do the trick, or do those flags not what I think they would? 2017-06-14 14:07:42 Im aware I could just extract the ramdrive image, but on that particular system thats a slow and unwieldly process, atleast to do more then once. Also the boot process itself might influence its state, for which i want to see its effects 2017-06-14 14:12:23 elegast: there are some fbsplash in initramfs init 2017-06-14 14:12:32 i havent really used it 2017-06-14 14:13:20 yes that was actually my first look, until I figured arch can do both openrc and plymouth, when I switched focus to plymouth 2017-06-14 14:14:22 but, I do think its easier to get fbsplash working on alpine, then it is to get plymouth working from referenced docs from either arch or gentoo 2017-06-14 14:14:40 hi, all distros are getting bigger and bigger include "small" AL 2017-06-14 14:14:40 a general query, is it a good idea to buy BluRay writer/media, seems the device is around 10yrs but the prices never came down, guessing less mass production. 2017-06-14 14:15:05 fabled: did you have patches to interact with fbsplash from openrc? 2017-06-14 14:15:27 ncopa, i had some rudimentary patches, but i never finished them properly 2017-06-14 14:15:45 i do use fbsplash, but no with the progress bar mode 2017-06-14 14:15:55 ok 2017-06-14 14:16:10 i suppose it should work without progressbar if you dont load fbcon 2017-06-14 14:16:27 fabled, would you mind sharing those? I don't mind getting my hands dirty if its not production quality 2017-06-14 14:17:48 vkrishn: i bought usb dvd player for old cdrom backups 2017-06-14 14:18:16 you mean writer 2017-06-14 14:18:28 also bluray, not dvd 2017-06-14 14:18:46 Isn't sony the only one on the bluray bandwagon though? 2017-06-14 14:19:07 I have 3 usb dvd writer, and now AL has 6 ARCHs :( 2017-06-14 14:19:57 Since media is getting bigger and bigger (e.g. 4k) I think the prudent move would be to wait and which medium prevails when the industry starts moving again. 2017-06-14 14:19:58 elegast, ncopa : now that i think of it, i believe i had it so that the bootsplash is showed during initramfs only. probably showing progress from apk installing packages to tmpfs 2017-06-14 14:20:50 makes sense 2017-06-14 14:22:00 ah, i'm looking more of a solution that hides all the scary messages behind a user friendly graphic that appeals more to people that well, aren't me. 2017-06-14 14:24:17 I'll keep thinkering with plymouth for a little while longer, I havent *really* explored all the nooks and crannys yet, it was just that there was a package in aports so i figured someone had gotten it working, but alas 2017-06-14 14:24:57 If it turns to be a dead end, I'm more confident about getting splashfs to work, though its a bit more archaic. 2017-06-14 14:25:20 Thanks for the help ncopa and fabled 2017-06-14 14:26:24 err, thats fbsplash obviously* 2017-06-14 14:38:51 On a completely different matter, any users here that have used alpine (or linux in general) for transparent network interception? I used to use netfilter + tproxy (and sometimes kzorp) to enforce certificate pinning, the occasional packet mangling and the like at the network layer, but nft has zero support for either of those things, and arent scheduled to be worked on in the near future either. Is the prudent choice to stick 2017-06-14 14:38:53 with iptables or move to nft with nfqueue, for which it does have support (though its more resource taxing)? 2017-06-14 14:39:39 with regards to plymouth and Alpine, it can work but if you run with grsec then you might have to do some extra magic to keep it working 2017-06-14 14:39:56 and you guys are right about plymouth documentation, it f**king sucks 2017-06-14 14:40:26 TBB, yes grsec is another problem, for now I've disabled it just to get things off the ground 2017-06-14 14:41:13 2017: the year we disabled kernel hardening in favor of a fancy boot screen 2017-06-14 14:41:15 with regards to docs, yes its not meant for end users to begin with, its assumed that the distributions get everything working and packaged 2017-06-14 14:41:28 as soon as you switch_root grsec starts reporting various chroot related blocks, so if you want to do that you'll have to disable some chroot limitations 2017-06-14 14:42:04 probably the unix socket 2017-06-14 14:42:06 elegast: true, but still... but then I guess there are only three real distros in the world (debian, rh and ubuntu) 2017-06-14 14:42:17 (and yes, I was being sarcastic) 2017-06-14 14:43:57 but yeh, I'll see if I can show the plymouth package itself some love at some point; I should do that as currently the package in testing (well, the last time I tried it anyway) still contains some systemd related files and dirs 2017-06-14 14:44:40 TBB, yeah, i installed the package and ended up with systemd related directories straight in the root after some script fired 2017-06-14 14:45:03 yup, those are simple to take care of in the apkbuild 2017-06-14 14:46:02 mhm. I like the apkbuild system, it's not pkgsrc, but it's a fine piece of work all the same! 2017-06-14 14:46:37 I like it too, although, coming from the world of RPM some things took some time to get used to 2017-06-14 14:47:11 TBB: like what? 2017-06-14 14:48:26 hehe, personally I never much liked RPM, I always ended up doing source compiles anyway, but when I got introduced to the BSD way of things, ports where just a magical thing 2017-06-14 14:48:29 I didn't exactly document them, but subpackaging is a bit different, the way files are chosen for packages is different (and IMO better done) in RPM packaging, etc 2017-06-14 14:49:26 but I appreciate abuild's flexibility; rpmbuild is quite a stiff way of packaging stuff 2017-06-14 14:49:26 yeh, that convention is not really tied to the ports system itself though, as far as I can tell 2017-06-14 14:49:39 its more of a distro choice i think 2017-06-14 14:51:24 personally though, I think the doc+licensing split from base packages, along with the dev packages is a good move, atleast for alpine 2017-06-14 14:52:12 the same with modularizing the larger packages 2017-06-14 14:52:19 you only install the binaries you really need 2017-06-14 14:52:31 oh nice, I just ran my own installer without thinking too much and I just wiped out the partition table of the system the installer's on 2017-06-14 14:52:51 oh, thats shitty 2017-06-14 14:52:57 it could be worse though 2017-06-14 14:53:23 the other day, I installed an older suse enterprise release on a system that a BSD system on a compact flash card 2017-06-14 14:53:25 fortunately I literally install Alpine tens of times every day so as soon as I get a squashfs image done of sysroot I'll just repartition and unpack 2017-06-14 14:53:39 I took great care to install it along side it without having it touch anything else 2017-06-14 14:53:54 then after it installed and booted and i ran the package updates 2017-06-14 14:54:12 it wiped the gpt+mbr of the bsd and replaced it with grub on its own. 2017-06-14 14:54:35 :D 2017-06-14 14:55:48 Some distros just assume the only thing in the world is Linux, or even just their distro. ugh. 2017-06-14 14:56:37 I've now been doing this tool for installing and building packages for various distros and this was the third time in 2.5 years I've managed to wipe out my workstation by choosing the wrong target disk 2017-06-14 14:56:48 Hehe, fix your installer though! 2017-06-14 14:57:44 it's been enough that I've implemented a check for that and it refuses to install on the boot media; however, for now you have to do that check in the install profile itself, it's not done automatically within the install command itself... 2017-06-14 14:58:39 Wiping your workstation on accident once a year is not the best track record though! Stop that :) 2017-06-14 15:00:17 of course everything important is backed up elsewhere... right? 2017-06-14 15:00:20 ... right? 2017-06-14 15:00:24 lol if only 2017-06-14 15:00:54 and installing a new workstation fortunately takes less than 30 seconds so it's not that much of a catastrophy 2017-06-14 15:02:30 True, as long as the important bits are safely tucked away, its no biggy. Reconfiguring is a bitch though, I rarely just restore the configs as is after such events 2017-06-14 15:03:07 I wrote a little mirroring tool the other day that, now that I think of it, takes most of the pain away from that 2017-06-14 15:04:02 Yeah, if you just mirror it offsite or atleast on a different machine, it's just a simple restore 2017-06-14 15:04:09 it's based on big base squashfs snapshots and differential snapshots, so if you want to restore such a system you just mount them all as an overlay, basically 2017-06-14 15:04:42 Though I only do that for servers and virtuals, not workstations. Their configuration fluctuates too much for my taste 2017-06-14 15:05:31 Its mostly just making sure code and keys are safe, which should be in some repository (minus the keys ofcourse) 2017-06-14 15:05:32 yup, developer workstations live a bit too much 2017-06-14 15:06:35 I love squashfs and overlayfs, those things have made my life so much easier in the last year or so 2017-06-14 15:07:30 hehe, yeah I havent' gotten all the ins and outs of overlayfs/aufs under my belt yet, I've only recently gotten back into the linux way of things 2017-06-14 15:08:16 oh yeah, you mentioned BSD above 2017-06-14 15:08:26 I used to be exclusively bsd and illumos for the decade or so 2017-06-14 15:08:31 last* 2017-06-14 15:09:45 I finally replaced my last BSD earlier this year; while I had 20 years of experience of especially FreeBSD, I just haven't had the time and the opportunity to actively work on it so administrative tasks were always like "okay, how the f**k did you do this on BSD again" 2017-06-14 15:09:48 But support for consumer hardware is scarse on those systems 2017-06-14 15:10:01 so I replaced my last FreeBSD with Alpine 2017-06-14 15:10:43 Hehe, yeah, servers are still either FreeBSD or Illumos based over here (especially when it comes to virtualization support). 2017-06-14 15:11:21 That won't change, anytime soon, lxc and kvm are nice, but they don't cut it just yet in my opinion 2017-06-14 15:12:06 But I'm likely replacing most of the things that have gui's and consumer hardware with alpine 2017-06-14 15:12:15 Or require a low footprint, for which its really great 2017-06-14 15:14:08 But yeah, solaris containers are still the do all be all solution for me in regards to feature set, design and stability over especially lxc or even bsd jails. 2017-06-14 15:14:37 the low footprint thing is really good wrt Alpine as well; my top end laptop runs KDE Neon on Ubuntu and you don't have to run it for long to have spent 2.5 GB of RAM 2017-06-14 15:14:50 I do run alpine inside some of the containers though, that goes really well 2017-06-14 15:16:02 Yeah my main laptop and workstation are debian + kde, and the couple of older netbooks im working on run alpline 2017-06-14 15:16:26 I haven't had a chance to do things in Solaris in a looong time, I left the telecom industry where it was the platform of choice over a decade ago 2017-06-14 15:16:44 I like how it can run a full gui and only use around a 100 megs of ram 2017-06-14 15:17:20 Ah, I imagine it still going strong there. Though i've moved away from pure Solaris after Oracle shat on it 2017-06-14 15:18:30 Im running mostly smartos and tribblix (which is an obscure distro, but is to illumos what alpine is to linux, sort of) 2017-06-14 15:19:49 I even had a chance to play with HPUX for a couple of years, and some of the systems were really massive; the Superdomes were systems with I don't even know how many cores and how much memory, but massive enough to run a lot of containerized stuff 2017-06-14 15:20:04 ah, ram usage has grown a bit actually, I think thats mostly just because plymouth and a few other things went into the ramdrive just now. 2017-06-14 15:20:21 Hehe HPUX, I never had the pleasure of working with that 2017-06-14 15:20:30 or rather displeasure, I'd guess 2017-06-14 15:20:40 I think pleasure is not the word to describe it really :D 2017-06-14 15:20:59 mhm :P 2017-06-14 15:21:37 No, Solaris (and OSX but that doesnt really count) are the only commercial unixen I've worked with 2017-06-14 15:21:52 a few others which are UNIX like, but don't really classify as such 2017-06-14 15:22:33 I even had some AIX systems to look after; that's one commercial UNIX I never really had a chance to get properly familiar with 2017-06-14 15:22:40 but it was ... different 2017-06-14 15:23:27 I imagine AIX being more like SunOS was before version 5 2017-06-14 15:24:11 Or idk, was SunOS 4 based on BSD still? I don't remember 2017-06-14 15:24:36 ancient history 2017-06-14 15:24:45 well before my time actually 2017-06-14 15:26:28 I liked CDE though, even now it still has its charmes 2017-06-14 15:26:46 If only it wasn't completely security hole riden 2017-06-14 15:26:49 I'd still use it 2017-06-14 15:27:37 Anyway, Im neglecting my duties, thanks for the chat TBB see you around :) 2017-06-14 15:28:25 see ya :) 2017-06-14 15:38:45 hi guys, can I ask one slightly off topic question? (related to terminal, gui) 2017-06-14 15:40:30 what do you think about this color combination > #efefef as bg, #595959 as fg 2017-06-14 15:42:39 also 'Blank Your Monitor + Easy Reading' addon for firefox is subjectively (in my opinion) incredible. To turn all the pages into given colors, makes internet browsing a kind of worth journey then 2017-06-14 16:28:14 I feel like elinks is a lot better than that. 2017-06-14 16:30:49 elinks is abandoned upstream 2017-06-14 17:40:30 has someone tried adjusting colors in X (X11/Xorg)? 2017-06-14 17:40:38 white balance, any other method? 2017-06-14 18:07:16 I think bluray should either get replaced or prices downed (mass production) 2017-06-14 18:07:16 looks more like patents and DRM issue 2017-06-14 18:07:16 long data storage (cold) is becoming an issue 2017-06-14 18:07:25 elegast, iirc, bluray disc capicity can grow, in TBs 2017-06-14 18:23:21 vkrishn: oh I had no idea BR disc capacity could grow into the TBs, though that's only of value when it's backward compatible with current generation readers, I suppose 2017-06-14 18:25:12 but like I said, besides only, only a very small part of the industry (and the world at large) supports bluray, and then we're not even talking about *using* it, which keeps costs extremely high (with regards to other media types) 2017-06-14 18:25:23 *besides Sony 2017-06-14 18:26:06 that fact hinders adoption 2017-06-14 18:26:12 and innovation for that platform 2017-06-14 18:29:10 But yes, I imagine that patenting and ugly things like drm are also contributing factors, like you say 2017-06-14 18:31:30 Either way, If you have a need for it, then by all means, buy a bluray writer. As far as my limitet understanding goes though, I don't expect prices to drop for years. 2017-06-14 18:31:41 limited* 2017-06-14 18:35:19 rewritable optical media degrades in like 15 years anyways 2017-06-14 18:35:26 I'll stick to NAND and spinning rust 2017-06-14 19:18:36 doen't NAND require pluggin the device for after certain time to keep data alive ? 2017-06-14 19:19:40 maybe stick with traditional hdd for a while and take extra handling care 2017-06-14 19:19:43 <_ikke_> it's non-volatile 2017-06-14 19:20:30 ok, thanks that clears it, I thought it required some internal charge in capicitors to hold the data 2017-06-14 19:20:55 new hdd, 10TB have helium in it... bravo 2017-06-14 19:22:59 maa look my hdd spins and flyies too 2017-06-14 19:25:51 gee ! what you talking about ... did you see an ufo 2017-06-14 19:32:08 nand memory loses its bits after a few years 2017-06-14 19:32:29 typical ssd will probably be unreadable after 15 years of unpowered storage 2017-06-14 20:56:39 ACTION cig 2017-06-14 21:28:34 lol what in the actual f is "windows 2006" 2017-06-14 21:29:13 vista probably 2017-06-14 21:29:35 yer the timeframe makes sense 2017-06-14 21:30:22 I just read the comment going through initramfs-init about supporting the option to tell the bios not to support windows 2006 and was like "wat?" 2017-06-14 21:31:25 yeah it's an ACPI thing 2017-06-14 21:31:40 some broken ass hardware breaks stuff unless you report your OS as Window 2006 2017-06-14 21:31:46 which presumably vista does 2017-06-14 21:32:43 hmm, doesn't the ! mark infront of the declaration mean *not* to support it? 2017-06-14 21:33:39 or does the exclamation mark emphasize the declaration, as to say we *must* support it 2017-06-14 21:33:48 oh yeah, you're right 2017-06-14 21:35:07 https://github.com/torvalds/linux/blob/eb4125dfdb1f64a5e41da8315bff48f67e5d1712/Documentation/admin-guide/kernel-parameters.txt 2017-06-14 21:36:05 https://github.com/torvalds/linux/blob/786856b6d111a30d51c98526cc7a283ecc73f00f/drivers/acpi/acpica/utosi.c#L96 2017-06-14 21:36:07 yeah like this 2017-06-14 21:36:18 Maybe there's some hardware vendors out there that had "vista certified" hardware that would assume vista will ONLY EVER run on this system 2017-06-14 21:36:20 linux claims to support the win2006 OS interface, which buggy hardware will respond to badly 2017-06-14 21:36:28 as it then assumes some quirks/os bugs 2017-06-14 21:36:31 i believe 2017-06-14 21:37:17 mhm makes sense 2017-06-14 21:37:44 so it was the other way around of what I initially said :p 2017-06-14 21:38:06 lol, its all good, vista has always been kinda upside down anyway 2017-06-14 21:41:30 another matter, though, the reason i was going through mkinitfs scripts. If I want to add extra binaries to the initrd image, I add "foo" to mkinitfs.conf, and then list the files in features.d/foo.files? 2017-06-14 21:41:51 just to verify 2017-06-14 21:42:35 yeah 2017-06-14 21:43:00 Allright, much appreciated :) 2017-06-14 21:45:03 hello, anyone know where I can find the redistributed license file on alpine? i'm trying to figure out what license it is distributed under (Apache, MIT, etc) 2017-06-14 21:46:17 alpine is a project, not a single piece of software 2017-06-14 21:46:38 we develop multiple pieces of software under varying licenses, and we package software under varying licenses 2017-06-14 21:46:44 anything in specific you want to know the license of? 2017-06-14 21:46:48 you can just assume any linux distrubution consists of a linux kernel and gnu user land, former is gpl2 and the latter ranges from (l)gpl2-3 2017-06-14 21:47:02 well that would be your first mistake 2017-06-14 21:47:05 alpine does not use a gnu userland 2017-06-14 21:47:21 right 2017-06-14 21:47:24 busybox and musl 2017-06-14 21:47:32 right, licenses are all over the place. But in order to get sign off from a legal team I need to have some form of license to show them before I can use alpine linux 2017-06-14 21:47:35 if we promote any license its mostly MIT 2017-06-14 21:47:35 and most of the alpine tools (apk-tools excluded) are BSD or similar 2017-06-14 21:47:53 and apk-tools is likely to become that way too 2017-06-14 21:48:01 I got this approved for debian because debian has a really good legal page with some base license for their kernel and things they develop 2017-06-14 21:48:05 it's definitely not going GPLv3 :P 2017-06-14 21:48:14 jgarr: hmm 2017-06-14 21:48:20 Its a good thing no one listens to me for legal advice :) 2017-06-14 21:48:25 i could probably make such a page 2017-06-14 21:48:34 of not eis that you can look up all package licenses on http://pkgs.alpinelinux.org 2017-06-14 21:48:37 of note is* 2017-06-14 21:48:48 e.g. http://pkgs.alpinelinux.org/packages?name=linux-vanilla&branch=&repo=&arch=&maintainer= 2017-06-14 21:48:52 Shiz: that would be great. Here's an example of debian's page https://www.debian.org/legal/licenses/ 2017-06-14 21:49:27 it doesn't say everything is license $foo but at least gives some idea and warm fuzzies for lawyers 2017-06-14 21:49:28 i can tell you this, looking at debian's page: 2017-06-14 21:49:41 we only ship OSI-approved software licenses 2017-06-14 21:50:06 unless you manually activate and buil dfrom the nonfree repository, which is neither a default nor trivial :p 2017-06-14 21:50:20 yes, actually alpine is more about software freedom than debian is 2017-06-14 21:50:28 TIL there IS a nonfree repo :) 2017-06-14 21:50:30 we do not distribute the contents of our non-free repo, you have to build it manually 2017-06-14 21:50:32 :) 2017-06-14 21:51:21 Shiz: want me to open a bug report somewhere for the request? If you could put that in writing I can pass it along to my legal team. Unfortunately, they won't take my IRC logs as "in writing" 2017-06-14 21:51:23 Hmm. I'm happily surprised most of it is MIT/BSD and that there is a general atmosphere of caution against GPL. I assumed alpine was another gpl house 2017-06-14 21:51:31 jgarr: sure thing 2017-06-14 21:51:47 if you just want a statement of 'everything in alpine's main/community/testing repo is OSI-approved', I can say it in that bug report 2017-06-14 21:51:50 would that be enough of a statement? 2017-06-14 21:52:04 (trying to figure out what's best/easiest) 2017-06-14 21:52:51 kniini: gpl has never been about freedom, at least not from the developers view 2017-06-14 21:53:08 (#alpine-offtopic would be better for gpl wars) 2017-06-14 21:53:08 Maybe a link/more info of what OSI-approved is/means. As much as lawyers love legal terms they don't know crap about OSS licenses categories 2017-06-14 21:53:15 jgarr: yes, it would include that 2017-06-14 21:53:23 mostly just wondering if a bug report comment would be enough for your lawyers 2017-06-14 21:54:03 elegast: there is a atmosphere of caution against GPLv3. GPLv2-only is a perfectly fine licensing situation 2017-06-14 21:54:06 I'm not sure. probably just because any registered user to the bug system could leave that comment. They may want something that looks more official or at least can be identified as coming from a maintainer 2017-06-14 21:54:16 jgarr: i do have admin status on the bug tracker 2017-06-14 21:54:18 :p 2017-06-14 21:54:19 Shiz: you're right, not trying to start a flamewar 2017-06-14 21:54:24 https://bugs.alpinelinux.org/users/2138 2017-06-14 21:54:27 but that's largely because FSF used the 'or later' clause as an attempt to weaponise the GPL 2017-06-14 21:54:28 see 'project lists' 2017-06-14 21:54:40 'projects'* 2017-06-14 21:54:56 kanini: yeah true, (l)gpl2 isn't as restrictive 2017-06-14 21:54:59 ok, I'll see if it passes through 2017-06-14 21:55:15 if it won't it's no problem, i'm sure we can make a simple disclaimer page 2017-06-14 21:55:36 but if a bug tracker comment suffices, it's easier and won't need any other dev's involvement to make that change to the site 2017-06-14 21:55:38 :p 2017-06-14 21:57:45 kaniini: yeah don't get me wrong, I'm not saying gpl is bad per say, its just not the be all end all license that fsf makes it out to be, and the fact that when someone mentions 'open-source' everyone assumes 'gpl' 2017-06-14 21:57:51 is kind of infuriating to me 2017-06-14 21:58:55 from a developer view that is, from a user perspective, its about as good as it gets. but I wouldn't call it freedom as in free speach 2017-06-14 21:59:12 Shiz: https://bugs.alpinelinux.org/issues/7423 Thanks again! 2017-06-14 22:00:21 Anyway, shiz has a point, this isnt the place for religious licensing talk 2017-06-14 22:00:52 Sorry to instigate a war. I just want to use Alpine :) 2017-06-14 22:01:20 no wars here :p 2017-06-14 22:01:22 Well, to be fair, as far as I'm concerned, we're just having a conversation that happens to be about licensing :) 2017-06-14 22:01:38 IN ALL CAPS!!! 2017-06-14 22:02:05 But its an invitation for someone to come in be offended in one way or another 2017-06-14 22:04:03 oh brb guys, im being evicted and my laptop is without battery 2017-06-14 22:04:28 so not to worry, I havent left in defeat :p 2017-06-14 22:14:36 jgarr: https://bugs.alpinelinux.org/issues/7423 2017-06-14 22:14:38 how's this? 2017-06-14 22:19:47 That's great. I'll pass it on to my legal team and hopefully they don't want anything else 2017-06-14 22:20:31 feel free to let us/me know if something else is needed, i'm sure we can get it done 2017-06-14 22:21:48 is there a source code repo that I could link to which constitutes the "base" OS. Or maybe a build system that assembles the pieces 2017-06-14 22:22:36 the build system is abuild together with the aports repository 2017-06-14 22:22:59 https://git.alpinelinux.org/cgit/abuild/ 2017-06-14 22:23:05 GPL2 as per https://git.alpinelinux.org/cgit/abuild/tree/abuild.in 2017-06-14 22:23:14 thanks 2017-06-14 22:23:27 https://git.alpinelinux.org/cgit/aports/ 2017-06-14 22:23:36 the repository containing the package recipes for alpine 2017-06-14 22:23:44 every one of which contains a license= line :) 2017-06-14 22:23:52 which is where pkgs.a.o gets its license info from 2017-06-14 22:30:34 allrighty 2017-06-14 22:31:37 we still fighting about licensing, or did the moment pass? 2017-06-14 22:34:17 lol 2017-06-14 22:34:36 all is well in the world 2017-06-14 22:35:20 no? fair enough. back to initrd. I was working on integrating plymouth, and I'm still iffy about how much cooperation is needed from the kernel. E.g. its configured to do early modestting and such, I've included the binaries in the ramdisk image and compiled the openrc plugin which I shamelessly stole from gentoo, Is it just a matter of having openrc start and interact with plymouthd, or is there some plymouth kernel module 2017-06-14 22:35:21 involved? 2017-06-14 22:36:01 i don't think plymouth uses a kernel module 2017-06-14 22:36:06 initramfs starts plymouth 2017-06-14 22:36:07 https://wiki.gentoo.org/wiki/Plymouth 2017-06-14 22:36:19 looks like it works with openrc 2017-06-14 22:36:43 yeh, no kernel modules, but to avoid having to mod plymouth messaging into init scripts you probably need that plugin you mentioned 2017-06-14 22:37:00 I guess that's what it is about, anyway 2017-06-14 22:37:44 Mhmm. Yeah, it looked just like regular rc script, except its C 2017-06-14 22:39:05 Allright, so, it *should* just come down to configuring and booting the image, then, I think 2017-06-14 22:40:24 hmm, interesting, that Gentoo page says versions 0.9.0 and later don't work with openrc 2017-06-14 22:41:10 Yeah, I saw that, it mentions multi seat, and, I *think* that one of the patches thats already in community/plymouth takes care of that 2017-06-14 22:41:10 but that seat rule mentioned, I think the Alpine package has that one, and it should fix this 2017-06-14 22:48:14 hmm. Instead of rebuilding the whole ramdrive image, can I just specify two? 2017-06-14 22:50:58 e.g. put INITRD initramfs-hardened initramfs-gfxboot 2017-06-14 22:51:22 don't think it can work like that 2017-06-14 22:51:25 no 2017-06-14 22:51:38 Hmm. I recall grub being able to do such things 2017-06-14 22:52:21 I'll look into that later, idk maybe my memory is fuzzy 2017-06-14 22:54:19 but how does this fit into our ultimate goal of a gnome 3 powered toaster running on nvidia drivers? 2017-06-14 22:54:47 oh i didn't know that was a goal 2017-06-14 22:55:07 all distributions expand until they can run on toasters 2017-06-14 22:55:07 I don't know about gnome 3 but I can definitly get the toaster thing to work 2017-06-14 22:55:15 Heh. 2017-06-14 22:55:39 I think all it takes is a kernel flag to disable cpu idling 2017-06-14 22:55:46 and it will make toast all day 2017-06-14 22:55:58 yes, especially on AMD Ryzen i hear 2017-06-14 22:56:04 so that would be, like, a toaster that looks like windows 10 but worse, and which because of the lack of hardware vendor provided drivers couldn't properly handle temperature 2017-06-14 22:56:28 Pretty much, yes. I think that meets said goal, no? 2017-06-14 22:56:46 I mean do we really need gnome 3 for that? 2017-06-14 22:57:19 I prefer shell access to _my_ toasters 2017-06-14 22:57:20 of course we do 2017-06-14 22:57:22 Cuz idunno about you guys, but I can do without client side window decorations 2017-06-14 22:58:00 Have you ever *tried* decorating a toaster? 2017-06-14 22:59:00 Hehe. You guys remember that flying toaster screensaver? 2017-06-14 22:59:25 kaniini: any clue when scipy and numpy will be hitting stable release? (if they haven't already) 2017-06-14 22:59:48 After Dark 2017-06-14 23:00:11 So much nostalgia 2017-06-14 23:00:19 /w/w 26 2017-06-14 23:00:47 https://www.youtube.com/watch?v=mjlusi_h_XA 2017-06-14 23:01:13 omg. 2017-06-14 23:01:14 yes. 2017-06-14 23:03:37 That should be the default plymouth theme 2017-06-14 23:31:16 Is it good practice to include /usr/lib files in initrd, or should I recompile packages to be in /lib? 2017-06-14 23:32:36 Its technically userspace stuff 2017-06-14 23:32:49 former is fine 2017-06-14 23:33:06 allrighty 2017-06-14 23:33:13 see also zfs.files :p 2017-06-14 23:33:33 oh yeh i see 2017-06-14 23:33:56 why do you need to include stuff in /usr/lib? 2017-06-14 23:34:05 mkinitfs will scan binaries for library deps automatically 2017-06-14 23:34:47 It also includes theme files, I dont think those would be picked up 2017-06-14 23:35:07 but I wasn't aware it would scan for lib files automagicly 2017-06-14 23:35:15 it uses elfscan? 2017-06-14 23:36:06 lddtree 2017-06-14 23:36:30 Ah ok 2017-06-14 23:42:31 ok cool, not having to specify any lib files makes it alot cleaner 2017-06-14 23:43:43 I didn't see any lib files in any of the other *.files, so I just assumed everything was statically built 2017-06-15 05:00:25 is there a way to connect to wifi that has no password protection? 2017-06-15 05:00:58 if iwconfig is available 2017-06-15 05:00:59 iwconfig wlan0 essid "networkname" 2017-06-15 05:01:21 haven't tried 802.11 on alpine though so not 100% sure, sorry 2017-06-15 05:01:56 and that would automatically get an ip addy from dchcp 2017-06-15 05:03:25 no, that'd just associate the interface to the network 2017-06-15 05:03:39 you'd still need to run a dhcp client or assign an ip manually 2017-06-15 05:04:06 so udhcpc -i wlan0 2017-06-15 05:04:14 ? 2017-06-15 05:04:21 yep should work 2017-06-15 05:04:36 word, thanks. time to test this out 2017-06-15 05:04:38 iwconfig will show signal strength & other information if the connection is successful 2017-06-15 05:04:48 you can verify that before running dhcp 2017-06-15 10:18:27 hi people 2017-06-15 10:39:44 hi 2017-06-15 10:40:16 hi, i noticed that the mono package was removed https://pkgs.alpinelinux.org/packages?name=mono&branch=&repo=&arch=&maintainer= . It previously existed and was used here... https://github.com/mattaitchison/alpine-mono/blob/master/Dockerfile Is there a way I can find out why, and if it can be re-instated? 2017-06-15 10:41:50 hmm 2017-06-15 10:42:20 chrismcv: https://git.alpinelinux.org/cgit/aports/commit/testing/mono/APKBUILD?id=2fb7351e135d135f0ad5a83f538764d696b01e7b 2017-06-15 10:42:24 kaniini likely knows 2017-06-15 10:53:29 :Shiz thanks 2017-06-15 13:11:49 hey guys, I'm confused, I'm running into a relocation error when I do this: cc -Bsymbolic-functions -fPIC -shared -Wl,-soname,plymouth.so plymouth.o -o plymouth.so -leinfo -lrc 2017-06-15 13:11:56 however, 2017-06-15 13:12:06 if I split the compile and linking command 2017-06-15 13:12:09 everything is fine 2017-06-15 13:12:11 e.g.: 2017-06-15 13:12:41 this: /usr/x86_64-alpine-linux-musl/bin/ld -shared -lrc -leinfo -fPIC -o plymouth.so plymouth.o 2017-06-15 13:12:43 works just fine 2017-06-15 13:13:56 well 2017-06-15 13:14:02 gcc will pass some extra arguments to the linker 2017-06-15 13:14:09 so what you pass to the linker is not identical to what gcc passes 2017-06-15 13:14:18 among others, your manual invocation doesn't include the -soname part 2017-06-15 13:14:31 and also likely doesn't include the -pie gcc likely passes 2017-06-15 13:14:37 oh, i think i copied the wrong commandline 2017-06-15 13:14:53 i did include the -soname part iirc 2017-06-15 13:15:46 how would I find out exactly which flags are passed? 2017-06-15 13:16:00 C is not my forte 2017-06-15 13:17:18 cc -v says my "man cc" 2017-06-15 13:17:48 ah, right, maybe I should install some manpages 2017-06-15 13:18:07 but thanks IcePic, thats what im looking for 2017-06-15 13:20:19 ah ok, thats alot of extra flags I hadn't in my manual compile 2017-06-15 13:20:58 well linking phase mostly, by the looks of it 2017-06-15 13:27:07 Whats the sensible thing to do here, should I just have the package split the compile and link phase so "it works". The prudent thing to do with be to identify the exact flag that breaks things, no? I imagine all those extra flags are there for good reason. 2017-06-15 13:29:22 I don't usually invoke the linker directly, even if linking is all I'm doing. 2017-06-15 13:29:42 Let the compiler driver (gcc) do its job. :) 2017-06-15 13:30:36 Yeah, I have to agree, though its not my fave passtime 2017-06-15 13:30:42 Oh, that's already what you're doing. I misread. 2017-06-15 13:31:39 do you link using cc or ld? 2017-06-15 13:33:10 First I just invoked cc in PATH to compile, and then I invoked musl's ld, this worked fine 2017-06-15 13:33:45 But then when I tried to have the cc do the linking it failed 2017-06-15 13:34:19 Try using cc to link, but use the exact same flags as when you're compiling? 2017-06-15 13:34:46 I did, but, cc includes a *whole* lot of other flags into mix 2017-06-15 13:35:40 Right. 2017-06-15 13:35:44 I mean pass it the exact same flags. 2017-06-15 13:36:05 Is this source public? I could give it a shot. 2017-06-15 13:36:25 It is, hang on 2017-06-15 13:38:09 https://github.com/aidecoa/plymouth-openrc-plugin.git 2017-06-15 13:38:58 2figuring out the flag that breaks would be the best idea, yes 2017-06-15 13:39:00 404. 2017-06-15 13:39:16 https://github.com/aidecoe/plymouth-openrc-plugin 2017-06-15 13:39:21 elegast made a typo 2017-06-15 13:39:38 I didn't expect packaging it up would've taken up so much effort after i ended up with a working binary 2017-06-15 13:39:44 =) 2017-06-15 13:40:57 btw 'musl's ld' is just binutils ld 2017-06-15 13:41:01 the same as on any other system 2017-06-15 13:41:03 :p 2017-06-15 13:41:33 Oh ok, I wasn't sure so was explicit about it 2017-06-15 13:42:12 I figured maybe it would be a busybox implementation 2017-06-15 13:42:31 But I think I confuzzled that with ld and ldd being the same thing for musl 2017-06-15 13:43:11 Thanks for taking look btw, Maxpm 2017-06-15 13:43:16 a look* 2017-06-15 13:43:21 Sure. Just give me a bit. 2017-06-15 13:44:03 Mhmm. It's only depency should be openrc-dev, i think 2017-06-15 13:44:32 ldd is something different, and actually part of musl :p 2017-06-15 13:45:47 Oh, I could've sworn I read ld checks if its argv0 is named ldd and switches mode 2017-06-15 13:46:50 https://pbs.twimg.com/media/DCUEBleUQAAqBU9.jpg:large 2017-06-15 13:46:51 Eh. I'm about as fuzzy as it gets sometimes 2017-06-15 13:47:14 lol 2017-06-15 13:47:25 Linux lies in general about memory, doesn't it? 2017-06-15 13:47:45 isn't that the whole reason theres this nasty oom thing 2017-06-15 13:48:27 what do you mean lies about memory? 2017-06-15 13:48:29 yes, it almost never says no to malloc() 2017-06-15 13:48:53 i mean 2017-06-15 13:49:19 elegast: What's the error, exactly? It seems to compile and link fine for me. 2017-06-15 13:49:30 no way 2017-06-15 13:49:35 :) 2017-06-15 13:50:24 oh ffs, my copy paste broke switching into lxc 2017-06-15 13:50:55 gimme a sec 2017-06-15 13:51:45 librc.a(librc.o): relocation R_X86_64_PC32 against symbol `stderr' can not be used when making a shared object; recompile with -fPIC 2017-06-15 13:51:48 It might be significant that I'm using a fresh openrc build. 2017-06-15 13:52:26 my openrc is from v3.6main 2017-06-15 13:52:32 binary install 2017-06-15 13:57:24 is your openrc fresh as in built from aports v3.6 head, or as in aports master? 2017-06-15 13:58:13 As in https://github.com/OpenRC/openrc master. :) 2017-06-15 13:58:34 I'm not actually on Alpine at the moment. 2017-06-15 13:58:42 ah 2017-06-15 13:58:49 But that error is confusing me. 2017-06-15 13:59:20 oh. so you're not running musl? 2017-06-15 14:00:06 No, but I can give that a shot. One minute. 2017-06-15 14:00:27 You're a boss :) 2017-06-15 14:08:33 I think theres parts being statically included that are generaded from assembly but not relocatable 2017-06-15 14:08:55 written in assembly* 2017-06-15 14:10:54 idk it's a bit beyond my understanding 2017-06-15 14:12:55 its pulling in a lot of musl *S.o related to c runtime, is that expected behaviour for building shared libraries? 2017-06-15 14:14:45 hey, anyone using lxd? 2017-06-15 14:15:02 no, just lxc 2017-06-15 14:21:46 my resolv.conf keeps being overwritten 2017-06-15 14:21:50 I guess some DHCP thing 2017-06-15 14:21:55 any idea how to get rid of that? 2017-06-15 14:22:11 elegast: With musl, I can't even compile openrc. Relocation errors. A bit over my head, I'm afraid. 2017-06-15 14:23:32 allright, thanks for trying though! And yesh same here, not a seasoned C programmer, let alone musl 2017-06-15 14:24:14 stevenroose: I think you're supposed to change the stuff in /etc/resolv.conf.d instead of counting on resolv.conf itself. 2017-06-15 14:24:16 stevenroose: what dhcp are you on? 2017-06-15 14:24:59 If you chosen the defaults, /etc/udhcpd.conf is a good place to start 2017-06-15 14:25:15 elegast, what do you mean? 2017-06-15 14:25:30 I guess the defaults, yes 2017-06-15 14:25:38 there's different dhcp clients, and i imagine they behaving differently 2017-06-15 14:26:44 elegast, I don't have /etc/resolv.conf.d 2017-06-15 14:26:56 I have the udhcp 2017-06-15 14:27:36 elegast, if I edit /etc/udhcpd.conf, do I have to reload something? 2017-06-15 14:27:42 rc-service udhcpd restart? 2017-06-15 14:28:18 er, my bad 2017-06-15 14:28:27 thats the daemon 2017-06-15 14:29:30 elegast, /etc/dhcp/dhclient.conf is that for a different client? 2017-06-15 14:29:36 then udhcpd? 2017-06-15 14:30:11 yeah, im not exactly sure, the d is confusing me, its either a daemon for the client, or a dhcp server, or both 2017-06-15 14:31:27 sorry, im having some issues of my own, I think my laptop just died 2017-06-15 14:32:52 im pretty sure udhcpd manages the dhcp client in the background though, either way "restarting" it, wont fix anything 2017-06-15 14:33:19 It would simply fetch a lease, and overwrite your resolv.conf again 2017-06-15 14:36:09 elegast, yeah but I guess just changing a config file without restarting anything also won't change anything :o 2017-06-15 14:36:10 sorry though, my equipment is catching fire here and im fairly new to alpine, maybe someone else can chime in and help you out. In general though, either resolv.conf.d, and /etc/networking/interfaces are places from where you hook the dhcp client, im not exactly sure on alpine though 2017-06-15 14:36:12 but ok tanks 2017-06-15 14:36:21 haha, no worries 2017-06-15 14:36:24 yes ofcourse 2017-06-15 14:36:28 go ahead and extinguish your fire 2017-06-15 14:38:36 I thought dhcpd was the server and dhcpcd was the client. 2017-06-15 14:41:15 I removed resolvconf on suggestion of a stackexchange page 2017-06-15 14:41:28 but I cant reboot for another few days, so I'll have to wait :D 2017-06-15 14:48:06 stevenroose: so you just want to make udhcpc not touch anything in /etc/resolv.conf, right? 2017-06-15 14:52:20 kahiru, uhu 2017-06-15 14:52:39 RESOLV_CONF=no 2017-06-15 14:53:06 ^ this 2017-06-15 14:53:14 I think that goes in /etc/udhcpc/udhcpc.conf 2017-06-15 14:53:16 into /etc/udhcpc/udhcpc.conf 2017-06-15 14:53:18 which likely doesn't exist yet 2017-06-15 14:53:19 exactly 2017-06-15 14:54:05 or you can put there dns="mydns1 mydns2" to get "nameserver mydns1" and "nameserver mydns2" entries in /etc/resolv.conf 2017-06-15 14:54:12 elegast, and then rc-service udhcpc restart? 2017-06-15 14:54:15 or reboot? 2017-06-15 14:54:26 just a service restart should suffice 2017-06-15 14:54:40 it doesn't have a standalone service afaik 2017-06-15 14:54:53 right, thats the dhcp daemon 2017-06-15 14:54:58 so I'd either play with ifup&ifdown or restart the whole networking 2017-06-15 14:55:11 well zap neworking service, or reboot i guess 2017-06-15 14:55:15 damn me 2017-06-15 14:55:31 I tried that before, but see now that I copied incorrectly. I had "ESOLV_CONF=no" 2017-06-15 14:55:35 oh 2017-06-15 14:55:37 bummer 2017-06-15 14:55:42 lol 2017-06-15 14:56:01 service `udhcpc' does not exist 2017-06-15 14:56:14 also is it even necessary to restart anything? shouldn't just putting correct nameservers into /etc/resolv.conf be enough? 2017-06-15 14:56:22 yes, its just udhcpd which is a dhcp daemon 2017-06-15 14:56:34 found it, thanks a lot 2017-06-15 14:57:29 but like kahiru says, resolv.conf changes are live as soon as the file is changed 2017-06-15 14:57:55 but I'm not sure whether udhcp rereads its configuration when it asks for a renew 2017-06-15 14:59:09 it looks like it doesn't, skimming over default.script 2017-06-15 15:01:25 actually, it doesn't look like udhcpc is a process that stays resident, so it should re-evaluate default.script 2017-06-15 15:07:58 ok so it looks like plymouth with openrc is working just fine, except for some packaging issues with regards to linking manually versus having cc do the link phase 2017-06-15 15:09:05 Hooray. 2017-06-15 15:09:08 yayy 2017-06-15 15:09:16 and the handoff from plymouth to xdm isn't seamless, but that's to be expected 2017-06-15 15:14:26 ah, i spoke too soon, its using text fallback 2017-06-15 15:15:45 still though, it looks spiffy :) 2017-06-15 15:42:30 Hello! 2017-06-15 15:43:49 I rewrote the article for installing Alpine on a Virtual Box on the wiki, as that was slightly outdated and did not link to some extensions. 2017-06-15 15:45:23 Do I need to inform anyone that I did so, to get the article published, or does that happen automatically? 2017-06-15 15:48:46 Bel-Chion: Yes, that is in the contract implied compliance and legal commitment to but partaking. Better to ask and ensure a 100% ascertained direction than having to argue against the prosecutor, the law fees are wearing on all parties, not to mention the time. Thank you. 2017-06-15 15:52:39 what 2017-06-15 15:52:55 Bel-Chion: if you edited it on the wiki, everyone should see the chnages immediately 2017-06-15 15:53:02 thanks for helping out, by the way! 2017-06-15 15:56:07 You are welcome. Alpine works really well, even as a beginner I had little problems trying it. 2017-06-15 16:01:33 thank you for the wiki addition; virtualbox installs have been one of the most frequently asked questions here during the time I've been on the channel 2017-06-15 16:34:57 Hi friends. What are you all planning to do now that grsecurity's no longer being released publicly? 2017-06-15 17:44:37 absolutely nothing 2017-06-15 17:44:46 see how things play out 2017-06-15 17:44:52 if something good comes along, maybe we offer it 2017-06-15 17:44:54 who knows 2017-06-15 18:59:25 I'm getting this error when trying to open audio mixer: GStreamer was unable to detect any sound devices. Some sound system specific GStreamer packages may be missing. It may also be a permissions problem. 2017-06-15 18:59:49 I've installed gstreamer packages, which don't seem to fix it 2017-06-15 19:01:01 specifically these: gstreamer1, gst-plugins-base1-dev 2017-06-15 19:04:04 you shouldn't need the -dev packages 2017-06-15 19:04:25 you probably want gst-plugins-good1 2017-06-15 19:04:46 and possibly gst-plugins-ugly1 if that doesn't work 2017-06-15 19:08:34 Shiz, nope 2017-06-15 19:08:47 neither wor 2017-06-15 19:08:51 work* 2017-06-15 19:11:41 i'll try rebooting maybe 2017-06-15 19:13:21 works now, thanks 2017-06-15 19:13:56 weird that it needs a reboot 2017-06-15 19:15:14 Sounds more like windows 2017-06-15 19:18:47 Shiz, maybe it was that i had to log in again 2017-06-15 19:48:54 What are the practical differences between the virtual and mini-root releases? 2017-06-15 19:49:51 <_ikke_> miniroot contains only the minimum userspace programs 2017-06-15 19:50:01 <_ikke_> mainly for things like containers 2017-06-15 19:50:04 no kernel or modules? 2017-06-15 19:50:17 <_ikke_> Not sure, but would not expect so 2017-06-15 19:50:39 <_ikke_> considering the size, it almost certainly does not contain a kernel 2017-06-15 19:50:45 Aha. 2017-06-15 19:50:51 Thanks. 2017-06-15 19:52:44 miniroot has no kernel or modules, no 2017-06-15 20:08:02 is there an atom package? 2017-06-15 20:11:40 <_ikke_> https://pkgs.alpinelinux.org/packages?name=atom&branch=&repo=&arch=&maintainer= 2017-06-15 20:12:21 guess not 2017-06-15 20:15:55 I can't seem to figure out how to connect to wifi that has no encryption. I can connect to wifi with wpa_supplicant easily. this has got me stumped. 2017-06-15 20:16:21 m4chm4n, encrypt your wifi 2017-06-15 20:17:14 Thats obvious.... But what if I have no control over that? ie public wifi hotspots.... 2017-06-15 20:17:35 tell them to encrypt it 2017-06-15 20:17:53 hahaha!! 2017-06-15 20:17:54 also dont use unencrypted public wifi lol 2017-06-15 20:18:01 thats a terrible idea 2017-06-15 20:18:33 well thanks Capt. Obvious!!! 2017-06-15 20:23:21 when attempting to use alpine-3.6.1 with https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage to build an iso for x86, the image is unable to load a kernel to boot, is there anything to try to get it to work? 2017-06-15 20:33:45 the script has a bug 2017-06-15 20:33:57 kernel_flavors="" will make it build no kernels 2017-06-15 20:34:00 delete that line 2017-06-15 20:34:49 k, thanks 2017-06-15 21:22:05 hello @ all 2017-06-15 21:22:44 you earlyer dicussed about the images ... 2017-06-15 21:23:22 What ist the diffrence between virt and standard? 2017-06-15 21:23:55 How come's that the virt iso is that small? 2017-06-15 21:24:33 They both run's under xen pretty good 2017-06-15 21:24:57 anyone having trouble connect to phones for sharing files with bluetooth? 2017-06-15 21:25:14 I am using bluez 2017-06-15 21:25:26 feuerteufel: the virt iso has a smaller kernel with less driver support outside what VM software emulates 2017-06-15 21:28:56 Xe: thank's, but it's al lot of difrence 36MB to 112MB ... 2017-06-15 21:29:10 feuerteufel: it's a lot less drivers :D 2017-06-15 21:31:28 so if I'm working under xen or what ever it's OK to use that one? 2017-06-15 21:31:37 yep 2017-06-15 21:32:14 Great, thank's again!! 2017-06-15 21:34:28 np 2017-06-15 21:40:47 do the OpenJDK packages include javaws? 2017-06-15 21:41:26 people still use java web start? 2017-06-15 21:42:14 Heh. 2017-06-15 21:45:14 im having trouble connecting to my phone with bluetoothctl. I cant seem to send files from my phone to my computer. 2017-06-15 22:54:56 cu 2017-06-16 00:51:43 <_spOOn_> interesting questions being thrown around on the alpine-devel mailing list... 2017-06-16 00:52:42 <_spOOn_> didnt realise there were options being investigated to change out some of alpines core components... 2017-06-16 00:53:09 That sounds scary. 2017-06-16 00:59:27 inb4 systemd 2017-06-16 01:01:50 D: 2017-06-16 01:07:14 not on my body 2017-06-16 01:28:53 _spOOn_: sounds scary to the point i don't even want to read the threads 2017-06-16 01:32:39 Wanna give us a summary? :P 2017-06-16 02:00:17 Couldn't figure out if there's a way to list back virtual package sets (i.e. apk add -t foo) anyone know? 2017-06-16 04:42:02 I'm getting permission denied errors as root, not sure why. This command demonstrates the issue. Why is root getting permission denied? 2017-06-16 04:42:02 docker run -it alpine:edge /bin/ash -c 'apk add npm --update; npm install -g brackets' 2017-06-16 04:42:23 sorry, wrong command. this one: docker run -t mhart/alpine-node /bin/ash -c 'npm install -g brackets' 2017-06-16 09:38:35 hI 2017-06-16 09:38:37 is anyone here? 2017-06-16 09:40:45 maybe 2017-06-16 09:42:52 oh god thanks! 2017-06-16 09:42:55 I need help 2017-06-16 09:43:08 so I'm using alpine linux to share a folder with samba 2017-06-16 09:43:14 and this folder is a fuse filesystem 2017-06-16 09:43:24 I can't make it mount at boot 2017-06-16 09:43:28 but after boot works fine 2017-06-16 09:43:46 where I can find the logs ? 2017-06-16 09:45:05 check out /etc/rc.conf 2017-06-16 09:45:10 there are logging options there 2017-06-16 09:45:41 I'm using rclone 2017-06-16 09:45:49 I have a file rclonemount.start 2017-06-16 09:45:58 in local.d 2017-06-16 09:46:17 why not do it from fstab? 2017-06-16 09:46:26 how? 2017-06-16 09:46:32 is that better? 2017-06-16 09:46:37 I'm new to alpine 2017-06-16 09:46:39 I usually use systemd 2017-06-16 09:46:47 ouch 2017-06-16 09:46:51 it has nothing to do with alpine 2017-06-16 09:47:23 fstab is default for all linux distro's afaik 2017-06-16 09:47:29 is pastebin down? 2017-06-16 09:47:33 I never had to use it before 2017-06-16 09:47:51 and rclone needs to read it's config file to mount the cloud storage 2017-06-16 09:48:16 <_ikke_> http://tpaste.us 2017-06-16 09:49:17 apk add tpaste 2017-06-16 09:49:27 to be true I has following this guide https://github.com/ajkis/scripts/issues/2 2017-06-16 09:49:39 idk if was because of different versions but almost everything he did has broken 2017-06-16 09:49:43 i figured out almost everything but this 2017-06-16 09:50:52 fstab is pretty standard for automounting on boot 2017-06-16 09:51:52 I usually only use systemd 2017-06-16 09:52:17 systemd also has fstab :P 2017-06-16 09:53:02 http://www.pclosmag.com/html/Issues/200709/page07.html 2017-06-16 09:53:28 it won't work for me 2017-06-16 09:54:48 there's nothing in rc.conf about logs 2017-06-16 09:55:21 there is: 2017-06-16 09:55:29 rc_logger= and rc_log_path= 2017-06-16 09:56:26 I'm going to reboot 2017-06-16 09:56:57 I'm seeing a lot of permission denied... 2017-06-16 10:00:04 ? 2017-06-16 10:01:15 can't be much help without more precise errors/info 2017-06-16 10:03:45 https://imgur.com/a/74U1Q 2017-06-16 10:04:43 it seems like you're trying to execute /var/log/rc.log in /etc/rc.conf? 2017-06-16 10:05:09 what does line 48 of rc.conf say? 2017-06-16 10:06:01 a /var/log/rc.log 2017-06-16 10:06:07 I need to insert A or wouldn't send 2017-06-16 10:07:15 make that rc_log_path="/var/log/rc.log" 2017-06-16 10:08:29 k so now I reboot without errors 2017-06-16 10:08:40 but I don't see anything in the log related to rclone 2017-06-16 10:08:50 or my file rclonemount.start in local.d 2017-06-16 10:08:56 is it chmod +x at least? 2017-06-16 10:09:00 the start file 2017-06-16 10:09:43 yes 2017-06-16 10:10:08 I would paste my file into pastebin but it's down 2017-06-16 10:10:35 hello guys! 2017-06-16 10:10:49 you can do: curl -F 'tpaste=<-' http://tpaste.us/ < /etc/local.d/rclonemount.start 2017-06-16 10:10:57 are there some instructions how to make Generic arm kernel work on raspberry ? 2017-06-16 10:11:29 http://tpaste.us/Be9b 2017-06-16 10:11:51 well that's not right, is it, all those newlines 2017-06-16 10:12:10 well at some point it did work 2017-06-16 10:12:16 like I said I'm following a guide 2017-06-16 10:12:39 https://github.com/ajkis/scripts/issues/2 2017-06-16 10:12:44 it's exactly the same he used 2017-06-16 10:12:48 you should add \ at the end of every line, from rclone .. to --log-level .. 2017-06-16 10:12:54 log-file* 2017-06-16 10:13:10 or just put it all on one line 2017-06-16 10:14:01 it would be nice to see at least a log saying " the file is corrupted / formatted wrong " 2017-06-16 10:14:17 this is the first linux distro I used with so poor logging 2017-06-16 10:14:18 well, it's odd that you didn't see any errors 2017-06-16 10:14:21 but I'll try thanks 2017-06-16 10:14:51 alpinelinux it isn't 'corrupted' 2017-06-16 10:15:03 its a shell script and github removed characters that should be visible in the script 2017-06-16 10:15:28 oh really? 2017-06-16 10:15:36 that's strange since github is there for ... code 2017-06-16 10:15:42 its in a comment 2017-06-16 10:15:45 the comments use markdown formatting 2017-06-16 10:15:54 which requires special care to display code properly 2017-06-16 10:15:56 https://paste.ee/r/ZHSeB 2017-06-16 10:16:01 thats what you need, those backslashes 2017-06-16 10:16:09 thats really just one command broken up onto multiple lines 2017-06-16 10:16:29 also make sure you actually have bash installed ;p 2017-06-16 10:16:37 it doesnt need to be bash 2017-06-16 10:16:41 you can change that to /bin/sh 2017-06-16 10:16:45 or that, sure 2017-06-16 10:16:47 oh it isn't installed by default? 2017-06-16 10:16:49 no 2017-06-16 10:16:54 the default shell is ash 2017-06-16 10:17:08 since when? 2017-06-16 10:17:13 since alpine linux 2017-06-16 10:17:19 not every distro uses the same shell by default 2017-06-16 10:17:23 I don't understand then 2017-06-16 10:17:27 that guide was written like that 2017-06-16 10:17:28 and it worked 2017-06-16 10:17:37 well it depends on how you executed the script 2017-06-16 10:17:45 the syntax in that github link is incorrect 2017-06-16 10:17:47 so I need to apk add bash 2017-06-16 10:17:48 ? 2017-06-16 10:17:49 it will not work as visible 2017-06-16 10:17:51 though the lack of backslashes should have never worked 2017-06-16 10:17:55 you can just change the top of the script to 2017-06-16 10:17:56 #!/bin/sh 2017-06-16 10:17:58 and it will work fine 2017-06-16 10:18:16 and isn't that the bash? 2017-06-16 10:18:20 no 2017-06-16 10:18:26 /bin/sh can be any number of different shells 2017-06-16 10:18:55 I put everything in a single line, didn't work it's probably what you are saying...will fix and seex 2017-06-16 10:19:50 /bin/sh is the lowest common denominator between other shells like bash, ash, ksh 2017-06-16 10:21:11 didn't work at all 2017-06-16 10:21:46 http://tpaste.us/bKwW 2017-06-16 10:22:51 ive never used rclone before but based on the example i'd expect rcloneremote to be a hostname 2017-06-16 10:22:55 is 'crypt' the hostname of your target? 2017-06-16 10:23:25 yes, it's the name of my remote in my rclone.conf 2017-06-16 10:23:27 you forgot the / after crypt: if that matters, also wonder if this is a potential network-is-not-up thing.. 2017-06-16 10:23:53 I do think you have to authenticate with my cloud provider to mount it 2017-06-16 10:24:00 i would also remove the --quiet option while testing/debugging just to get more information 2017-06-16 10:24:01 with systemd I start it after the network 2017-06-16 10:25:19 does local.d start after or before the network? 2017-06-16 10:25:31 maybe I should add some waiting time... 2017-06-16 10:25:50 local.d occurs after all other startup items afaik 2017-06-16 10:25:52 yes it starts after everything 2017-06-16 10:27:44 oh I might have found it 2017-06-16 10:27:49 now at least it's logging to my file 2017-06-16 10:27:55 seems I can't use -q with logging 2017-06-16 10:32:19 all fine! 2017-06-16 10:32:23 Thanks for the awesome support guys! 2017-06-16 10:32:58 but the whole goal of this was to use a system I could fit in less than 500 MB and I ended up with almost 700 MB because of all stuff I installed haha 2017-06-16 10:33:40 i'd imagine you can slim it down, 700MB is a *lot* 2017-06-16 10:34:47 well all I really need is SSH, samba & rclone 2017-06-16 10:34:51 what do you suggest me to remove? 2017-06-16 10:35:17 I had to install both wget and curl too because for some reason it would fail with the 2 urls I had to donwload 2017-06-16 10:37:26 as you can guess it's installed in virtualbox so I'll need to research how to shrink the virtual hd 2017-06-16 10:48:46 k guess everyone got afk right? 2017-06-16 10:59:32 hey, are there any instructions how to run generic arm kernel, if someone knows i'll appreciate :) 2017-06-16 11:00:16 anyone home? 2017-06-16 11:00:42 geeze 2017-06-16 11:13:38 I need to go 2017-06-16 11:13:43 thanks for the awesome support guys! 2017-06-16 12:44:50 why does this produce "permission denied" when it's being run as root? shouldn't root have privilege to do everything? docker run -t mhart/alpine-node /bin/ash -c 'npm install -g brackets' 2017-06-16 12:45:30 <_ikke_> sabrehagen: The kernel might still block root from doing things, especially in hardened kernels (not sure what the issue is here though) 2017-06-16 12:46:41 _ikke_: interesting, thanks. hadn't run into that case before. 2017-06-16 12:47:14 the internet says you can add --unsafe-perm to your npm install 2017-06-16 12:48:45 also install doesn't actually run everything as root 2017-06-16 12:49:51 Bun: --unsafe-perm fixes it! amazing! 2017-06-16 12:53:01 it seems --unsafe-perm defaults to false when run as root, as opposed to defaulting to true when running as a regular user 2017-06-16 14:20:36 Shiz: Where is it? 2017-06-16 14:20:40 How is it useful? 2017-06-16 14:20:49 ? 2017-06-16 14:21:11 Shiz: quickfix 2017-06-16 14:21:15 Do you use it on :make? 2017-06-16 14:21:43 you sure you shouldn't be talking to consus instead? 2017-06-16 14:21:45 :p 2017-06-16 14:21:48 Or if you just run :cc there will be error? 2017-06-16 14:21:52 Wait 2017-06-16 14:21:53 ACTION checks 2017-06-16 14:22:01 ops 2017-06-16 14:22:05 He isn't here 2017-06-16 14:23:27 pickfire: I can send you a screenshot if you want :D 2017-06-16 14:23:50 consus: That would be great! 2017-06-16 14:24:35 consus: http://ix.io/xA8.png 2017-06-16 14:24:38 Previous message 2017-06-16 14:24:49 Err 2017-06-16 14:24:54 Unreadable 2017-06-16 14:25:18 How come? 2017-06-16 14:25:22 dunno 2017-06-16 14:25:27 Looks like a binary junk 2017-06-16 14:25:42 consus: That's a png, of course unreadable. 2017-06-16 14:25:47 Err 2017-06-16 14:25:48 now that I've been writing a little reposync tool for my own use, I notice new packages keep arriving to the mirror almost every time I run it 2017-06-16 14:25:50 Don't have image viewer? 2017-06-16 14:25:54 Well 2017-06-16 14:26:02 Why is my firefox does not want to open it =/ 2017-06-16 14:26:24 Yeah, firefox fails to open that, mime error 2017-06-16 14:26:28 and I wonder if the repositories actually stay in sync if files keep coming in with this interval... 2017-06-16 14:26:29 I think someone mentioned it back then. 2017-06-16 14:26:34 :D 2017-06-16 14:26:38 Use anon.img 2017-06-16 14:26:42 *anonimg 2017-06-16 14:26:52 consus: Have feh? 2017-06-16 14:26:58 I read it 2017-06-16 14:27:04 But the next time use anonimg 2017-06-16 14:27:08 It would be simpler 2017-06-16 14:27:08 What is that? 2017-06-16 14:27:12 A hosting 2017-06-16 14:27:15 Ah 2017-06-16 14:27:27 or just imgur lol 2017-06-16 14:27:40 curl anonimg.com 2017-06-16 14:27:41 Redirecting⏎ 2017-06-16 14:27:48 Not useful at all and took me so long. 2017-06-16 14:30:18 http://i.imgur.com/LL4RCCV.png 2017-06-16 14:30:49 That thing in in the bottom -- quickfix 2017-06-16 14:31:06 Hi folks, how do you guys deal with lxc where the containers are *not* in the standard location (/var/lib/lxc) ? 2017-06-16 14:31:23 And it contains two defs 2017-06-16 14:31:24 But I don't get to know how you get that thing in the bottom. 2017-06-16 14:31:25 In a.c 2017-06-16 14:31:27 I'm looking for a global config switch to change the basedir, but I can't find any 2017-06-16 14:31:27 And in b.c 2017-06-16 14:31:32 Wow, edit compile test in the same vim. 2017-06-16 14:31:42 It's a plugin 2017-06-16 14:31:46 second 2017-06-16 14:32:15 Plugin 'ronakg/quickr-cscope.vim' 2017-06-16 14:32:34 http://i.imgur.com/LL4RCCV.png 2017-06-16 14:32:45 I get that when I do :make 2017-06-16 14:32:51 Aha 2017-06-16 14:33:02 Make uses this quickfix too 2017-06-16 14:33:06 consus: So you do testing in vim ass well? 2017-06-16 14:33:06 CScope uses it 2017-06-16 14:33:08 Latex uses it 2017-06-16 14:33:09 How? 2017-06-16 14:33:13 Latex as well? 2017-06-16 14:33:18 Yep 2017-06-16 14:33:25 pdflatex print it's shit there 2017-06-16 14:33:28 Never used Latex in quickfix list. 2017-06-16 14:33:38 It prints errors there 2017-06-16 14:33:41 Like make 2017-06-16 14:33:42 How? 2017-06-16 14:33:55 internal vim api I guess 2017-06-16 14:34:00 I do pdflatex pdf myself in another terminal most of the time. 2017-06-16 14:34:04 Oh 2017-06-16 14:34:05 Well 2017-06-16 14:34:15 There is a big fat plugin for vim 2017-06-16 14:34:17 That handles latex 2017-06-16 14:34:27 I do not remember the name 2017-06-16 14:34:32 I used a while ago 2017-06-16 14:34:36 *I used it 2017-06-16 14:34:42 there's about 3 i think 2017-06-16 14:35:02 I think have seen quite a few big fat latex plugin, didn't even give a second thought to try it. 2017-06-16 14:35:13 well im sure theres many more tex plugins, but 3 big ones 2017-06-16 14:35:15 Ah 2017-06-16 14:35:17 It was good 2017-06-16 14:35:28 I guess vim-latex 2017-06-16 14:36:19 you're using pdflatex though, does that have unicode and opentype support yet? 2017-06-16 14:36:31 I'm not using it now 2017-06-16 14:36:35 ah ok 2017-06-16 14:36:45 Maybe it was luatex 2017-06-16 14:36:49 I really can't remember 2017-06-16 14:36:58 It was 5 years ago 2017-06-16 14:37:12 hehe, ancient history 2017-06-16 14:37:20 Damn I'm old 2017-06-16 14:37:25 =/ 2017-06-16 14:37:26 =\ 2017-06-16 14:37:48 T_T 2017-06-16 14:38:10 its ok, its the journey that matters, not the destination 2017-06-16 14:38:13 I am still using latex nowdays but can't get it to work in multipple language. 2017-06-16 14:38:27 //:=| 2017-06-16 14:38:41 I mean I wrote my cv in latex but I separate en and zh into different files. 2017-06-16 14:38:42 what do you mean, multiple language? 2017-06-16 14:38:52 elegast: en, zh_CN 2017-06-16 14:38:55 fr + en 2017-06-16 14:38:57 For example 2017-06-16 14:39:29 So the quickfix is done by plugins, no wonder. 2017-06-16 14:39:49 quickfix is populated by plugins 2017-06-16 14:40:12 You mean having multiple encodings in the same source file? 2017-06-16 14:40:26 elegast: Yeah. 2017-06-16 14:40:33 Using utf-8 works well right? 2017-06-16 14:41:07 Well, most ide's for latex, especially the older ones, lack any form of unicode support 2017-06-16 14:41:15 thats the first problem 2017-06-16 14:41:20 Oh 2017-06-16 14:41:26 elegast: But vim doesn't right? 2017-06-16 14:41:34 Vim has exceptional unicode support 2017-06-16 14:41:49 The second problem though, its the tex distribution you use 2017-06-16 14:42:39 I am bad in knowledge of tex, it's just complicated and I don't even know how texlive, pdflatex, xelatex, luatex differs. 2017-06-16 14:42:56 So normally if latexmk fails, I try either pdflatex or xelatex. 2017-06-16 14:42:56 pdflatex and latex, for example, need packages to support a single type of encoding 2017-06-16 14:43:11 I dont think you can mix and match as those are set in the preamble 2017-06-16 14:43:13 elegast: So if utf-8 it works then right? 2017-06-16 14:43:22 Yes but *only* utf-8 2017-06-16 14:43:30 or *only* chinese 2017-06-16 14:43:32 elegast: What other encodings are useful? 2017-06-16 14:43:40 You can only use one at a time 2017-06-16 14:43:41 however, 2017-06-16 14:43:46 Ah 2017-06-16 14:44:29 I think xelatix is more capable, it has full opentype support, and I think it also has better encoding support 2017-06-16 14:45:23 xelatex is included in texlive, I think 2017-06-16 14:45:37 You could also try https://tectonic-typesetting.github.io/en-US/ 2017-06-16 14:46:58 oh that's nice 2017-06-16 14:47:03 Yeah 2017-06-16 14:47:30 yes, its basically a wrapper around texlive and xelatex, from the looks of it 2017-06-16 14:48:25 I guess its a modernized package manager for livetex 2017-06-16 14:48:42 consus: Do you normally test your build in another term or in vim as well? 2017-06-16 14:48:49 Looks quite nice. 2017-06-16 14:50:51 pickfire: you want xelatex, don't let the lack of pdf in its name fool you, it can produce better pdfs then pdflatex does 2017-06-16 14:51:03 Ah 2017-06-16 14:51:11 xelatex has actually decent font support 2017-06-16 14:51:12 also it doesn't need the intermediate step from text -> dvi -> pdf 2017-06-16 14:51:13 which is a big plus 2017-06-16 14:51:15 tex* 2017-06-16 14:51:18 elegast: I do use xelatex sometimes, if it fails then pdflatex. 2017-06-16 14:51:30 has xelatex has full opentype support and also good unicode 2017-06-16 14:51:33 Shiz: What does that means? 2017-06-16 14:51:40 yes xelatex* 2017-06-16 14:51:51 Opentype relates to the font features you can use 2017-06-16 14:51:58 unicode obviously the encoding 2017-06-16 14:53:35 That means you can just use *regular* fonts in your tex files 2017-06-16 14:53:44 and not have to jump through hoops with metafonts and such 2017-06-16 14:54:18 and you can use the fonts rasterizer and binary processors 2017-06-16 14:55:35 in laymens term think of it like a virtual machine inside the font, processing its own kind of bytecode 2017-06-16 14:55:43 its what makes it looks and render nicely 2017-06-16 15:05:03 Well, latex is known for it's font rendering 2017-06-16 15:05:17 Ah 2017-06-16 15:06:00 No, it excels at typesetting, not really the rendering, thats why metafonts where invented 2017-06-16 15:06:54 But the problem was you couldn't use system fonts, e.g. the fonts people use and see in daily life, xetex fixed that with opentype support 2017-06-16 15:11:07 The "renderer" part of latex, in modern days, is actually postscript 2017-06-16 15:12:03 and has nothing to do with tex 2017-06-16 15:12:30 and of course pdflatex and xelatex output to pdf by default 2017-06-16 15:13:07 yes, though, pdflatex turns it into dvi first, iirc 2017-06-16 15:13:35 its been awhile though, not exactly sure on that 2017-06-16 15:13:57 but yeh, pdf = postscript 2017-06-16 15:14:55 yeah pdflatex uses a whole bunch of intermediary files including a .dvi stage 2017-06-16 15:17:27 I see no reason to use system font. 2017-06-16 15:17:45 But latex's font is too good to be read on a paper. 2017-06-16 15:18:21 Well, the part about it being a system font is just a welcome surprise, its the fact that its a modern opentype font thats the big killer feature 2017-06-16 15:19:11 It means you no longer are limited to just the few fonts made specifically for tex 2017-06-16 15:20:18 Which may not matter to us much, ofcourse, we just want legible docs, but does to alot of publishing agencies 2017-06-16 15:23:37 Getting it working nice and good is my target, not the system fonts so I see no diffence in those. 2017-06-16 15:24:59 Well, its easier to install an opentype font on your system, then it is to take that font, converting it to a format usable for tex, and then packaging it up in a tex package 2017-06-16 15:25:40 besides the obvious fact that you loose a lot of features, that are "just there" in opentype, and not anymore in tex, unless you go through a lot of fiddling 2017-06-16 15:28:10 Ofcourse, you're right, if the default fonts in tex suffice, it adds little value, except for say, Litegatures and all those other cool opentype features we didn't know existed 2017-06-16 15:28:38 trust me, there's quite a lot 2017-06-16 15:32:51 the thing is, font foundries, don't make fonts for tex anymore (if they ever did) and though you might think theres alot of them around since theres a million fonts going around, theres only 3 or 4 in the world that actually produce quality fonts 2017-06-16 15:34:44 its added value that those can now be used in documents typeset with tex, especially to publishers and designers 2017-06-16 15:53:40 ACTION bus stop 2017-06-16 15:59:39 ACTION got it now 2017-06-16 15:59:42 elegast: Thanks 2017-06-16 16:12:27 I was trying to noop a command by linking it to /bin/true but busybox returns "applet not found" any ideas? 2017-06-16 16:13:09 mgoodwin: busybox uses argv[0] to determine which applet to run 2017-06-16 16:13:23 so if you ln -s something to a busybox symlink, it will try to use the name of whatever you're symlinking 2017-06-16 16:13:32 which likely doesn't exist as a busybox applet 2017-06-16 16:13:48 hmm 2017-06-16 16:13:57 ime the fastest noop is 2017-06-16 16:14:02 touch /some/bin && chmod +x /some/bin 2017-06-16 16:14:05 :p 2017-06-16 16:14:15 I get it, /bin/true is itself a link to bb 2017-06-16 16:14:18 yeap 2017-06-16 16:14:39 Alright, thanks 2017-06-16 16:24:41 install -m 0755 /dev/null /usr/bin/git :P 2017-06-16 16:25:31 also works, although i'd be wary of device files 2017-06-16 16:25:33 :p 2017-06-16 16:59:38 i wonder if it would be better to have a separate /bin/true and /bin/false 2017-06-16 17:07:53 /bin/maybe 2017-06-16 17:11:35 Heh. 2017-06-16 19:55:20 Hey, is there a way to automatically enable ssh for installs and the boot medium? I don't have any visual output on my target machine 2017-06-16 19:57:36 <_ikke_> ln -s /etc/runlevels/default/sshd /etc/init.d/sshd ? 2017-06-16 19:57:55 <_ikke_> not sure how to get that on the boot medium though 2017-06-16 20:00:08 hm, could I create the boot drive, mount it on another machine and chroot into it? _ikke_ 2017-06-16 20:00:58 <_ikke_> the default isos are iso9660, which is read-only 2017-06-16 20:03:36 bummer 2017-06-16 20:10:11 What might cause setup-disks to show "No disks found"? 2017-06-16 20:10:26 The devices aren't even in /dev. Strange. 2017-06-16 20:18:07 <_ikke_> what kind of disks do you have? 2017-06-16 20:20:09 It's just one SSD. It's a VPS. This is from a slightly old CD, 3.3.1. 2017-06-16 20:22:23 Tyr Have you looked at the Install Alpine on Rackspace article in the wiki? If I understand it correctly, it describes a similar problem to what you want to do (create install medium with ssh running): https://wiki.alpinelinux.org/wiki/Install_Alpine_on_Rackspace 2017-06-16 20:23:31 Hm. Rebooting fixed it. Weird. 2017-06-16 20:31:14 Bel-Chion: I did not, thanks I will check it out 2017-06-16 20:35:11 So, when booting, a large Linux distro first loads an initramfs image into RAM, mounts the real root and then `switch_root`s into it. Can it go the other way, too? Assuming I have enough RAM, can I `switch_root` a running system back into RAM? 2017-06-17 00:00:29 Is there an officially sanctioned way to install Alpine to disk from an existing Linux distro (i.e. without booting the live CD)? 2017-06-17 00:01:35 Ah, https://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting looks helpful. 2017-06-17 00:03:15 Wait, no, that still requires the live CD. 2017-06-17 00:33:24 Well, my horrible method is to un-cpio the initramfs, chroot into it and install from there. :| 2017-06-17 00:33:55 I'd just use the mini root filesystem release, but that doesn't appear to include the installation tools. 2017-06-17 00:41:09 I did migrate a live Gentoo box to Alpine once. Don't recommend the procedure. 2017-06-17 01:04:55 Hey guys, anyone running alpine on lxc, where networking is setup in the lxc config? How to properly deal with the networking service *inside* the host in that case? 2017-06-17 01:05:28 wtf 2017-06-17 01:05:30 why am i guest 2017-06-17 01:05:53 ACTION is stumped 2017-06-17 01:07:38 Maxpm: I think there's a howto on the wiki somewhere, regarding mini-root-fs 2017-06-17 01:08:03 not sure if the one you referenced is the same one 2017-06-17 01:08:26 I dont think so 2017-06-17 01:08:48 but yeh, like TBB says, I imagine its a flaky procedure 2017-06-17 01:11:26 anyway, I want my lxc network config defined on the host, the alpine guest shouldn't be allowed to mess things up. But then the networking service in the guest starts failing with ioctrl on an existing network interface, and permission issues, which is obvious 2017-06-17 01:11:46 but if i wipe /etc/network/interfaces, the service still fails 2017-06-17 01:12:06 should I install a dummy rc.d provider for net? 2017-06-17 01:12:52 Or is there something obvious I'm missing? 2017-06-17 01:14:57 The alternative would be to accept the fact that the guest can reconfigure networking at will 2017-06-17 01:15:02 I dont particularly like either solution. 2017-06-17 01:24:37 Hi, I used https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage as a guide and created an alpine iso with some apks cached. I was hoping to have a few installed by default and avoiding an apkovl to do it upon every startup, so I was attempting to use 'abuild-apk add --root "$APKROOT" --arch "$ARCH" ' within a section the custom profile script defines... that does install the package, but the installa 2017-06-17 01:28:02 the more i learn about lxc, the less impressed I end up ... 2017-06-17 01:28:34 its starting to feel like a toy compare to competing tech 2017-06-17 01:30:56 which sucks because its practically the only thing on linux that comes close to containerization 2017-06-17 01:31:48 well except docker 2017-06-17 01:32:36 maybe. 2017-06-17 01:32:55 I guess the problem is there's containers-for-convenience and containers-for-security, and most people only care about the former, which robs momentum from the latter. 2017-06-17 01:33:15 Yeah, pretty much 2017-06-17 01:33:29 I don't get why its not been done from the ground up though 2017-06-17 01:33:51 I mean, if you try and mimic stuff from other platforms, atleast do it right. 2017-06-17 01:36:16 My assumption about setting up networking from outside the guest was wrong though, if you configure networking there, the guest is still allowed to change it at will 2017-06-17 01:38:46 okeuday_bak, not that I have any experience with apkovl's, but you might receive more help if you finish your question :) 2017-06-17 01:43:46 Well, lxc suffices for an easy ad-hoc build environment, but that's about it, really. I don't really need it for anything else, or for it to be airtight, but the fact that it can't be done but is advertised as such just royally pisses me off 2017-06-17 01:44:37 elegast: what was the last there? I think it got cut off 2017-06-17 01:45:02 okeuday_bak: ".. package, but the installa" and then it cut out 2017-06-17 01:45:12 elegast: its basically like https://unix.stackexchange.com/questions/370895/install-packages-into-a-custom-alpine-linux-iso 2017-06-17 01:45:27 ah ok 2017-06-17 01:46:10 Hi, I used https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage as a guide and created an alpine iso with some apks cached. I was hoping to have a few installed by default and avoiding an apkovl to do it upon every startup, so I was attempting to use 'abuild-apk add --root "$APKROOT" --arch "$ARCH" ' within a section the custom profile script defines. 2017-06-17 01:46:12 that does install the package, but the installation is not in the iso that is created. Is some other command(s) required, and is this a bad approach for some reason (that requires everything use an apkovl upon startup)? 2017-06-17 01:46:50 You *dont* want to have to use apkovls right? 2017-06-17 01:46:55 elegast: (thanks for mentioning that, not sure why it gets cut off, the irc client doesn't complain) 2017-06-17 01:47:26 elegast: I don't mind using the apkovls, I just want to make sure some apks are installed by default to avoid the startup cost 2017-06-17 01:48:14 elegast: I know the hit is minimal, due to being cached, but it would be nice to make sure the image has everything ready before boot time 2017-06-17 01:48:18 Well, if its to be included in the install medium, thats not what apkovls are for, as far as my understanding goes 2017-06-17 01:48:46 Its basically just meant to store runtime changes across reboots 2017-06-17 01:49:01 elegast: well, I could add some config change to do "apk add" upon startup, I just wanted to move that into iso image creation 2017-06-17 01:51:24 Eh, if you're making the images yourself, you should just be able to add extra packages to be installed when you run it. Lemme go over it though, I haven't used those scripts yet 2017-06-17 01:52:50 obvious first question: you've exported the profile name for the build profile you made, assuming you followed that howto? 2017-06-17 01:53:16 before running mkimage 2017-06-17 01:54:28 idk if it fails if you missed that step, or just assumes a default profile 2017-06-17 01:54:41 elegast: yeah, I got a profile script working to make an iso, so it is fine that way, and I have it using an apkovl included in the iso, which is nifty, I just wanted to avoid relying on the apkovl for the "apk add" commands, but it seems like that approach is expected/typical 2017-06-17 01:55:17 elegast: the howto wiki has some errors, the EOF marker isn't used properly, and there is a line that needs to be deleted, let me look 2017-06-17 01:56:31 elegast: you have to delete the kernel_flavors line, not sure if that is meant to mislead beginners and lead to consulting revenue :-) , but it is a current problem in the wiki 2017-06-17 01:57:22 ghehe 2017-06-17 01:58:38 allright, lemme just make an iso and see what happens, from the looks of it, adding your packages to $apks should be it 2017-06-17 01:59:05 elegast: that only caches the apks, it still requires an "apk add" on the resulting iso, after it has booted 2017-06-17 02:01:13 ah ok if thats expected behaviour, then I imagine its just a matter of adding a setup_xxx script 2017-06-17 02:01:22 or adjusting the setup_alpine script 2017-06-17 02:03:01 elegast: well, I wanted to have it have the installation ready upon startup, and alpine uses the local service for custom stuff instead of rc.local, and I could add "apk add" there to solve the problem, but that seems like a hack that makes stuff slower 2017-06-17 02:05:08 ok, im not exactly sure, but I think what you want to look at is setup-alpine with the -a flag 2017-06-17 02:05:21 as setup-disk checks for overlay files 2017-06-17 02:05:25 during installation 2017-06-17 02:05:52 elegast: but that requires a script be ran on the custom iso while I am trying to make sure the custom iso is ready before login 2017-06-17 02:06:32 you're building a live image? 2017-06-17 02:06:44 sorry I thought you meant making an install medium 2017-06-17 02:06:45 elegast: yeah, trying to make a livecd with some apks 2017-06-17 02:08:28 well, I suppose one could make an rc service to run apk add 2017-06-17 02:08:51 elegast: yeah, but isn't that bad, or is that seen as a good thing from a security perspective perhaps? 2017-06-17 02:09:53 well, its installed from the cache, not the network, so the package can't be intercepted during install, but even then it would be a signed package 2017-06-17 02:10:29 that said though, ideally you'd just want the contents of the package file be apart of the rootfs 2017-06-17 02:12:09 elegast: that was my perspective, and that question in the link above, because that is pretty normal for a livecd, but perhaps alpine wants to keep the apkovl as a more mutable way of influencing images that are meant to be more immutable, so perhaps that is the way it is suppose to be 2017-06-17 02:13:52 elegast: I think I will go the apkovl route since it seems like the normal path with the least resistance 2017-06-17 02:14:41 true, building a custom image implies immutability though 2017-06-17 02:15:01 atleast that's how i'd see it 2017-06-17 02:15:18 yeah, it is different 2017-06-17 02:18:30 I'm pretty sure there's a hook somewhere to do what you want though, I haven't found it yet 2017-06-17 02:20:10 elegast: everything I have seen refers back to the apkovl file, having that retrieved from a remote or local source, the addition of alpine-base is hidden within update-kernel and I still haven't found where the apk world file is created with that entry, so some things are still unclear 2017-06-17 02:20:41 I guess the apk add does it automatically 2017-06-17 02:21:18 yeah, but it also implies including kernel modules and such, atleast in that wiki page, those need to be included in the ramdrive, or atleast the rootfs, somehow 2017-06-17 02:21:44 that wouldn't work at all with apkovls 2017-06-17 02:21:51 elegast: yeah, all the kernel stuff is coming from apks (it appears) so that makes sense 2017-06-17 02:23:33 elegast: I mean looking in initramfs-init was mainly about apkovl stuff (https://git.alpinelinux.org/cgit/mkinitfs/tree/initramfs-init.in#n454) 2017-06-17 02:25:18 oh ok. yeah. that would work. didn't expect it to be done like that 2017-06-17 02:25:53 elegast: yeah, it is sort of neat, but I have it working with it specified in the iso, so I will try to stick with that 2017-06-17 02:26:00 though i can't tell if the apks are extracted before or after the switch-root 2017-06-17 02:26:07 apkovls 2017-06-17 02:26:49 speed wise, I don't think it matters much 2017-06-17 02:27:32 extract ramdrive -> extract apokvl -> init 2017-06-17 02:27:54 elegast: yeah, it is quick when I try it, so it probably doesn't matter relying on the apkovl, thanks for the feedback though, just not use to alpine yet 2017-06-17 02:28:09 yeah, I'm still learning also 2017-06-17 02:28:15 your welcome though 2017-06-17 02:29:47 elegast: it seems like pkgs as a kernel option handles automatically adding, so I will try that first 2017-06-17 02:30:14 comma separate list 2017-06-17 02:30:21 yeah just pretend they're kernel modules :) 2017-06-17 02:31:28 I think they'll only exist during ramdrive init phase though 2017-06-17 02:31:52 after that the rootfs is mounted and switched in 2017-06-17 02:31:57 *I think* 2017-06-17 02:32:00 but heck, try it 2017-06-17 02:32:38 or idk maybe its taken that into account 2017-06-17 02:32:54 yeah, not sure, it looks neat, so I need to try it :-) 2017-06-17 02:34:41 oh I'm not sure but I think I recall reading about specifying apkovls as a kernel commandline option also 2017-06-17 02:36:23 elegast: yeah, I found that in there, it is automatically discovered on the iso somewhere, so I will stick with that, it helps to avoid making it more complex 2017-06-17 02:37:26 yerr, if you put apkovl in the APPEND line, it goes through /tmp/apkovl 2017-06-17 02:37:29 elegast: it is nice having the kernel option too though, it keeps things flexible 2017-06-17 02:38:31 lol can even specify an url 2017-06-17 02:38:36 thats cool 2017-06-17 02:39:06 yeah, that can help with managing diskless stuff 2017-06-17 02:39:25 though you'd have to setup networking at that stage also 2017-06-17 02:40:00 well, there's dhcp ofcourse 2017-06-17 02:40:40 yeah, it seems to handle the ifup, since it isn't relying on tftp 2017-06-17 02:41:39 that is weird, since the apkovl normally would have the config though 2017-06-17 02:43:37 elegast: yeah, the pkgs kernel option is the way to do it, odd though it may be 2017-06-17 02:43:42 well, linux kernel as well as modern bootloaders can setup networking 2017-06-17 02:43:57 mhm yeah, I just noticed that one too 2017-06-17 02:44:15 it runs apk add for whatever is on pkgs = ... 2017-06-17 02:44:41 string of packages seperated by comma 2017-06-17 02:45:53 I think you can specify the commandline in the mkimage profile, no? 2017-06-17 02:46:18 yeh kernel_cmdline 2017-06-17 02:46:22 elegast: yeah, did that and that worked, so all is well 2017-06-17 02:46:31 there ya go! 2017-06-17 02:46:33 :) 2017-06-17 02:46:56 :-) thanks for looking into it, should be set for now 2017-06-17 02:47:02 Ok this is some pretty interesting stuff 2017-06-17 02:47:23 Yeah you're welcome, thanks for having me check it out, learned some new insights 2017-06-17 09:36:37 I'm having some interesting issues with rsyslog, potentially in combination with logrotate: I log mail.* to a different file, but after a while they end up in messages anyway? and now it's even weird, postfix is being logged to maillog but dovecot to messages 2017-06-17 09:38:44 oh what the hell, busybox syslogd is also running 2017-06-17 09:40:17 the syslog service is not in any runlevel so I wonder what's starting it 2017-06-17 10:06:30 what's with the syslog-ng package? the init.d concats a bunch of stuff together, and those configs all have "do not edit!" in them 2017-06-17 10:56:05 how can I run a command as user within a script in local.d 2017-06-17 10:56:06 ? 2017-06-17 10:57:11 su -c 'command' user ? 2017-06-17 10:58:46 oh I was trying 2017-06-17 10:58:49 su user command 2017-06-17 10:59:54 so it will be something like this 2017-06-17 10:59:55 https://pastebin.com/BfaaUPQ5 2017-06-17 10:59:56 ? 2017-06-17 11:00:47 I guess; you can also make it su - -c '...' user if you want to "clear environment, go to home dir, run shell as login shell" 2017-06-17 11:03:28 now I just want to execute it as the user samba 2017-06-17 11:03:47 so the folder created ins't taken over by root 2017-06-17 11:03:52 and breaks my samba share 2017-06-17 11:04:21 guess you can just run it and try-and-see 2017-06-17 11:05:42 didn't worked 2017-06-17 11:05:48 now the fuse mount is not mounted at all 2017-06-17 11:05:54 it worked before without anything 2017-06-17 11:07:04 well, it seems unlikely samba will be able to log to /mnt and so on 2017-06-17 11:07:20 nologin: this account is not available 2017-06-17 11:07:20 it worked before 2017-06-17 11:07:25 and the reason I'm getting this is 2017-06-17 11:07:27 nologin: this account is not available 2017-06-17 11:07:35 I created the samba user with nologin 2017-06-17 11:07:41 well there you go 2017-06-17 11:07:56 is there a workaround? 2017-06-17 11:08:04 not setting it to nologin? 2017-06-17 11:08:12 i think that's not secure 2017-06-17 11:08:26 never had a issue with this setup in ubuntu before 2017-06-17 11:08:57 su -s /bin/ash -c foo user 2017-06-17 11:10:30 foo not found? 2017-06-17 11:10:41 replace foo with your command 2017-06-17 11:11:14 in my script? 2017-06-17 11:12:08 foo is where the reclone bit goes 2017-06-17 11:12:43 it works only if i use ' before and at the end of the command 2017-06-17 11:12:47 but thanks 2017-06-17 11:13:12 well yes, the command is an argument to the -c option 2017-06-17 11:13:36 i like how there is always a workaround with linux 2017-06-17 11:13:38 :p 2017-06-17 11:17:00 2017/06/17 11:14:55 mount helper error: fusermount: failed to open /etc/mtab: Permission denie 2017-06-17 11:17:06 i still get this error in my logs 2017-06-17 11:17:15 but the mount is created and working ok as of now 2017-06-17 11:18:58 /etc/mtab should be a symlink to /proc/mounts 2017-06-17 11:24:17 I'm not sure how to check that 2017-06-17 11:25:30 # ls -ld /etc/mtab 2017-06-17 11:25:32 lrwxrwxrwx 1 root root 12 May 25 18:57 /etc/mtab -> /proc/mounts 2017-06-17 11:31:28 yeah it looks like that 2017-06-17 11:31:31 whats now? 2017-06-17 11:32:22 dunno, maybe some hardened issue, but if it works it works 2017-06-17 11:37:12 ok 2017-06-17 11:37:13 thanks 2017-06-17 12:26:11 I'm evaluating some options for replacing ftp access with scp (or sftp) the thing is that ftp is chrooted, giving the customer a whole different layout then say when I'd use scp which confuses customers a great deal. Chrooting sftp implies root owned home directions, which is arguably worse. Has anyone used the trick of an unused non-standard root owned homedirectory shared by all sftp users, and using sshd's chroot-user-dir option to chroot them back 2017-06-17 12:26:13 to a *user* owned homedir in /home? 2017-06-17 12:26:49 using ChrootDirectory /home/%u 2017-06-17 12:28:49 but then the passwd file (or whatever password database mech is used) has all the users in say /my/root/owned/userdir, which has some obvious side-effects for users running applications on that machine 2017-06-17 12:31:55 I'm likely to inherit a small ISP in my chores one of these months, getting rid of ftp access without disrupting the eco-system is one of the things high on the list 2017-06-17 13:52:18 elegast: it's been a while but I've done that before, the catch to how it works is that you set the chroot directory, and then inside that, you create a home structure 2017-06-17 13:53:48 I had a whole separate directory tree just for sftp, something like /srv/sftp/home/%u 2017-06-17 13:55:15 unfortunately, that system is long gone, otherwise I'd share my config 2017-06-17 14:00:56 qman__: thanks, I'm not entirely sure yet how flexible I can be in changing the directory layout for this trick to work, with the existing infra, and still having minimal impact, but knowing that it *can* be done, is a good thing 2017-06-17 14:01:44 one trick I used to address the change in directory structure is bind mounts 2017-06-17 14:02:00 I had a shared folder I wanted all users to have access to, so I did a bind mount inside each of their homes to it 2017-06-17 14:02:09 it's kind of complex but it does work 2017-06-17 14:05:56 Mhmm. Yeah, I was leaning towards bind mounts also 2017-06-17 14:07:37 Though there's also a lot of bsd infra along side, I think nullfs mounts would do the same trick there 2017-06-17 15:26:18 Is there no apk for the kernel itself? 2017-06-17 15:28:10 Maxpm: there is, the package is called: linux-hardened or linux-vanilla depending on the flavor you want to use 2017-06-17 15:28:55 Ah. It didn't show up in https://pkgs.alpinelinux.org/packages?name=linux&branch=&repo=&arch=&maintainer= 2017-06-17 15:37:36 Maxpm: needs a wildcard 2017-06-17 16:03:11 There appears to be some kind of error in mkinitfs-3.1.0-r0.trigger: https://pastebin.com/rhNYrV9B 2017-06-17 16:03:26 It looks like it's piping the output of a failed command to cpio. 2017-06-17 16:11:56 not an issue in the trigger but in mkinitfs itself 2017-06-17 16:19:12 Mm. 2017-06-17 16:35:42 good old unix and its lack of pipefail 2017-06-17 16:52:43 say what you want about bash, but at least it's got that 2017-06-17 17:07:07 Speaking of shells, I'm surprised there's no (t)csh package in aports 2017-06-17 17:07:46 or did some joker put it in a package named "bsd-crap" 2017-06-17 17:09:11 would there be desire for such a package? 2017-06-17 17:09:48 the former, mind you 2017-06-17 17:10:46 <_ikke_> elegast: Even if it's usefull for you, alpine is pretty inclusive with pacakges 2017-06-17 17:10:51 <_ikke_> packages 2017-06-17 17:11:54 allright, I'll see how much effort it is to compile and package up, its been on my todo list for a while now 2017-06-17 17:42:56 great, the only sane bug report on the entire web is hidden behind a redhat paywall 2017-06-17 17:43:50 D: 2017-06-17 17:46:45 someone feel like sharing their credentials? 2017-06-17 17:47:32 I'm tall, handsome, intelligent and a monster at a pool table 2017-06-17 17:47:47 ... 2017-06-17 17:47:53 I'll get me coat 2017-06-17 17:48:47 Heh, those are fine creds 2017-06-17 17:49:18 dont share your redhat creds though, i'll cancel your subscription first chance ill get 2017-06-17 17:49:27 sorry I can't be of much help with that paywall 2017-06-17 17:49:54 bug report for what 2017-06-17 17:50:01 nah it's fine, there's some other bug reports about the subject 2017-06-17 17:50:07 tcsh out of memory error 2017-06-17 17:50:37 can you strace itt 2017-06-17 17:50:37 https://access.redhat.com/solutions/53622 2017-06-17 17:50:44 yep 2017-06-17 17:50:59 lemme just find a bin to paste in 2017-06-17 17:53:06 https://pastebin.com/t7jbsZg9 2017-06-17 17:53:51 just to distract you guys a little bit, it seems there's been a movement against csh and tcsh for quite some time now 2017-06-17 17:54:38 lemme guess, I should use bash? please 2017-06-17 17:55:01 ash :) 2017-06-17 17:55:54 I doubt that specific movement is advocating use of ash though 2017-06-17 17:56:12 But yes, ash trumps bash 2017-06-17 17:56:33 https://mx.gw.com/pipermail/tcsh-bugs/2015-July/000963.html 2017-06-17 17:56:43 well I'm addicted to some inputrc features, so alas stuck with bash 2017-06-17 17:57:27 https://mx.gw.com/pipermail/tcsh-bugs/2015-July/000967.html 2017-06-17 17:57:51 And there's a solution already, much abliged ncl :) 2017-06-17 17:58:17 no actually, it doesn't trump bash, at all 2017-06-17 17:58:58 its more secure and less bloated 2017-06-17 17:59:00 What does modloop do? 2017-06-17 17:59:02 it adheres to posix 2017-06-17 17:59:14 bash is none of those things 2017-06-17 17:59:17 but it does has lots of features 2017-06-17 17:59:57 I really don't feel like coding in features with ash that are already there in Bash just to make scripting a less of a pain in the butt 2017-06-17 17:59:59 so yer, on a day to day machine, I'd likely run bash over ash 2017-06-17 18:00:37 that said though. For scripting purposes, I *want* my scripts to be cross platform and bash-ism free 2017-06-17 18:01:21 but let me clarify my position slightly - I have no problem writing 100% sh when the scripts I do don't need the features of bash 2017-06-17 18:01:44 the only feature of bash i ever missed was pipefail 2017-06-17 18:01:51 and ${a/b/c} but ash does that too 2017-06-17 18:01:52 and seriously, if a platform can't handle bash then I would seriously consider not touching that platform :) 2017-06-17 18:02:14 Anything I can't do with plain sh, I do in Python 2017-06-17 18:02:26 or anything more then say 20 lines 2017-06-17 18:02:48 I agree on that last part though, TBB 2017-06-17 18:03:01 There's a lot of stuff that assumes Bash in the wild 2017-06-17 18:03:03 I wouldn't do that; OS level scripting is just so much easier and clearer in shell 2017-06-17 18:03:09 so yes. often you just need bash 2017-06-17 18:03:32 Yeah, there's some exceptions to the rule ofcourse 2017-06-17 18:03:58 and... well, pipefail and arrays make bash useful. then there's of course zsh but I never really got what its major selling point was 2017-06-17 18:04:17 lol, idk, its what all the cool kids use 2017-06-17 18:04:51 at the end of the day though, my shell is usually tcsh :P 2017-06-17 18:05:45 (and now of course as always when speaking of shells) now all I'd like bash to have is the kind of scoping that would enable "object" scripting and that other feature I always forget when talking about what I'd like to have in bash, and I'd probably never touch another shell again... 2017-06-17 18:06:57 idk, sounds like you're describing powershell :P 2017-06-17 18:07:21 nothing of the sort 2017-06-17 18:07:26 just scoping 2017-06-17 18:07:45 just doing local/global is ... meh 2017-06-17 18:12:23 I use fish because I'm a wimp. :) 2017-06-17 18:13:43 as far as interactive shells go xiki is probably the most insane, but I'm pretty sure outside of that it's useless 2017-06-17 18:16:44 lol that reminds me of the wish shell 2017-06-17 18:17:42 Oh, I forgot, thanks ncl, you pointed me to a working solution! 2017-06-17 19:22:20 I've noticed sometimes during the pkg phase of apkbuild assumes pkg/$pkgname but copies from sources where its src/$pkgname-$pkgver, how to deal with this inconsistency? 2017-06-17 19:23:28 should I just redeclare $pkgdir? 2017-06-17 19:23:36 or is there something I'm missing here 2017-06-17 19:25:58 no, that doesn't seem to affect anything 2017-06-17 19:27:34 oh its the Makefile of the package itself, that assumes this, nevermind 2017-06-17 19:56:37 :P 2017-06-17 20:15:07 does one of you got a (currently) cubietruck running alpine? 2017-06-17 20:15:14 i'm finally where I can reinstall them 2017-06-17 20:21:23 elegast: at a certain point the pkgs kernel option becomes a length problem, I switched to the ovl kernel option which automatically adds everything from the apkovl, which is better 2017-06-17 20:44:55 the apkovl kernel option I mean 2017-06-17 22:23:58 okeuday_bak: Glad to see you got things sorted! 2017-06-17 22:27:26 In the alpine release cycle, is the *current* state of packages in the repository tracked in aports git, the v3.6 branch does not currently reflect the state of the repositories. Can I assume that's the stable branch, ahead of the repositories, and the tags, e.g. v3.6.1 reflects the state of the repository as it *was* during release? 2017-06-17 22:29:55 Short of dropping the official repositories, and bulk building every package myself (or, atleast, every package included in my deployments), how would I get the state of aports reflecting the current state of release 2017-06-17 22:34:53 A moving reference pointing to the current commit that represents the state of official packages might be something worthwile to have 2017-06-17 22:36:46 3.6-stable does reflect the current state of the 3.6 repository 2017-06-17 22:37:04 https://git.alpinelinux.org/cgit/aports/?h=3.6-stable 2017-06-17 22:37:33 by repository i mean, the package mirror, not the git repository, if i wasn't clear 2017-06-17 22:37:33 master reflects edge, and the v3.x tags reflect the packages contained in the downloads for that respective version 2017-06-17 22:37:38 as do I 2017-06-17 22:37:41 oh ok 2017-06-17 22:37:46 builders directly build from 3.x-stable and push 2017-06-17 22:37:56 so unless a builder is broken, 3.x-stable always refers to the latest packages for 3.x 2017-06-17 22:38:30 main/uwsgi-* 3.x-stable is version 2.0.15, though the mirror reports 2.0.14 2017-06-17 22:38:53 Hm. It could ofcourse be that the mirror is behind a bit 2017-06-17 22:38:55 https://git.alpinelinux.org/cgit/aports/tree/main/uwsgi/APKBUILD?h=3.6-stable 2017-06-17 22:38:57 no? 2017-06-17 22:38:59 :P 2017-06-17 22:39:03 it's 2.0.14 in 3.6-stable 2017-06-17 22:39:10 o.O 2017-06-17 22:39:47 its 2.0.15 in the master branch, which is edge 2017-06-17 22:39:49 im currently on branch 3.6-stable 2017-06-17 22:40:12 git branch -a: * 3.6-stable 2017-06-17 22:40:56 APKBUILD pkgver reports 2.0.15 2017-06-17 22:41:38 all i can tell you is that my local 3.6-stable branch also says 2.0.14 2017-06-17 22:41:54 Yeah, no I believe you 2017-06-17 22:42:54 tried git reset --hard origin/3.6-stable ? 2017-06-17 22:42:58 I guess instead of checking out and branching 3.6 I checked out master instead 2017-06-17 22:43:35 probably leaving out the origin/, creating a new branch of the head 2017-06-17 22:44:08 i think # git checkout -b 3.6-stable would creat ethe branch from origin/ by default 2017-06-17 22:44:12 err 2017-06-17 22:44:15 just # git checkout 3.6-stable 2017-06-17 22:44:20 -b would create a new branch... 2017-06-17 22:44:33 Yeah, thought so too 2017-06-17 22:44:56 My bad :) 2017-06-17 22:45:48 I didn't even notice until I started building the dependencies as opposed to installing from repository 2017-06-17 22:47:23 Well all good now, thanks for the assist, Shiz 2017-06-17 22:47:43 np 2017-06-17 22:47:58 Eh. I'll start over tommorrow, time for bed 2017-06-17 22:48:01 C ya guys 2017-06-18 00:12:50 do any of the alpine images provide a 32 bit efi image 2017-06-18 01:37:12 can't find any alpine linux "script" command, how can you save output from abuild to a file? 2017-06-18 01:39:07 guess older ">&" is appropriate 2017-06-18 02:37:39 My time is showing up as 18:37, but it's actually 22:37. I remember choosing busybox ntpd during setup, did I do something wrong? 2017-06-18 02:38:26 Timezone? 2017-06-18 02:39:01 America/NY 2017-06-18 02:39:25 I mean, maybe your timezone isn't set correctly? 2017-06-18 02:40:01 I recall setting it, but how can I double-check anyway? 2017-06-18 02:41:01 cat /etc/timezone 2017-06-18 02:41:19 Er, wait. 2017-06-18 02:44:01 ls -l /etc/localtime 2017-06-18 02:44:13 Should be a symlink to something in /etc/zoneinfo. 2017-06-18 02:44:39 http://termbin.com/hems 2017-06-18 02:44:48 That's the output, looks like it's set to New York. 2017-06-18 02:45:24 ...huh, when I ran that command, it fixed itself 2017-06-18 02:45:43 Go figure. 2017-06-18 02:46:30 now let's see if it's still fixed after I reboot 2017-06-18 02:48:40 ...yup, it's fixed. 2017-06-18 02:48:44 wow, that's really weird. huh. 2017-06-18 03:37:06 Just attempted multi-boot install of alpine. I followed the dual boot guide @ https://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting 2017-06-18 03:37:21 Everthing seem to work fine. 2017-06-18 03:37:45 I did not follow the grub2 boot loader modification. 2017-06-18 03:38:12 as I already had grub configured from another linux os. 2017-06-18 03:38:30 I have efi partion on sda1 2017-06-18 03:38:52 I went back into my other linux and ran upgrade-grub 2017-06-18 03:39:16 It picked up alpine-linux 2017-06-18 03:39:40 so when i booted into alpine linux it was a blk screen. 2017-06-18 03:40:00 what have I done wrong. 2017-06-18 03:40:49 my last input was "setup-disk -m sys /mnt" 2017-06-18 03:40:55 then I rebooted 2017-06-18 03:41:24 was I meant to install xorg at this point or after the reboot 2017-06-18 07:16:23 Hi, has anyone run into problems running OpenSMTPD on Alpine? I'm getting a "tlsv1 alert decrypt error", and my search fu is apparently weak 2017-06-18 09:09:44 Hi guys, is there anyone around to assist, Thx in advance ! 2017-06-18 14:38:47 Hello all. Quick inquiry: I am setting up a mail server using cyrus-imapd and postfix as MDA/MTA. Cyrus is not listed in alpine packages. What is the standard way of dealing with this: Compile on my own, compile and share package, request an official build, or something else perhaps ? Thanks 2017-06-18 14:39:43 <_ikke_> user1: compile + share package 2017-06-18 14:41:13 Sounds fine. I'll check the wiki about that. How fast are package accepted ? I will build missing modules on a separate VM from the actual server, using the package manager tere would be superb 2017-06-18 14:41:23 there* 2017-06-18 14:41:49 <_ikke_> AlexIncogito: Depends a bit how much time the devs have, but it can be rather quick 2017-06-18 14:42:10 Super, thank you 2017-06-18 14:43:39 <_ikke_> In the mean time, it's not that hard to setup your own repo either 2017-06-18 14:47:49 I think that would work 2017-06-18 15:22:40 I have a question about package policies regarding versioning, but also about the way apk works; 2017-06-18 15:22:54 Lets say I have a piece of software, of which there are two major versions 2017-06-18 15:24:57 Both versions need to be available, and I also *must* ensure that if someone does an apk upgrade, the greater version does not automatically install, as dataformats are with some risky effort forward upgradable, but *never* downgradable 2017-06-18 15:25:18 Is this a valid usecase where the major version becomes a part of the package basename? 2017-06-18 15:27:50 That's what I would expect, elegast. 2017-06-18 15:29:47 Maxpm: Mhmm I think so too! but I'd like these packages, after we've verified them to work as they should, to eventually be commited upstream, I want them to be consistent with the alpine way of doing things, so I wanna be sure 2017-06-18 15:31:48 There's two approaches basically, make it the responsible of the system administrator by using version pinning to ensure a certain version, or to move this responsibility to the package maintainers, both would have merit 2017-06-18 15:33:34 Though Im not sure if you can direct apk to install say basename-pkgver1.0 while basename-pkgver2.0 is also in the repository, in the former case 2017-06-18 15:35:01 My initial tests suggest thats *not* exactly how apk works, I can only seem to pin locally installed versions to prevent higher versions from installing on top, but cannot perform a clean install from the lower index as it seems to not be available anymore in the index 2017-06-18 15:37:23 Which is fine, most package managing systems work that way 2017-06-18 15:38:49 Can the two versions of your package coexist? 2017-06-18 15:39:30 They can, though they might want to read the same config, which is not always guarenteed to work 2017-06-18 15:40:00 I suppose that would be a requirement, or atleast a strong incentive? 2017-06-18 15:41:55 I don't personally agree with the approach of versioned config files btw (or default configs being installed as active configurations for that matter), its a sysadmin responsiblity imho 2017-06-18 15:43:50 Also as a general rule, you want one over the other, not both, atleast not on the same system, but you want the choice of either 2017-06-18 15:47:28 Default configs being distributed in say /usr/share/baseversion1 and /usr/share/baseversion2 I wouldn't be opposed to. 2017-06-18 15:48:56 But /etc/baseversion1/* and /etc/baseversion2/* is a no no, /etc/*.conf is the preferred choice 2017-06-18 15:49:09 Though I'm aware thats a highly personal view :) 2017-06-18 16:46:18 How bad an idea is running apk upgrade in a cron job? 2017-06-18 16:46:30 Well, apk update && apk upgrade. 2017-06-18 16:52:23 Depends on the level of risk you're willing to take honestly. 2017-06-18 16:52:30 And your range of comfort. 2017-06-18 16:53:10 <_ikke_> If you follow a specific release, it should be pretty safe 2017-06-18 16:54:26 also you can do apk --update upgrade 2017-06-18 17:06:39 Maybe there's a way to only auto-upgrade stuff from main? 2017-06-18 17:15:45 <_ikke_> Maxpm: it should not matter. community might be less 'stable', but it's also less critical 2017-06-18 17:17:02 <_ikke_> And I think it's not even less stable, just shorter maintenance period 2017-06-18 17:18:13 Perhaps. 2017-06-18 18:27:10 firefox appears to be uninstallable on 3.6; it requires so:libvpx.so.4, but I can only find libvpx-1.5.0-r0 which only provides libvpx.so.3 2017-06-18 18:46:03 <_ikke_> It is updated on edge 2017-06-18 18:47:06 How would you properly request a package be ported? 2017-06-18 18:48:12 <_ikke_> fastest would be to provide the package 2017-06-18 18:48:17 <_ikke_> Reeckz: what package? 2017-06-18 18:48:35 ProxySQL - http://www.proxysql.com/ 2017-06-18 18:57:18 TML: I've been running firefox on 3.6 just fine 2017-06-18 18:57:57 *i think* 2017-06-18 18:57:59 lemme check, brb 2017-06-18 19:00:41 yeah, it's real slow on my 1st gen atom, but it works 2017-06-18 19:24:34 <_ikke_> Reeckz: I'm trying to build it 2017-06-18 19:40:07 Sweet, thanks @_ikke_ 2017-06-18 19:44:03 elegast: I'm sure it would run fine, it's just not installable for some reason. Which version of libvpx are you using? 2017-06-18 19:46:27 TML: libvpx-1.5.0 2017-06-18 19:46:48 odd 2017-06-18 19:46:55 though, 2017-06-18 19:47:03 firefox-esr-52.1.2 2017-06-18 19:47:07 oh 2017-06-18 19:47:18 I was trying to "apx add firefox" 2017-06-18 19:47:24 what is "firefox-esr"? 2017-06-18 19:47:50 they changed the name a while back, I think esr is the lts release 2017-06-18 19:47:59 can't remember exactly but my guess is extended support release 2017-06-18 19:48:08 ye ye thats it 2017-06-18 19:49:08 ok, that at least installed - thanks 2017-06-18 19:49:31 though if you run apk add firefox, it would just complain about no such package, not libvpx being the wrong library version 2017-06-18 19:49:32 I still have constant problems with Firefox on Alpine tho 2017-06-18 19:50:02 elegast: apk info firefox 2017-06-18 19:50:02 firefox-54.0-r0 description: 2017-06-18 19:50:04 Firefox web browser 2017-06-18 19:50:16 as I do with other gtk-using software as well; the desktop environment themes don't reach some of the gtk enabled applications 2017-06-18 19:50:35 odd. I don't have that package in my repo 2017-06-18 19:50:57 edge/testing 2017-06-18 19:51:02 maybe that's why 2017-06-18 19:51:18 and 64-bit only 2017-06-18 19:51:22 oh right, yeh I disapled edge/testing 2017-06-18 19:51:39 except for a few local packages 2017-06-18 19:51:46 ah, I thought I disabled it too, but left one of the uncommented 2017-06-18 19:51:49 TBB: thanks 2017-06-18 19:52:06 and yeh, it also works on 64bit 2017-06-18 19:53:02 though I did compile it from source, its built just like the one in the repositories 2017-06-18 19:57:49 hmpf. Do we have a bsdmake in ports somewhere? 2017-06-18 19:58:15 it could possibly be named nmake or pmake or something like that. 2017-06-18 19:58:39 ah, thanks, I haven't tried that yet 2017-06-18 19:59:15 no, just imake and cmake 2017-06-18 20:00:02 imake sounds like it would have some bsd heritage though 2017-06-18 20:00:35 on bsd, "man make" says its derived from pmake, that's why it could have that name 2017-06-18 20:00:37 oh, nop, its an x11 specific build system 2017-06-18 20:00:45 imake is for x11 I think 2017-06-18 20:01:00 mhm google agrees with you 2017-06-18 20:01:03 (and dont look too closely, that output scares kittens to death) 2017-06-18 20:01:09 the imake one, that is 2017-06-18 20:01:17 ya, I figured 2017-06-18 20:01:33 its like "perl2c" output, except for x stuff 2017-06-18 20:01:56 ew 2017-06-18 20:02:03 on both accounts 2017-06-18 20:02:45 I love X, but the innards, I don't wanna go near it 2017-06-18 20:03:46 guess I'll bring in bmake too 2017-06-18 20:04:12 I don't feel like converting makefiles 2017-06-18 20:05:41 the netbsd one, that is 2017-06-18 20:07:27 if netbsd is similar to openbsd, then much magic will be in all the included /usr/share/mk/bsd.*.mk, and not so much in the code 2017-06-18 20:07:46 they have lots of macros and stuff so the per-binary makefiles get really short and neat 2017-06-18 20:08:13 I dont know, I've used openbsd exactly once, and didn't like it 2017-06-18 20:08:46 But yes, there's a lot of bsd.*.mk 2017-06-18 20:10:43 I feel like I'll end up bootstrapping pkgsrc by the end of the day, or something .. 2017-06-18 21:47:26 What does modloop do? 2017-06-18 22:02:51 iirc it is a squashfs containing kernel modules which is mounted at /lib/modules sometime during the (early?) boot 2017-06-18 22:07:17 That's funky. 2017-06-18 22:07:45 Are normal kernel modules too boring for us? :) 2017-06-18 22:07:57 kernel modules are a security liability 2017-06-18 22:08:27 ? 2017-06-18 22:13:41 normal kernel modules are in /lib and not available if you mount the root a bit later 2017-06-18 22:13:44 etc 2017-06-18 22:14:15 or, more funky, if your root won't mount 2017-06-18 22:14:19 after all it's just linux 2017-06-18 22:14:25 can't expect stuff to work :) 2017-06-18 22:50:43 Maxpm: it's not like they're different kernel modules 2017-06-18 23:31:27 hey guys, I was going over wiki/Wishlist; what exactly is meant by "building distros on top of alpine"? Running the *build* process ontop of Alpine, or *running* the userspace ontop of Alpine core? 2017-06-18 23:40:12 both 2017-06-18 23:40:32 anything nontrivial in the latter implies the former anyway 2017-06-18 23:53:29 Interesting, thanks Shiz 2017-06-19 00:00:10 awk, my routers failing over still isn't seamless .. 2017-06-19 00:00:36 blegh 2017-06-19 00:07:50 hmpf, is it sane that cloning aports from github/aports would take much much longer then cloning of alpinelinux/aports? 2017-06-19 00:08:14 github puts more layers in between, so no surprise :P 2017-06-19 00:08:50 lol, its been going for 20 minutes or so, not even at 50 percent yet 2017-06-19 00:09:20 its probably this new gui client, too .. 2017-06-19 00:09:52 cloning off alpines cgit usually takes like a minute 2017-06-19 00:09:55 at most 2017-06-19 00:10:07 ok well, a few actually. but still 2017-06-19 00:11:14 I just tried gitkraken, its not scoring any points 2017-06-19 00:11:27 its already non-free 2017-06-19 00:11:33 I think its going in the bin 2017-06-19 00:12:56 Its sales pitch movie thingy even claimed "if you use it with github, its even faster!" 2017-06-19 00:13:01 the audacity. 2017-06-19 00:16:14 geez. the cloning progression bar's z-index is above the menu bar even ... the window is literally immovable, much like the progress meter its drawing 2017-06-19 00:16:33 I just love these apps that do window management themselves 2017-06-19 00:16:58 I'll stop ranting now. 2017-06-19 00:19:02 anyone around to assist, Pls ! 2017-06-19 00:19:08 thx in advance ! 2017-06-19 00:19:29 don't ask to ask, just ask :) 2017-06-19 00:20:18 :0 2017-06-19 00:20:21 :) 2017-06-19 00:20:34 no one bas been around for 18 hrs 2017-06-19 00:21:09 well i've not tried for about 8hrs 2017-06-19 00:21:34 I'm GMT +10 2017-06-19 00:21:56 ok.. 2017-06-19 00:22:16 Installed alipne on sda3 2017-06-19 00:22:26 dual boot 2017-06-19 00:22:58 I followed https://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting 2017-06-19 00:23:20 dear lord that is some old wiki page :( 2017-06-19 00:23:37 :) 2017-06-19 00:23:51 my last input was setup-disk -m sys /mnt 2017-06-19 00:24:19 I already grub installed with other arch 2017-06-19 00:24:32 elegast: using git protocol might be slower than https, not sure why 2017-06-19 00:24:34 grub picks up alpine 2017-06-19 00:25:00 when I boot alpine I receive blk screen 2017-06-19 00:25:06 nothing happens. 2017-06-19 00:25:33 was I meant to install xorg at the same time as the os 2017-06-19 00:25:38 or after reboot 2017-06-19 00:26:22 with file manager I'm able to see all the right files are in the partition of sda3 2017-06-19 00:26:23 if xorg isn't installed you should see a console prompt to log in 2017-06-19 00:26:36 Thats what I thought ! 2017-06-19 00:27:12 suggestions Pls ! 2017-06-19 00:27:52 Grub might be funny 2017-06-19 00:28:33 if it's not doing display right 2017-06-19 00:28:41 I'm certainly no techie, what do you suggest.. 2017-06-19 00:28:57 do you /etc/fstab set up? 2017-06-19 00:28:59 medling with grub kind of scares me 2017-06-19 00:29:11 does Grub show anything at all? 2017-06-19 00:29:35 i have efi-boot disk onsda1 2017-06-19 00:29:38 i have efi-boot disk on sda1 2017-06-19 00:29:53 fat32(boot.esp) 2017-06-19 00:29:57 i didn't ask that - do you ever see grub at all when you boot up? 2017-06-19 00:30:05 yes 2017-06-19 00:30:10 alpinelinux 2017-06-19 00:30:44 ok, sounds like it is not a problem with Grub 2017-06-19 00:30:52 yes 2017-06-19 00:31:12 the blk screen has subtle hue to it 2017-06-19 00:31:13 are you doing anything esoteric? Using LVM? dm-crypt full disk encryption? 2017-06-19 00:31:19 not completly blk 2017-06-19 00:31:30 no 2017-06-19 00:31:41 :) 2017-06-19 00:31:54 beyond my skill set 2017-06-19 00:33:24 Fuck.. More attacks in london 2017-06-19 00:42:18 let's not get offtopic, or ahead of ourselves 2017-06-19 00:44:02 duncan: what do you suggest i do to rectify my issues 2017-06-19 00:45:20 first decide if the kernel itself actually works - when you select 'alpine' in grub, what happens? do you see scrolling text where the kernel is loading? 2017-06-19 00:45:55 no 2017-06-19 00:46:34 nothing, no attempt to boot.. It's like having the wrong gpu driver and rebooting 2017-06-19 00:47:26 i'm on the same pc with arch on sda2 2017-06-19 00:47:43 I can access the partition 2017-06-19 00:49:20 Any other monitors plugged in? 2017-06-19 00:49:40 no, just one default 2017-06-19 00:50:03 Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz 2017-06-19 00:50:04 | RAM: 1033 MiB / 11894 MiB 2017-06-19 00:50:13 Hp Sff 8300 2017-06-19 00:50:34 intel onboard gpu 2017-06-19 00:51:51 Does anything happen if you do Ctrl+Alt+F1 or Ctrl+Alt+F2? 2017-06-19 00:52:27 no tried that could not even reboot ctrl alt delete 2017-06-19 00:52:45 had to hold power button for 5 secs 2017-06-19 00:56:27 Well, I'm out of ideas. :| 2017-06-19 00:57:25 ok thx 2017-06-19 01:08:25 Just to be sure, you did change the UUID from that howto to yours? 2017-06-19 01:08:37 oh, nvm 2017-06-19 01:09:52 mystified quit 2017-06-19 01:10:01 they should have stuck around 2017-06-19 01:24:30 Now we're left hanging! :( 2017-06-19 01:58:44 Eep, downloads are slow right now. Had to try 5 mirrors before finding one faster than 300kBps 2017-06-19 02:00:27 leaseweb is hovering around 700 2017-06-19 02:13:50 mystified probably didn't adjust the ramdrive image from -grsec to -hardened 2017-06-19 02:14:48 I thought grsec was hardened? 2017-06-19 02:15:27 it is, but the suffix changed with 3.6 2017-06-19 02:15:40 Oh. 2017-06-19 02:15:51 that howto he was following referenced the old name 2017-06-19 02:17:32 ugh, time for bed, its like 4am 2017-06-19 02:18:03 later lads 2017-06-19 05:58:53 is there a "branding" page on the website with the alpine logo designs? 2017-06-19 08:03:36 luxio: https://alpinelinux.org/alpinelinux-logo.svg ? 2017-06-19 14:54:26 why the 64 hardened kernel in 3.6 is so much bigger than the grsec in 3.5 ? 2017-06-19 14:54:31 nearly double the size 2017-06-19 14:57:09 Hi, is there an how to about preparing (packaging) a kernel modules not provided in base image? 2017-06-19 15:26:29 <_ikke_> untoreh: the kernel itself has become larger (modules etc) 2017-06-19 15:36:17 _ikke_: only the x86_64 got bigger ? 2017-06-19 15:40:05 clandmeter: Is that logo designed by a designer? 2017-06-19 18:13:32 pickfire, ncopa arranged it. 2017-06-19 18:15:12 I don't know if my messages were sent, I got disconnected : 2017-06-19 18:15:15 Hi all. I am building a new package for cyrus-imapd. It has been a smooth ride so far, all the dependencies were already available. That's great, thanks. Now, I want to copy the default configuration files to /etc/cyrus/. I have been looking at postfix's APKBUILD as an example, but I cannot, for the life of me, figure out where these files are actually handled/copied, aside from the obvious in-file sed 2017-06-19 18:15:16 fix in prepare() : https://git.alpinelinux.org/cgit/aports/tree/main/postfix/APKBUILD 2017-06-19 18:15:19 (and the optional pcre/*sql/... maps) 2017-06-19 18:47:51 Ah, unless it is in make install 2017-06-19 18:56:42 The spirit of IRC fraternity helping, I resolved my issue 2017-06-19 18:57:17 Regarding libressl, is it a) required b) recommended that packages use it instead of openssl ? 2017-06-19 18:58:10 I surmise openssl alpine package is the actual implementation, not some sort of proxy ? 2017-06-19 19:06:34 Ok, I believe I have a grasp of how this is done. The "alpine" way of doing it would be to use libressl-dev and patch the code (mostly headers) where appropriate. Correc t? 2017-06-19 19:07:18 (s/headers/preprocessor defs/) 2017-06-19 19:11:45 AlexIncogito: yes 2017-06-19 19:12:02 you should only use openssl if the package really really really needs it and it can't be easily patched 2017-06-19 19:12:05 which should be almost never if never 2017-06-19 19:12:14 would be nice if we could get openssl out entirely by 3.7 imo 2017-06-19 19:15:09 libs arg would be "-lssl -lcrypto" correct ? 2017-06-19 19:20:54 it shouldn't be needed.. 2017-06-19 19:21:02 unless you're building your own software 2017-06-19 19:21:24 no ? But how will it distinguish between openssl and libressl library filenames ? 2017-06-19 19:21:56 libressl-dev installs the libressl .so it will link against 2017-06-19 19:22:25 Of course, makes sense 2017-06-19 19:22:55 Alright, so I'll just look into patching the ifdefs 2017-06-19 19:23:33 should be all that's needed, if anything 2017-06-19 19:23:38 usually not even that is needed 2017-06-19 19:24:36 Well I grepped for OPENSSL_VERSION_NUMBER and it returned a bunch of them, won't that be an issue ? 2017-06-19 19:24:50 Depends on what it does of course 2017-06-19 19:24:56 no 2017-06-19 19:25:00 not necessarily 2017-06-19 19:30:05 Well it can't find some function ref now, so I suppose I'll have to bite the bullet 2017-06-19 19:37:02 Shiz, are you sure we still ship openssl? 2017-06-19 19:37:33 i thought openssl pulls in libressl 2017-06-19 19:38:23 yeshttps://git.alpinelinux.org/cgit/aports/tree/main/openssl/APKBUILD 2017-06-19 19:38:26 https://git.alpinelinux.org/cgit/aports/tree/main/openssl/APKBUILD * 2017-06-19 19:38:26 Well, on latest alpine, my package builds with openssl-dev but not libressl-dev. Something must be different 2017-06-19 19:42:50 Shiz, ah ok, i was mistaken. 2017-06-19 19:44:00 AlexIncogito: likely conditional checks on OPENSSL_VERSION_NUBMER to use openssl 1.1.0 apis 2017-06-19 19:44:11 fix is to add || defined(LIBRESSL_VERSION_NUMBER) or something like that 2017-06-19 19:44:18 since libressl doesnt have those apis either 2017-06-19 19:45:23 yes, you can grep aports for similar patches. 2017-06-19 19:57:27 Kind of a weird question. If I'm trying to run docker on Alpine with a container that's tripping a grsecurity protection (in this case rancher/server is getting denied when mmaping some memory pages), is there a way to run paxctl to whitelist the executable INSIDE the container? 2017-06-19 20:04:08 sure 2017-06-19 20:04:16 just... do so :P 2017-06-19 20:04:50 pick apart the layers in /var/lib/docker you mean and run paxctl on the java binary? 2017-06-19 20:05:17 never worked with grsec before so the concept is kind of new as to how it turns on/off protections 2017-06-19 20:09:01 well, that appears to have worked but I feel like it's really hairy to do it that way since I'm modifying a docker container post-build. I might have to derive the container to run paxctl -c, paxctl -m as a build step 2017-06-19 20:40:06 In case anyone is curious, this is how you fix up rancher/server to run on a grsec kernel: https://pastebin.com/1A2Ejmpd 2017-06-19 20:59:09 Is there a prebuilt iso that has the serial/console port enabled? 2017-06-19 21:00:05 Of Alpine I mean 2017-06-19 21:09:49 nate you can adjust the cmdline to enable serial 2017-06-19 21:10:02 He quit 2017-06-19 21:11:03 thx 2017-06-19 21:12:35 How would one include check() in a package ? 2017-06-19 21:13:15 its just a function called by abuild 2017-06-19 21:13:22 abuild check reads that this will soon be mandatory 2017-06-19 21:13:30 you can do whatever you want in it 2017-06-19 21:13:55 some makefiles provide check/tests 2017-06-19 21:14:14 So basically I just add check() { make check } ? 2017-06-19 21:14:17 if you have no way to check, options="!check" 2017-06-19 21:14:32 yes 2017-06-19 21:14:39 if makefile supports it of course. 2017-06-19 21:14:45 Indeed 2017-06-19 21:14:59 Almost too easy, thank you 2017-06-19 21:15:28 yw 2017-06-19 21:16:25 Might have to add testing dependencies though I suppose 2017-06-19 21:16:55 if the tests needs them add them to makedeps 2017-06-19 21:20:29 cunit is in edge/testing, abuild cannot find them, what's the fix for this ? 2017-06-19 21:26:03 you have edge/testing it in your repository file? 2017-06-19 21:26:29 Oh, is it as simple as that ? What about when I upload the package ? Will it be stuck in testing ? 2017-06-19 21:26:39 well edge/testing 2017-06-19 21:28:02 define upload a package 2017-06-19 21:28:18 Commit my work for your appoval 2017-06-19 21:28:52 ah ok 2017-06-19 21:28:55 yes 2017-06-19 21:29:11 new aports should to to edge/testing 2017-06-19 21:30:02 ones you have verified it works as expected we/you can move it. 2017-06-19 21:30:18 Ok, so it goes into edge/testing regardless of whether I am bound to a package there or not. Not familiar with the workflow,sorry 2017-06-19 21:30:40 For now, I'm testing using apk add, is that incorrect ? 2017-06-19 21:31:13 I thought that would made more sense, commiting untested work seemed like an heresy 2017-06-19 21:31:27 make 2017-06-19 21:31:55 testing by apk add? 2017-06-19 21:32:05 you mean testing if the package can be installed? 2017-06-19 21:32:21 Testing if it is corretly installed, runs correctly, etc. 2017-06-19 21:33:19 the procedure is you build it first local, verify it works, submit pr to testing, wait for it on the mirrors, install it from the mirrors, test it again, confirm its ok if you want it to be moved. 2017-06-19 21:34:30 Ok, good, that's what I had in mind 2017-06-19 21:34:43 of course we will review if you have done basis sane packaging, but we wont test if the app actually does what it suppose to do 2017-06-19 21:34:48 clandmeter: if you are trying to test with a custom iso, it can be hard to verify before the package gets merged into aports, right? 2017-06-19 21:35:15 why is that? 2017-06-19 21:35:33 you can add repo's when running from custom iso and install the pkg 2017-06-19 21:35:35 clandmeter: since you would need to go through the process of making your own index to be used during iso creation 2017-06-19 21:36:06 ok like that. 2017-06-19 21:36:16 you want to include your new pkg into your iso? 2017-06-19 21:36:23 clandmeter: yeah 2017-06-19 21:37:07 abuild will make the index for you (it stores pkgs in ~/packages...) 2017-06-19 21:37:55 clandmeter: k, I may need to look into the iso creation script to see if it can accept a local index file, or something, to use that 2017-06-19 21:38:14 i think it can, but i didnt use the new scripts that much 2017-06-19 21:38:40 i think somebody (fcolista) recently added a howto how to handle the new iso scripts. 2017-06-19 21:39:17 clandmeter: does it take awhile for new aports to get merged in, if they are only going into testing for now? 2017-06-19 21:39:40 you have to make sure you include your apk pub key in the iso or it wont install. 2017-06-19 21:40:08 that depends on the aport 2017-06-19 21:40:10 clandmeter: yeah, I got the iso creation process working, it is nice, once it is ready/setup 2017-06-19 21:40:21 and if reviewers have time to review. 2017-06-19 21:40:45 clandmeter: k, if I don't hear back in X days, should I bug them? 2017-06-19 21:40:55 simple aports mostly dont take much time to get approval. 2017-06-19 21:41:17 clandmeter: so it is mainly about the number of dependencies and potential complexity? 2017-06-19 21:41:51 deps are not a factor here, more about complexity and possible risk or other breakage. 2017-06-19 21:42:04 clandmeter: k, makes sense 2017-06-19 21:43:18 if you touch libs or system stuff, then we tend to be more careful. if its just a new aport in testing, then its probably only about complexity and quality. 2017-06-19 21:44:05 clandmeter: k 2017-06-19 23:00:31 _ikke_: Any luck with ProxySQL? 2017-06-20 02:26:08 in 3.6.2 (or whenver i missed it) we end up with the linux hardened stuff now it seems? 2017-06-20 02:26:30 that is the going stale grsec branch i suppose? 2017-06-20 02:27:14 i just wanna be sure i'm aware whenever it's only the "new" upstream thing and the 30% they don't take over vanish 2017-06-20 02:27:20 or if that happened 2017-06-20 02:28:10 because, from what i've seen over the last weeks it seems there's quite a gap between what grsec shrugged off and will now start hitting 2017-06-20 02:28:45 i've grown quite used to the fact that 9/10 (or more) linux issues just don't affect an alpine box 2017-06-20 02:28:52 could't scratch it 2017-06-20 02:30:10 :D 2017-06-20 02:30:47 and what i honestly find irritating if apk search grsec returns things like linux-hardened 2017-06-20 02:30:54 and apk info has no reference to grsec 2017-06-20 02:31:19 this does not seem, uh, respectful 2017-06-20 02:32:48 i believe it shows up because it provides the linux-grsec package 2017-06-20 02:32:51 as an upgrade path 2017-06-20 02:33:08 ah ok that'd explain it 2017-06-20 02:33:17 thank you 2017-06-20 02:33:28 https://git.alpinelinux.org/cgit/aports/tree/main/linux-hardened/APKBUILD?h=3.6-stable#n143 2017-06-20 02:33:30 yep that's it 2017-06-20 02:33:31 i should slee and not try to think 2017-06-20 03:28:48 are dynamic library dependencies for a package included in the package? they don't appear to get mentioned explicitly in the depends list 2017-06-20 03:29:09 they're usually automatically detected 2017-06-20 03:29:45 scv: I guess I am asking should I put library dependencies into the depends section if they get dynamically linked 2017-06-20 03:30:07 dynamic stuff is detected as well 2017-06-20 03:30:16 unless it's loaded with dlopen during runtime or stuff like that 2017-06-20 03:31:07 scv: nah, nothing weird, I saw the output at the end of abuild, which didn't complain and spit a bunch of linkage info, it just seemed strange since some of the -dev packages were not installed, since the library was coming through a separate package 2017-06-20 03:31:46 scv: will depend on the auto-detect for now, that seems simpler 2017-06-20 04:28:25 <_ikke_> Reeckz: Some compilation issues 2017-06-20 04:29:00 Resolveable? 2017-06-20 04:30:02 <_ikke_> Reeckz: Probably, but I'm not certain how at this moment 2017-06-20 04:30:13 Alright, thank you so much for trying (: 2017-06-20 04:30:19 <_ikke_> mysql_connection.cpp:1572:30: error: 'mysql_close_no_command' was not declared in this scope 2017-06-20 05:58:04 clandmeter: ncopa pay someone to design it? 2017-06-20 08:17:35 Hi guys! I'm having trouble building varnish on alpine, does anyone have any tips? I get this error when the build gets to sphinx: Segmentation fault (core dumped) ../../bin/varnishd/varnishd -x cli > include/cli.rst 2017-06-20 08:23:07 kirreen, does dmesg say something? 2017-06-20 08:23:27 pickfire, not that i know. 2017-06-20 08:24:02 clandmeter: yes: [ 5423.586934] varnishd[955]: segfault at 559780c00d08 ip 00007f2df9728863 sp 00007ffc821c3b60 error 4 in libjemalloc.so.2[7f2df9725000+22d000] 2017-06-20 08:24:22 and another row: varnishd[2035]: segfault at 564ef3c001d8 ip 00007f4313ad2863 sp 00007ffd173996a0 error 4 in libjemalloc.so.2[7f4313acf000+22d000] 2017-06-20 08:26:12 you need jemalloc? 2017-06-20 08:27:12 clandmeter: Actually not sure, but I just installed gdb and it seems to build fine - wait a min 2017-06-20 08:27:31 looks like per default we disable jemalloc 2017-06-20 08:31:06 clandmeter: alright, gdb didn't fix it, but it shouldn't have anyway... So I need to enable jemalloc for alpine? 2017-06-20 08:31:47 it seems to crash in libjemalloc, so you could disable it and try. 2017-06-20 08:31:59 i dalready install jemalloc-dev and jemalloc before building, does that change anything? 2017-06-20 08:32:06 we have varnish in aports, not sure why you are compiling your own? 2017-06-20 08:32:47 uninstall jemalloc and add --without-jemalloc 2017-06-20 08:33:19 or you could try to debug the issue and fix it. 2017-06-20 08:33:33 option #1 is probably easier :) 2017-06-20 08:33:35 clandmeter: because the current version installed with apk is 4.1.3, I want a newer and immutable version 2017-06-20 08:34:07 I'll try uninstalling it and making with --without-jemalloc! thanks for the help by the way :) 2017-06-20 08:35:32 clandmeter: wait, where do I add --without-jemalloc? 2017-06-20 08:35:46 oh, in the configure 2017-06-20 08:36:57 which version are you using? 2017-06-20 08:39:50 clandmeter: of varnish? 5.1.2 2017-06-20 08:40:27 clandmeter: ok, might've built fine without jemalloc, trying to do a full container setup 2017-06-20 08:43:25 kirreen, how did you fix the missing backtrace? 2017-06-20 08:43:42 clandmeter: yezzzz, got further this time, thanks a lot! now for compiling the modules 2017-06-20 08:44:17 clandmeter: I think it was a missing dependancy 2017-06-20 08:44:30 clandmeter: Yeah, I installed something with apk, let me se what 2017-06-20 08:44:39 you used libexecinfo? 2017-06-20 08:45:13 clandmeter: Yup, that's probably the one - I install it anyway 2017-06-20 08:45:37 clandmeter: I've had a break from this project for about a month so I don't remember the exact process for solving all of my problems 2017-06-20 08:47:11 we patch out the backtrace crap because musl doesnt support it. 2017-06-20 08:56:03 clandmeter: my full docker image generation works now that i removed jemalloc, is it fine to run varnish without it? 2017-06-20 08:56:24 it will use musl own malloc 2017-06-20 08:56:37 its probably fine 2017-06-20 08:56:40 clandmeter: alright, great! Thanks for the help :) 2017-06-20 08:57:17 i think musl malloc is already good, i dont know if there are comparisons available. 2017-06-20 09:01:29 kirreen, looking at alpine varnish, i wonder if your custom build will work properly. we seem to adjust stacksize or else it will probably crash on musl. 2017-06-20 09:02:24 clandmeter: oh, okay. I'll make a test run soon, so we'll see what happens 2017-06-20 09:02:59 https://git.alpinelinux.org/cgit/aports/tree/main/varnish/fix-stack-overflow.patch 2017-06-20 09:03:54 could be you already run into the issue with using jemalloc. 2017-06-20 09:05:16 I have that patch downloaded, where and when should i run it? in the varnish directory prior to make? 2017-06-20 09:05:34 it wont work, its against 4x 2017-06-20 09:05:41 I run patch -p0 < /tmp/musl-mode_t.patch there 2017-06-20 09:06:27 clandmeter: oh, right, I think I realized before which is why it's not used. But you think I need to do those changes to the new varnish ver? 2017-06-20 09:08:20 im not sure, it doesnt look like the code is fixed. and it doesnt seem ncopa submitted the fix. 2017-06-20 09:08:46 so if you know how to apply it properly, give it a try. 2017-06-20 09:18:01 clandmeter: Ok, i'll give it a try later 2017-06-20 09:36:49 yeah, we got a few errors running it, I won't look into it until after lunch, but here are the errors if you're interested: https://hastebin.com/ihofahovuf.pl 2017-06-20 09:41:35 kirreen, edge now has 5.1.2 2017-06-20 09:45:54 can't find it on https://pkgs.alpinelinux.org/packages?name=varnish&branch=edge&repo=&arch=&maintainer= ? 2017-06-20 09:46:48 patience :) 2017-06-20 10:40:28 Hi everyone. 2017-06-20 10:43:12 I am still working on making cyrus imapd package. I have enabled cunit tests, and some of them fails. I am trying to tackle them one by one. Starting with regexp, after lookup in musl docs, it seem the ABI for regexp does not work on 64 bit arch. So I suppose that mean I would have to compile in 32 bit regardless of whether this is a x86_64 or x86_32 arch. How would I go about that ? I couldn't find gcc 2017-06-20 10:43:14 lib32 in the repository 2017-06-20 10:44:43 (Or perhaps is there another solution altogether ?) 2017-06-20 11:07:25 we do not support multilib 2017-06-20 11:16:43 how can a software kit not work on 64 bit in 2017? 2017-06-20 11:16:47 ACTION <- amazed 2017-06-20 11:26:08 IcePic: There is certainly a way to make it work, I would rather likely put that on my not knowing about it 2017-06-20 11:27:01 There are many errors, I guess I'll look into switching to dovecot. I have been using cyrus for only a few months 2017-06-20 11:28:38 well, many good things have been said about dovecot, so its probably not a bad move either way, but given that 64bit unix machines have been available since early 90s and cheap x86_64 since 10-15 years by now, it would be very strange for someone to code something up and not being able to test on a 64bit host and document it as just-not-working-on-64. 2017-06-20 11:29:09 the github page comments suggest they compile for x86_64 2017-06-20 11:29:22 https://github.com/cyrusimap/cyrus-imapd 2017-06-20 11:29:33 Yes cyrus does work on 64, I have it working on opensuse 64 2017-06-20 11:29:47 I'm unable to make the tests working with musl libc 2017-06-20 11:30:23 Which may or may not be related to the ABI issue I read about in musl doc 2017-06-20 11:32:11 is this what you're referring to? 2017-06-20 11:32:13 "glibc's regex uses a 32-bit regoff_t even on 64-bit archs. This is non-conforming to POSIX and precludes giving correct output for strings larger than 2GB. musl uses a correct type, but this renders the ABI of the regex functions incompatible on 64-bit archs. The ABI is compatible on 32-bit." 2017-06-20 11:32:26 Exactly 2017-06-20 11:32:55 as I read it, glibc's implementation is non-standard, and therefore not compatible with musl's - it isn't broken, just not compatible 2017-06-20 11:34:21 To be fair, this is one of 27 other unit tests failing, this seem to involve a lot of patching, which may involve several days work.. That seem too high a price 2017-06-20 11:35:24 I do like cyrus, especially it coming with CalDAV and CardDAV, but it seem much more reasonable to just switch to dovecot which is also a good piece of software 2017-06-20 11:36:02 yeah, larger projects that have been based on glibc forever can be quite difficult to port to musl 2017-06-20 11:38:50 An alternative would be to host this single service on anoter distribution, say CentOS. The plan was already to run network isolated services. But that does defeat the point of having a low attack surface, and security improvements blessed with alpine 2017-06-20 11:39:35 Plus, I'd rather only manage one distribution, and alpine seem great 2017-06-20 11:39:56 not necessarily, but yes, you'd have to spend some time stripping and locking down the other distro 2017-06-20 11:41:00 True 2017-06-20 11:42:23 I love using alpine, so I try to use it where I can, but there are some software packages I need that simply don't run on it 2017-06-20 12:01:07 AlexIncogito, if the tests give you too much trouble you could skip a few and go on. if you submit a pr just mention what you did. 2017-06-20 13:54:17 Ah, okay. 2017-06-20 17:10:43 @clandmeter: My issue lies on the side that if these many tests (and some crucial ones) do not run, this will certainly impact the software itself in production 2017-06-20 17:12:55 https://github.com/OpenRC/openrc/blob/master/src/rc/openrc-init.c 2017-06-20 17:12:58 openrc has an init now 2017-06-20 17:36:40 huh cool! 2017-06-20 17:59:14 did alpine have KDE packages at some point? I swear I remember seeing KDE packages..... 2017-06-20 17:59:36 I remember seeing kde packages in aports at some point 2017-06-20 18:01:28 actually.... what desktop environments does alpine have? I got xfce running, but ehh 2017-06-20 18:02:05 it would be very cool if there was an official alpine desktop image 2017-06-20 18:03:10 how hard is it to make a custom alpine ISO? 2017-06-20 18:03:33 There are three or four desktops mentioned on the wiki. 2017-06-20 18:04:14 there's mate at least 2017-06-20 18:04:33 didn't really impress me, it's broken like xfce too, only in a different way 2017-06-20 18:04:46 XFCE, Mate, Awesome, Gnome, EyeOS (is that really a desktop?) 2017-06-20 18:05:34 what is a desktop environment besides a lot of bloat anyway? 2017-06-20 18:05:37 Bel-Chion: yeah, I saw that page 2017-06-20 18:05:51 i'm running alpine on my desktop, and am happy 2017-06-20 18:06:12 I'm not entirely sure if you're trolling or wanting to be trolled :) 2017-06-20 18:06:26 all you really need is a window manager lol 2017-06-20 18:06:28 neither. 2017-06-20 18:06:39 agreed. wm is all you need. 2017-06-20 18:07:13 firefox, openoffice, and all the other apps all work, independent of the "desktop environment" be it kde/gnome, or anything else 2017-06-20 18:07:35 cause they use the wm inside of those big des 2017-06-20 18:07:48 fair enough I guess, if you want to stay in the 1990's :D 2017-06-20 18:08:16 youd be surprised there are some really cool wms that look alot better then des 2017-06-20 18:08:20 come to think of it, I think I remember seeing enlightenment in the packages before, too 2017-06-20 18:08:25 but its a matter of opinion ofc 2017-06-20 18:08:39 but I don't see it now. 2017-06-20 18:09:30 maybe it was removed due to lack of interest by the pkg maintainer 2017-06-20 18:09:39 TBB: what exactly do you mean with 90ies? besides the bloat that got added? 2017-06-20 18:09:42 arch3y: yeah, that's my guess 2017-06-20 18:10:53 Well, better a package removed than an age-old unmaintained package shipped. 2017-06-20 18:11:06 hmm, maybe alpine needs another repo called "desktop" 2017-06-20 18:11:23 that would be better suited as a group 2017-06-20 18:11:28 not a repo imho 2017-06-20 18:12:00 oh and another note: I <3 the wifi support in setup-alpine 2017-06-20 18:12:13 So what would be a good window manager, efficient for usage and possible "good looking" ? 2017-06-20 18:12:21 i3wm 2017-06-20 18:12:45 worked OOTB on the old laptop I installed it on (atheros wifi) 2017-06-20 18:12:47 seconded 2017-06-20 18:12:49 bloat is a term used by people who really got stuck in the terminal window, that's all. there's a wide range of people with varying needs with regards to their computing 2017-06-20 18:13:15 thats your opinion but all you really need is a web browser and terminal 2017-06-20 18:13:23 i3 seemed nice 2017-06-20 18:13:25 and that's your opinion 2017-06-20 18:13:36 not to mention that "bloat" is much less of a concern if you have a fast enough computer ;) 2017-06-20 18:13:56 I used to have a launcher app called keypirhina on windows, but I couldn't find something equivalent... 2017-06-20 18:14:03 this debate could go on forever we all have our own prefs and ways of doing things 2017-06-20 18:14:12 however, some additional meta packages that install more desktop extras for those who want it ... is not such a bad idea 2017-06-20 18:14:23 Not to mention that it strongly depends on the work you do with your computer. 2017-06-20 18:14:28 $work bought me a maxed-out dell xps 13 :D 2017-06-20 18:14:32 maybe they need to be pkged 2017-06-20 18:14:40 after that it's a matter of whether you -want/need- to install them 2017-06-20 18:14:59 arch3y: my point exactly; we all have our own preferences 2017-06-20 18:15:14 and we were just explaining that a DE is not needed 2017-06-20 18:15:20 With Alpine being rather small, it would probably better to concentrate on one desktop and doing that well instead of installing half a dozend, that all work somewhat? 2017-06-20 18:15:46 Bel-Chion: as far as project resources are concerned, yes 2017-06-20 18:15:52 ppl are always gonna have a different opinion but it will take time to grow teh community to get things built 2017-06-20 18:15:55 Bel-Chion: and XFCE is pretty much just that right now 2017-06-20 18:16:07 cause yes resources of the project is a prioirty always 2017-06-20 18:16:25 I kinda wonder if it wouldn't be better to have a 3rd party build a desktop distro on top of alpine 2017-06-20 18:16:30 or it could be based on their own pref they pkg what they like to use for an env and leave the rest to the community 2017-06-20 18:17:12 I don't know about that, shodan45. Working as a part of a larger community would probably help such efforts 2017-06-20 18:18:05 TBB: true. And this magical "3rd party" doesn't exist, anyway. 2017-06-20 18:19:34 arch3y: I believe that's how things are now for Alpine - people have packaged what they see as useful. There's always room for more package contributors. 2017-06-20 18:20:13 hah. the topic needs to be updated - 3.6.2 was released 3 days ago 2017-06-20 18:20:14 for sure however most ppl will take and take from an open source distro and not always give back 2017-06-20 18:20:30 and that puts more stress on the devs to keep up with demand and so on 2017-06-20 18:20:42 but more ppl pkging is always a good thing 2017-06-20 18:20:51 and maintaining 2017-06-20 18:21:56 Also, if you want a distribution that allows you to install everything under the sun, you can always use Arch. 2017-06-20 18:22:37 some of that is due to the AUR which is not supported by the Arch lol 2017-06-20 18:22:58 and those can be disasterously bad ideally you learn to pkg things yourself and help support tehm 2017-06-20 18:23:02 cause its not that difficult 2017-06-20 18:25:08 Arch, or Gentoo, or even *shiver* Ubuntu 2017-06-20 18:25:29 I was basically a Gentoo user for 16 years 2017-06-20 18:26:09 then I woke up to the fact I needed my spare time for things other than emerge 2017-06-20 18:26:32 I have a stupid question. How do I add a user to different groups? Is it just like any other linux OS? 2017-06-20 18:26:33 No idea, I never used Gentoo. Only OpenSuSe for some time, and then installed Alpine in a virtual machine for testing how to set up a server a little while ago. Mostly, I use Windows.. 2017-06-20 18:26:45 m4chm4n: not quite if you're going with the default install 2017-06-20 18:26:55 m4chm4n: addgroup 2017-06-20 18:27:42 sweet!! Thank you. 2017-06-20 18:27:45 I did things in RedHat professionally for some years so for me coming up with my own packages for Alpine when needed was not that much of a problem 2017-06-20 18:29:45 This is the first time I have attempted installing a linux distro this technical. 2017-06-20 18:36:01 I have one more question. How do I get my usb drives and sd card drive to mount when usb/sd gets plugged in? 2017-06-20 18:38:13 are you using xfce? 2017-06-20 18:39:02 apk add gvfs udisks2 thunar-volman 2017-06-20 18:39:14 if I remember the pkg names correctly 2017-06-20 18:39:35 yes and thanks once again for the help 2017-06-20 18:39:44 np 2017-06-20 18:41:22 I might be interested (if I can only find the time for it) in participating in some sort of a desktop effort, although I'm probably not the right guy to do it; I mean, I haven't even in the space of one year managed to solve that XFCE + Gtk theme problem that still bothers me to no end 2017-06-20 18:42:08 but on the other hand I've been banging my head against Alpine for 1.5 years now professionally so I understand some parts of it 2017-06-20 18:43:08 those apps did the trick TBB 2017-06-20 18:43:29 excellent, so my dementia is not that far developed yet :) glad to hear that 2017-06-20 18:44:01 lol 2017-06-20 18:48:44 Yea I am slowly learning the ins and outs of Alpine...... 2017-06-20 18:56:07 I just wish I was good enough to port over some bluetooth apps. the ones in the repos are shit. 2017-06-20 19:21:26 I'm actually one of the (rare?) opensuse users 2017-06-20 19:21:35 when it comes to desktop/laptop stuff 2017-06-20 19:22:07 their build service is usually excellent 2017-06-20 19:22:39 well, the stuff available from the build service. I haven't used OBS myself. 2017-06-20 19:28:02 I have, back in the day when it was in its infancy and a pain in the ass to use 2017-06-20 19:49:26 obs is pretty cool once you figure out your way around its ui 2017-06-20 19:57:56 Hi guys, how make TZ works properly with alpine? 2017-06-20 19:58:59 take a look https://dpaste.de/0ag4 2017-06-20 20:13:41 there's a setup-timezone script in the "regular" alpine version, not sure how it should be handled in containers 2017-06-20 20:14:21 I'll check this, just install tzdata didn't work.. 2017-06-20 20:15:55 did you try ln -sf /usr/share/Europe/Amsterdam /etc/localtime ? (for europe/amsterdam) 2017-06-20 20:16:27 tzdata package contains the stuff in /usr/share, /etc/localtime needs to be a symlink to the right one 2017-06-20 20:17:45 I wanted to use variables for this 2017-06-20 20:19:46 the one with TZ variable seems to work for me if you have the tzdata package installed 2017-06-20 20:19:50 docker run -ti -e TZ=Asia/Tokyo alpine:3.6 /bin/sh -c "apk add -U tzdata && date" 2017-06-20 20:20:25 just the stock alpine image doesn't contain the tzdata 2017-06-20 20:22:05 latest = 3.6 , right? 2017-06-20 20:23:04 should be 2017-06-20 21:19:07 maybe the problem is in my PHP configuration then 2017-06-20 21:25:00 Why does my chroot report an error using nslookup? I created an alpine chroot with these commands: http://paste.debian.net/hidden/4db6547d/ 2017-06-20 21:27:44 no reverse lookup for 127.0.0.1 i think 2017-06-20 21:27:59 do you have a localhost entry in /etc/hosts proper 2017-06-20 21:28:14 or does the nameserver lack one for itself 2017-06-20 21:28:37 might also be because you got no domain/search entry but i think it's just localhost 2017-06-20 21:28:49 well, one of them should be it, not sure which ;-) 2017-06-20 21:29:29 I tried domain/search in resolv.conf as well, didn't help. hosts has a line for 127.0.0.1 localhost localhost.localdomain. 2017-06-20 21:29:47 I get the same error with the nameserver set to 8.8.8.8, which I hope would be configured correctly. 2017-06-20 21:29:48 the order is wrong for one 2017-06-20 21:29:54 That's how it came. 2017-06-20 21:29:59 127.0.0.1 localhost.localdomain localhost 2017-06-20 21:30:02 well 2017-06-20 21:30:06 it's not working right 2017-06-20 21:30:11 so why assume anything is right 2017-06-20 21:30:12 :) 2017-06-20 21:30:32 I flipped it, no change. 2017-06-20 21:30:35 ::1 in it? 2017-06-20 21:31:07 no other ideas unfortunately, strace next hope 2017-06-20 21:33:53 Seems like a bug. 2017-06-20 21:34:21 https://hastebin.com/dupazegidu.css 2017-06-20 21:34:23 Here's the strace if you're curious. http://paste.debian.net/hidden/7a2399d0/ I'll open a report. 2017-06-20 21:34:31 it's something in your dns i'm afraid, just no idea what 2017-06-20 21:34:42 Oh? curious. 2017-06-20 21:35:50 from the output it's really the lookup of the dns server that fails 2017-06-20 21:35:51 It prints that line before it even opens a socket to try network lookup. 2017-06-20 21:35:54 but idk why it is null 2017-06-20 21:38:47 line 7 is just bizarrely early. It hasn't even opened resolv.conf yet. I verified the gpg signature for the tarball, so it's "correct". 2017-06-20 21:41:27 <- sleep 2017-06-20 21:41:30 gl ;) 2017-06-20 21:41:43 good night. I'll open a bug report. 2017-06-20 21:44:32 nevermind, someone already has reported it as #6902 2017-06-20 21:49:42 uh weird problem 2017-06-20 21:50:44 I'm installing using the default image. I am trying to create an ext4 filesystem, but "mkfs.ext4" is not found. I thougt this package was provided by mke2fs, but I can't find it. I think I'm being really silly? 2017-06-20 21:50:56 :( 2017-06-20 21:52:41 it's in e2fsprogs as a symlink to mke2fs 2017-06-20 21:56:08 ah yes, how silly of me 2017-06-20 22:04:27 ahh, Linux 2017-06-20 22:04:50 tried to find a WLAN adapter for my access point, and did 2017-06-20 22:05:04 of course, not supported by the kernel directly 2017-06-20 22:05:23 there's a driver from the manufacturer that's out of date and several updated versions by the community 2017-06-20 22:05:52 ... none of which work on Alpine, which suggests the source is pretty effing terrible 2017-06-20 22:06:28 they compile allright, but loading the kernel modules either do nothing or even jam the kernel :D 2017-06-20 22:09:27 TBB: Intel, Atheros work OK 2017-06-20 22:10:57 yup. but you rarely know for sure what chipset your wlan adapter uses, especially wrt USB ones 2017-06-20 22:14:23 I've only had good experiences with Intel and Atheros really 2017-06-20 22:49:13 Actually that stands for most intel hardware: igpu, netw, chipsets, ... If only there were no ME 2017-06-20 22:59:30 ok I have a proper question now. I've installed the kernel dependencies (I did apk add kernel-hardened-dev, also isntalled GCC, ncurses, ncurses-dev), but when I do "make menuconfig, I get "scripts/basic/fixdep.c:105:23: fatal error: sys/types.h: no such file or directory" - it's obviously checking for the build dependencies, but I'm not sure what I'm missing 2017-06-20 23:34:31 bison or something like it? 2017-06-20 23:50:47 Good evening good people. I'm running into some issues trying to compile the OSSEC agent. It's throwing this error at me: " error: unknown type name 'u_int8_t' " as well as " error: unknown type name 'u_int16_t' " 2017-06-21 01:22:29 fixed 2017-06-21 01:23:26 now I've got problems compiling on GCC6. GCC 5 works, but GCC 6 goes bezerk when dealing with the size overflow protector plugin 2017-06-21 01:24:16 Same behavior on hardened gentoo (with musl). I wonder what causes it. It could be my config file. 2017-06-21 01:55:09 duncan^: need linux-headers 2017-06-21 01:55:13 fwiw 2017-06-21 01:56:00 I solved that, I've got all the prerequisites now, the problem i just described is present with linux-headers. 2017-06-21 03:45:21 Managed to navigate my way around the issue I had. 2017-06-21 09:08:30 Do I really need root to mount stuff? 2017-06-21 09:08:53 <_ikke_> pmount can mount removable drives as a normal user 2017-06-21 09:08:55 lizzzy: can't you use udisks or something? 2017-06-21 09:09:11 <_ikke_> oh, alpine has no pmount 2017-06-21 09:10:58 So, I can't mount drives like on other flavors of *nix? 2017-06-21 09:11:23 <_ikke_> on other flavors you also need to be root to mount drives.. 2017-06-21 09:11:52 <_ikke_> unless the option user is specified in fstab 2017-06-21 09:12:12 the ownership of the dir-to-be-mounted-over might also affect your ability, no? 2017-06-21 09:40:02 Eh? 2017-06-21 09:52:42 I'll look into it later. 2017-06-21 09:58:03 https://bugs.alpinelinux.org/issues/7372 Can somebody have a look at this? Java is quite broken in the 3.6 release, Jenkins unable to start ( https://github.com/jenkinsci/docker/issues/508 ), our webapp image generation completely broke (silently), etc... Can we raise the priority of the ticket? 2017-06-21 11:11:22 Trying to go through the wiki to find how the package lifecycle works but I can't find anything 2017-06-21 11:11:42 Anyone know where that info resides? 2017-06-21 11:14:19 <_ikke_> What kind of lifecycle? 2017-06-21 11:18:57 Documentation on how packagers upgrade their packages and then how that upgrade moves between releases? I'm a fedora packager so I'm accustomed to that lifecycle and I'm new to alpine. Perhaps there isn't one? Is it just that edge is snapshotted and that's how new versions of packages are introduced, or is it up to the maintainer's discretion? 2017-06-21 11:19:46 <_ikke_> mgoodwin: Yeah, edge is following the master branch in the aports repository, and releases are tagged from that 2017-06-21 11:20:05 <_ikke_> each release should only get minor updates 2017-06-21 11:22:22 Take this for example: https://pkgs.alpinelinux.org/packages?name=monit (since it happens to be relevant to me) Is there a policy that governs how versions are introduced to the latest release? (latest branch from edge) 2017-06-21 11:22:52 That is, I'm not very clear on whether that package will eventually be upgraded to 5.23 to match edge, or whether that will happen with 3.7? 2017-06-21 11:23:52 <_ikke_> mgoodwin: You can upgrade master/edge at any time 2017-06-21 11:23:57 <_ikke_> or the maintainer can 2017-06-21 11:24:34 Well he has, but 3.6 is still old 2017-06-21 11:24:44 Hence why I wanted to understand how the alpine lifecycle works 2017-06-21 11:25:08 <_ikke_> mgoodwin: Because resources are limited, there is no one making sure all packages are updated to the latest version 2017-06-21 11:26:21 <_ikke_> mgoodwin: You can flag the package out-of-date to notify the maintainer 2017-06-21 11:26:48 <_ikke_> mgoodwin: but 3.6 would not get a new major version, that will only happen in 3.7 2017-06-21 11:28:19 And I would imagine what determines a major version is at the discretion of the maintainer 2017-06-21 11:29:15 In the end what I'm looking for exactly is "When can I expect monit 5.23 to reach alpine:latest (docker) which is currently at 3.6" 2017-06-21 11:29:51 And I believe the answer is "When 3.7 arrives" 2017-06-21 11:30:03 <_ikke_> Right 2017-06-21 11:30:15 but I wanted a deeper appreciation of the package lifecycle too 2017-06-21 11:30:21 Thanks 2017-06-21 11:34:53 mgoodwin: typically major upgrades are not done within a stable branch 2017-06-21 11:35:05 mostly upgrades in a stable branch are minor updates 2017-06-21 11:35:13 this can of course depend on the urgency of said update 2017-06-21 11:35:40 Is what determines a major version clearly defined? 2017-06-21 12:06:07 https://hastebin.com/raw/ecekafuqun wow Natanael really has is hands full! 2017-06-21 12:34:23 :P 2017-06-21 12:34:33 there is no strict definition no 2017-06-21 12:39:30 <_ikke_> mgoodwin: It even depends on the project itself what consists of a minor or major update 2017-06-21 13:06:26 Is there a (easy) way to run a VM on linux like a "transparent" hypervisor? As in you start up the machine and it boots the VM which owns the console and when the VM shuts down the host shuts down. It can be as noisy as necessary on the console before and after VM as long as its fully automatic. I'm currently thinking basic alpine install + bridged networking + qemu in x11 w/ no WM. Pass as much hardware 2017-06-21 13:06:32 through to the VM as is reasonable/possible with vt-d/iommu. Write some custom ACPI daemon that is aware of the situation and pass events to qemu over monitor. Automatic shutdown when qemu exits. 2017-06-21 13:07:15 As far as why, my father has a machine with failing hardware, I gave him a win10 box, but he hates it and unplugged it to plug the old one back in. 2017-06-21 13:12:31 you could do that with a custom session with virtualbox 2017-06-21 13:14:33 I don't follow, googling for virtualbox "custom session" isn't illuminating. 2017-06-21 13:16:57 Do you mean a custom Xsession after like lightdm? I was thinking boot up runs a custom xinit script that just runs qemu w/ no host-side authentication. 2017-06-21 13:18:35 yep, a custom xinit script is easy enough 2017-06-21 13:22:26 tw: take a look at Xen Dom0/DomU 2017-06-21 13:22:34 for a concept 2017-06-21 13:22:49 Basically what Qubes uses 2017-06-21 13:27:01 I'll look into it, thanks. It seems like DomU is a PV container and not a HVM? I'll have to do some research, I'm not as familiar with xen as qemu. 2017-06-21 13:39:20 tw: domU is usually a PV container, but there is a new type called PVHVM, which is, um, different: https://wiki.xen.org/wiki/DomU_Support_for_Xen 2017-06-21 13:39:42 it basically does the IO for PV but puts the rest on top of HVM drivers 2017-06-21 14:35:18 That sounds a lot like virtio. I think I've got enough information that I need to make an afternoon project to concrete my understanding before moving forward. 2017-06-21 14:56:46 tw: a domU is an unprivileged VM - dom0 is privileged 2017-06-21 14:57:04 the difference is if it is allowed to make any hw calls that do not go through the hypervisor 2017-06-21 14:57:07 emulated or not 2017-06-21 14:57:37 PV means it uses hypercalls to directly get stuff (i.e. memory accesses or idk what really) from the hypervisor 2017-06-21 14:57:55 HVM means it uses cpu hardware virtualization to do most of this stuff 2017-06-21 14:58:19 so it can't work unless cpu helps - early days the PV stuff was substantially faster than cpu hacks but that is long past 2017-06-21 14:58:43 still, there was inefficienncies, i.e. it is stupid to emulate a system timer for every vm 2017-06-21 14:58:50 PV didn't need to do that it just asked 2017-06-21 14:59:04 and PVHVM is the attempt to merge these concepts 2017-06-21 14:59:37 so make use of a "smart" hypervisor but launch as an emulated/hw-assisted/non-virtualization-aware system 2017-06-21 14:59:50 newish intel hw now also provides a hw emulated timer in the vLAPIC 2017-06-21 15:00:08 that way the system init ends up like normal hardware, no shitty code to make it work 2017-06-21 15:00:20 and then it can use whatever the hypervisor can do better later on 2017-06-21 15:00:26 royger: at last, that's good 2017-06-21 15:00:36 the timer thing is the killer difference for scalability 2017-06-21 15:00:56 so if you don't just care about "how fast is a VM" but "how fast are 60 vms that are all busy" 2017-06-21 15:01:12 royger: nice to see you, i heard you moved from barcelona 2017-06-21 15:01:26 was gonna visit at the last oneconf and then you were MIA :) 2017-06-21 15:03:13 oh yes, I moved closer to my home town, and now I'm in Cambridge for a while :) 2017-06-21 15:03:51 probably until the winter 2017-06-21 15:11:25 Hi guys, is alpine-rpi limited to RPi's 2 and 3, or also the older 1 B+ models? 2017-06-21 15:15:43 The wiki mentions version Alpine 3.2 or greater for the RPi2 but it doesn't hint wether this is from the point of view where it's otherwise compatible with the RPi1 or RPi3 2017-06-21 15:16:11 it should work with any of them 2017-06-21 15:16:30 Allright, thanks kaniini! 2017-06-21 15:20:37 btw, does anyone have any info to what extent the recently discovered stack-clash exploit impacts alpine linux? 2017-06-21 15:21:54 Other distro's apply fixes to glibc (which Alpine deployments usually do not have) but there's also patched applied to the kernel on those distro's. I don't think the exploit is limited to glibc per say 2017-06-21 15:22:26 https://grsecurity.net/an_ancient_kernel_hole_is_not_closed.php 2017-06-21 15:23:39 That should be an interesting read 2017-06-21 15:24:15 elegast: it's part of the kernel, they all implement similar stack 2017-06-21 15:25:09 Seems like it, it's also not limited to Linux kernels as certain BSD's are also impacted 2017-06-21 15:25:36 It can only really be fixed by compiling userspace with stack probing. 2017-06-21 15:27:25 I don't think it affects -hardened ekrnels at the very least 2017-06-21 15:27:52 I think if you place a huge gap between the heap and stack and have ulimits in place (which would only work on 64bit arches where you can have large holes), then you could be safe also 2017-06-21 15:28:27 IcePic, safer or safe? 2017-06-21 15:29:33 in terms of security, I think the only word usable would be "safer". You close one hole, or randomize one parameter and you become safer, until a new type of exploit is discovered, then you are unsafe again 2017-06-21 15:29:37 huge gap between heap and stack is standard way to layout process memory 2017-06-21 15:29:54 True, but I mean with regards to this specific exploit in its currently known form 2017-06-21 15:30:07 xmux: yes, but making sure they never meet and that neither can grow more than so it reaches into the other 2017-06-21 15:30:57 You just need an application which can make a stack allocation larger than gapsize 2017-06-21 15:31:26 xmux: sure, but you can ulimit that so the size is reasonable, but still less than the gap. I guess. 2017-06-21 15:32:45 if you ulimit the stack to less than the gap, you can still hop the gap 2017-06-21 15:32:59 stackprobing is the solution. but that only solves large stack allocations hitting the guard page. but from the other end the heap can also jump the guard page, so allocations also need some kind of probing i assume? 2017-06-21 15:33:58 How could that happen? 2017-06-21 15:34:17 you malloc something that jumps the guard page? 2017-06-21 15:36:30 A correctly implemented malloc won't allocate memory outside of the memory provisioned to it by the kernel (ie by brk or mmap), and the kernel won't provision memory that overlaps with other memory 2017-06-21 15:36:53 yeah, that part seems easy 2017-06-21 15:36:55 stack clashing happens because the kernel doesn't get the opportunity to intervene 2017-06-21 15:37:14 the stack pointer is just moved into memory that isn't part of the stack 2017-06-21 15:37:40 normally there wouldn't be any memory mapped there, and then the kernel would fault in more pages for the stack 2017-06-21 15:37:41 oh. right, thx! 2017-06-21 15:38:11 but wouldn't it be hard to make an existing program to go "hey, for this function, I think I need a petabyte" ? Not impossible if you get to create the program yourself, but getting some other existing or even running program to do it? 2017-06-21 15:38:22 assuming a petabyte gap, for instance. 2017-06-21 15:39:28 variable length arrays and alloca() intrinsic are two ways of creating allocations on the stack with sizes that depend at runtime on ...something 2017-06-21 15:40:08 I just read on a concept called defensive programming, as I was going through about a related topic called design by contract 2017-06-21 15:40:49 If you use either VLA or alloca in your program, it's because the size of the array will be different every time, maybe for a reason the attacker can control 2017-06-21 15:41:20 and a friend pointed out an old document which might be from as far back as 1968, a panel in a NATO conference with regards to safe programming 2017-06-21 15:42:55 to me, "safe programming" sounds so generic it doesn't convey much information that helps a programmer with a deadline tomorrow. 2017-06-21 15:43:16 or defensive programming 2017-06-21 15:43:45 it is generic because I wasn't willing to go for a full length article about it here on IRC :) 2017-06-21 15:46:34 they're pretty wide concepts ranging from design to testing, from project management to choosing tools, etc etc 2017-06-21 15:51:32 its almost as if not-doing-it has very negative words describing what coders do when not following a design, when not checking return values, when not testing. 2017-06-21 15:52:08 which is why it can be hard to convey a meaningful improvement by saying to your team of coders, "we shall now start defensive programming!". =) 2017-06-21 15:52:56 hi, I've installed Alpine edge in lxc container in a Ubuntu 16.04 server VPS, and I'm trying to make OpenVPN run on it. Already have /dev/net/tun from host available. 2017-06-21 15:53:24 Problem is that OpenVPN tries to modprobe tun 2017-06-21 15:53:29 Starting openvpn ...modprobe: can't change directory to '4.4.0-81-generic': No such file or directory 2017-06-21 15:53:47 TUN/TAP support is not available in this kernel 2017-06-21 15:53:59 Any ideas on how to solve this? 2017-06-21 15:54:49 modprobe it on the host? 2017-06-21 15:55:43 modprobe tun iirc 2017-06-21 15:56:44 Bun: if that's the solution, then there's some problem with ubuntu 2017-06-21 15:56:53 as # modprobe tun 2017-06-21 15:56:58 does nothing 2017-06-21 15:57:13 # lsmod | grep tun 2017-06-21 15:57:19 returns nothing 2017-06-21 15:57:47 <_ikke_> lsmod | grep tun 2017-06-21 15:57:49 <_ikke_> tun 28672 2 2017-06-21 15:58:00 <_ikke_> libertas: Did you upgrade your kernel? 2017-06-21 15:58:05 <_ikke_> by chance 2017-06-21 15:59:18 _ikke_: # uname -r 2017-06-21 15:59:20 4.4.0-81-generic 2017-06-21 15:59:22 or does the init script always try to modprobe and then fail? that would be silly 2017-06-21 16:01:16 <_ikke_> libertas: ls /lib/modules/4.4.0-81 2017-06-21 16:01:32 <_ikke_> libertas: or even just /lib/modules/4.4.0 2017-06-21 16:02:17 _ikke_: no such file, need to install kernel headers? 2017-06-21 16:04:10 <_ikke_> no 2017-06-21 16:04:14 <_ikke_> ls /modules 2017-06-21 16:04:19 <_ikke_> /lib/modules 2017-06-21 16:04:27 https://serverfault.com/questions/429461/no-tun-device-in-lxc-guest-for-openvpn 2017-06-21 16:05:04 the problem is not the modules if you have /dev/net/tun on the host 2017-06-21 16:05:32 # ls /lib/modules 2017-06-21 16:05:33 4.4.0-79-generic 4.4.0-81-generic 2017-06-21 16:05:45 <_ikke_> right, that looks good 2017-06-21 16:06:28 I've got /dev/net/tun mounted on Alpine lxc with 2017-06-21 16:06:31 # grep bind /etc/fstab 2017-06-21 16:06:31 bindfs#/dev/net /lxc_containers/vpn/rootfs/dev/net fuse force-user=100000 0 0 2017-06-21 16:07:42 <_ikke_> the init script only does modprobe if the tunnel device does not exist 2017-06-21 16:07:46 correction: when lxc starts dev/net/tun not visible 2017-06-21 16:08:04 _ikke_: correct, see my last line 2017-06-21 16:08:29 <_ikke_> So it's more like a matter of order? 2017-06-21 16:10:10 don't know. Please see Stuart Cardall answer in https://serverfault.com/questions/429461/no-tun-device-in-lxc-guest-for-openvpn, also posted by Bun 2017-06-21 16:10:15 gives errors too 2017-06-21 16:29:25 lxc.mount.entry = /dev/net dev/net none bind,create=dir seems to work 2017-06-21 16:40:28 $ lxc-start -n vpn -F 2017-06-21 16:40:31 lxc-start: utils.c: safe_mount: 1751 Invalid argument - Failed to mount /dev/net onto /usr/lib/x86_64-linux-gnu/lxc/dev/net 2017-06-21 16:41:11 Bun: it appends dev/net to /usr/lib/x86_64-linux-gnu/lxc instead of /lxc_containers/vpn/rootfs as it should 2017-06-21 16:42:25 Bun: do you have ${LXC_ROOTFS_MOUNT} set? 2017-06-21 16:56:23 occasionally I end up in a state where packages refuse to upgrade, usually ffmpeg: https://my.mixtape.moe/fvnvbg.png 2017-06-21 16:56:48 it's kinda scary because I always discover it by accident, meaning I might have missed out on security updates for a while 2017-06-21 16:57:20 any way to detect this when it happens, and what's the correct way to fix it? 2017-06-21 17:15:54 libertas: no 2017-06-21 17:16:13 http://sprunge.us/GeUW this is the full config 2017-06-21 17:18:27 granted this is an ubuntu container, but I don't think that should matter 2017-06-21 17:21:44 trfl: out of curiosity, does apk --update upgrade -i -a downgrade/change anything? 2017-06-21 17:29:24 Bun, nah that's a no-op 2017-06-21 17:32:42 I did use ffmpeg from testing at some point because of a bug in stable, it was probably still @test when I upgraded from 3.5 to 3.6 (just in case that's somehow relevant) 2017-06-21 17:33:21 on a clean install objdump -T /usr/lib/libavformat.so.57 | grep transfer_int works 2017-06-21 17:36:22 that made me realize I have both libavformat.so.56 and 57, however transfer_int is not referenced in any of them 2017-06-21 17:36:36 /usr/lib/libavformat.so.57 -> libavformat.so.57.56.101 2017-06-21 17:36:58 libavformat.so.56.40.101 and libavformat.so.57.41.100 in my case 2017-06-21 17:36:59 seems like something is b0rked then 2017-06-21 17:38:27 I wonder if apk audit --system might be useful in this case 2017-06-21 17:38:59 probably not very 2017-06-21 17:39:13 while that's running, I noticed P:ffmpeg2.8-libs in /lib/apk/db/installed 2017-06-21 17:39:20 along with the regular ffmpeg-libs 2017-06-21 17:39:41 and nope, nothing of interest in the audit (thanks for the suggestion tho) 2017-06-21 17:40:08 so if you get rid of that one and reinstall ffmpeg-libs it might work? 2017-06-21 17:41:12 yep! turns out vlc was holding it back 2017-06-21 17:41:16 thanks a lot :> 2017-06-21 17:41:33 that's what you get for usign vlc ;) 2017-06-21 17:43:27 hm, noticed it removed 2.2.4-r2 which was a 3.5 leftover, where it used to be in the main repo 2017-06-21 17:43:36 it moved to community in 3.6, maybe that's what left me in limbo? 2017-06-21 17:44:56 apk doesn't like deleting or downgrade packages it seems 2017-06-21 17:45:20 (or telling you about them) 2017-06-21 17:45:27 that last part is sort of a downer :p 2017-06-21 18:17:48 Hello, I would like to known on what license the alpine logo is registered 2017-06-21 18:24:29 <_ikke_> h3x4: Not sure if it's released under an explicit license 2017-06-21 18:25:16 <_ikke_> h3x4: This is the repo, and there is no license included. https://git.alpinelinux.org/cgit/alpine-mksite/tree/ 2017-06-21 18:25:42 I just want to known if the logo of alpine can be used with aqemu (a GUI management for qemu) 2017-06-21 18:26:57 to be more specific it's regarding this -> https://github.com/tobimensch/aqemu/pull/37 2017-06-21 18:27:58 <_ikke_> right. I guess you have to wait for ncopa 2017-06-21 18:28:16 Ok I will thanks for the info 2017-06-21 18:28:26 <_ikke_> or clandmeter 2017-06-21 18:28:32 <_ikke_> I guess he maintains the site 2017-06-21 18:30:26 <_ikke_> https://git.alpinelinux.org/cgit/alpine-mksite/commit/?id=cd0cb2e673bbc7cd4f4f8f45a3c52ac936cdfe45 the commits that added the logo 2017-06-21 18:33:25 According to the SVG, could we say this logo is under common creative license ? 2017-06-21 18:33:37 I'm refering to : xmlns:cc="http://creativecommons.org/ns#" 2017-06-21 18:34:54 that's just the XML namespace for some metadata 2017-06-21 18:35:10 <_ikke_> right, it's not a declaration of license 2017-06-21 18:36:31 <_ikke_> .. 2017-06-21 18:38:00 <_ikke_> https://creativecommons.org/ns 2017-06-21 18:38:29 <_ikke_> so no license is defined in there 2017-06-21 18:39:31 h3x4: alpine logo is NOT under creative commons license, there are specific requirements to use the logo 2017-06-21 18:39:32 _ikke_: ok thanks for thoses informations, I'm not familiar with this kind of stuff (license / law...), so if no license, it can be used in aqemu without getting legal problem or DMCA similar thing ? 2017-06-21 18:40:11 h3x4: that is probably fine 2017-06-21 18:40:16 <_ikke_> ianal, but the default is that the work is copyrighted, and you have no license to reproduce 2017-06-21 18:40:23 if there's no license most of the time it's not a good idea to use it, though depending on things using it may be fair use 2017-06-21 18:40:56 like in the US logos are not covered under copyright but under trademark law 2017-06-21 18:41:22 generally, you can do what you want with the logo as long as you are not trying to make it appear that the project is endorsing some other product or service 2017-06-21 18:41:54 <_ikke_> Bun: "However, copyright protection may be available for logo artwork that contains sufficient authorship. In some circumstances, an artistic logo may also be protected as a trademark. " 2017-06-21 18:41:58 we should probably document what rights are granted to reproduce the logo and other design elements 2017-06-21 18:42:06 <_ikke_> https://www.copyright.gov/help/faq/faq-protect.html#title 2017-06-21 18:42:09 but generally speaking 2017-06-21 18:42:15 yeah, speak to a lawyer ;p 2017-06-21 18:42:27 <_ikke_> Bun: so logos can be copyrighted 2017-06-21 18:42:30 it would be nice if alpine publishes the guidelines 2017-06-21 18:42:32 <_ikke_> according to that 2017-06-21 18:42:36 as long as you're not saying, in this instance, that alpine is endorsing or suggesting aqemu should be used 2017-06-21 18:42:40 then it is OK 2017-06-21 18:42:40 the use of this logo is only to custom the virtual system in aqemu, as for arch / gentoo or openbsd / windows... the plan aren't to steal the design or any other malicious thing with it 2017-06-21 18:42:47 yes, that's fine 2017-06-21 18:42:47 copyrighted things can still be used under fair use, sometimes 2017-06-21 18:42:52 <_ikke_> correct 2017-06-21 18:43:16 just do not say it's under creative commons license :P 2017-06-21 18:43:53 <_ikke_> would be a good idea to put a license on it so that it's clear what can and cannot be done 2017-06-21 18:44:35 ok I will not :) I will just reply that the logo haven't any license and it should be ok to merge it into aqemu without getting legal problem 2017-06-21 18:44:42 that's sound fine ? 2017-06-21 18:45:30 sure 2017-06-21 18:45:48 ikke: yes, it just hasn't been gotten to yet 2017-06-21 18:46:05 <_ikke_> right 2017-06-21 18:49:04 ok thanks for your time about this matter ppl 2017-06-21 18:57:05 fcolista is part of the dev team ? 2017-06-21 19:01:00 <_ikke_> h3x4: iirc, he is 2017-06-21 19:24:59 yes 2017-06-21 19:25:00 he is 2017-06-21 19:33:20 kaniini: ok I asked this since he was the one opening the pull on github, will add it on my comment 2017-06-21 19:33:21 thanks 2017-06-21 20:04:34 as an example, https://insights.ubuntu.com/wp-content/uploads/a5e4/Screenshot-from-2016-02-10-10-33-59.png would be an unacceptable usage of our logo 2017-06-21 20:04:56 as it is included with competitors at a much lower DPI rendering 2017-06-21 20:11:14 will make a screenshot on how the logo will be used on aqemu for make it clear on how and for what purpose exactly 2017-06-21 20:11:21 one moment 2017-06-21 20:11:38 i'm quite certain your use is fine 2017-06-21 20:11:57 you're not writing a hit piece about us and then rendering our logo at very low DPI to make us look even more shitty :) 2017-06-21 20:13:07 but the real question is 2017-06-21 20:13:15 would this be an acceptable use of the alpine linux logo? http://img.asie.pl/O1wI 2017-06-21 20:14:46 of course 2017-06-21 20:15:50 http://www.imgup.ru/images2/25c6x321011.png 2017-06-21 20:16:01 but probably not an acceptable use of the ubuntu or docker logos 2017-06-21 20:16:02 haha 2017-06-21 20:16:37 see, this is why you always use the SVG version of a logo 2017-06-21 20:16:44 if all DPIs are equal, nobody can complain about DPIs 2017-06-21 20:16:52 they can still complain about logo sizes though 2017-06-21 20:17:28 h3x4: yeah that's fine. you're just noting that the VM is an alpine VM, which is a factual statement 2017-06-21 20:18:00 it's not alphabetically fair, though. you seem to be sorting in a case-sensitive manner 2017-06-21 20:18:10 it appears you're trying to promote arch linux as higher, and thus better, than alpine 2017-06-21 20:18:28 I just created those virtual system for the screenshot 2017-06-21 20:18:45 quiet asie 2017-06-21 20:18:49 :D 2017-06-21 20:29:24 asie: :D 2017-06-21 20:31:46 it's not really promoting arch, just showing that it's only good on a Virtual System 2017-06-22 01:52:04 how I can use i386 image? 2017-06-22 01:52:21 I still can't find this in DOC 2017-06-22 01:56:48 That would be "x86" in the downloads. 2017-06-22 01:56:55 Other than that, there's no difference, right? 2017-06-22 02:18:17 made this work now... is i386:alpine I was trying inverse 2017-06-22 13:23:14 What's the policy with regards to aports for packages that are prefixed inside /opt/, are these barred from going upstream? 2017-06-22 13:24:31 or /opt/ for that matter 2017-06-22 13:41:49 yes 2017-06-22 13:41:56 /opt is not for distro packages 2017-06-22 13:45:46 Hi all, I've been experiencing a "out of memory" problem with a software compiled on alpine in conditions where such error shouldn't happen. I was hoping someone with better knowledge of musl libc could provide some feedback on this. I've compiled https://github.com/lh3/bwa statically on alpine and now run it on a Centos 6.6 machine. While running the software has a more or less constant 10GB memory 2017-06-22 13:45:48 footprint. It runs for a few hours and then fails with http://dpaste.com/2E33SS4 2017-06-22 13:46:03 The line where it fails is a realloc() call. 2017-06-22 13:46:29 I should also say that the system where this is run has 1TB of RAM so memory shouldn't be an issue. 2017-06-22 13:48:36 I've been reading on reasons that could lead to realloc() failing and one of them seems to be memory fragmentation. It's a longshot guess but does anyone know how muslc handles these kinds of allocations? 2017-06-22 14:07:16 1TB of RAM, oh brother 2017-06-22 14:07:44 Unode: it may be useful to check the memory usage patterns at crash time 2017-06-22 14:07:49 Is edge snapshotted only for minor version upgrades like 3.6->3.7 or also sub-minor version upgrades like 3.6.2->3.6.3 too? 2017-06-22 14:08:05 maybe through /proc/$whatever/maps 2017-06-22 14:08:09 mgoodwin: only minor 2017-06-22 14:08:22 Shiz: any suggestions on how to go about that? 2017-06-22 14:08:27 whenever 3.x.0 is released, a new 3.x-stable branch is created 2017-06-22 14:08:37 and all 3.x.y releases are snapshots of that branch 2017-06-22 14:09:03 Unode: gdb :) 2017-06-22 14:09:13 Ok so 3.6.2 is frozen, and 3.6-stable gets fixes at that becomes 3.6.3 eventually (also frozen) 2017-06-22 14:09:16 to make sure it doesn't at least die when the process crashes 2017-06-22 14:09:18 Allright, thanks shiz. I'll keep all the packages going to /opt for myself :P 2017-06-22 14:09:21 Shiz: was 'afraid' you'd say that :D 2017-06-22 14:09:40 mgoodwin: right -- keep in mind that patch versions are really only relevant for release media 2017-06-22 14:09:42 like the iso 2017-06-22 14:09:53 and docker tags? 2017-06-22 14:09:59 the repos for any 3.x.y install are always 3.x. 2017-06-22 14:10:08 i don't know if docker has patch tags, but if so then probably yes 2017-06-22 14:10:09 Probably no surprise that I'm mostly docker focused (that's how I was exposed to alpine) 2017-06-22 14:10:19 i've only ever seen alpine:3.x used 2017-06-22 14:10:21 :P 2017-06-22 14:11:01 elegast: that's likely why you had to specify options="!fhs" in the first place 2017-06-22 14:12:21 mgoodwin: point being -- it's only relevant for the .apk files on install/boot media 2017-06-22 14:12:29 the repos are not segmented by patch version, only by minor version 2017-06-22 14:12:38 Shiz: Oh I hadn't begun compiling them yet, to be honest. I'm still in the process of splitting version numbers and segregating extension modules 2017-06-22 14:13:04 there's a check that disallows any files in /srv, /opt and /usr/local 2017-06-22 14:13:09 as they are not fit for distro packaging 2017-06-22 14:13:18 unless you add !fhs to options= 2017-06-22 14:13:24 But I'll keep in mind I'll have to specify those options, thanks! 2017-06-22 14:18:16 Unode: I ran into a similar issue the other day, the fix was using sysmalloc, instead of the assumption of using malloc in (gnu) libc. 2017-06-22 14:19:05 elegast: do you know/have any documentation that explains the differences? 2017-06-22 14:19:36 No, I'm afraid not, luckily someone else had pointed me in that direction 2017-06-22 14:19:54 elegast: thanks! I'll give it a go. At this point I'm still trying to reproduce the issue 2017-06-22 14:20:14 Good luck, Unode 2017-06-22 14:20:21 elegast: just to clarify, what you saw was malloc related or also realloc and the whole collection of alloc -related functions? 2017-06-22 14:21:43 It segfaulted with out of memory issues, caused by malloc specifically, I'm not sure if or how this woudl relate to other functions in that order, say re-alloc, but I wouldn't be surprised if its closely related 2017-06-22 14:22:39 There's bits of info about the issue I had here: https://mx.gw.com/pipermail/tcsh-bugs/2015-July/000967.html 2017-06-22 14:25:11 Though, after rereading your messages, it might not be related, since it seems your piece of software had been running for a while, only to fail on the realloc, it seems odd to use say libmusl malloc, and then remalloc using a different implementation 2017-06-22 14:25:47 it's not libmusl, just musl 2017-06-22 14:25:51 libc is what it implements :p 2017-06-22 14:25:54 and libm, and libpthread... 2017-06-22 14:26:06 right right, my bad 2017-06-22 14:26:26 I'm not too knowledgeable about those things just yet :p 2017-06-22 14:40:01 Unode: there atleast seems to be a USE_MALLOC_WRAPPERS flag during compile time for bwa that might affect things 2017-06-22 14:41:52 elegast: yeah, the thing is I'm fully guessing now since I don't see any memory usage that might explain the "out of memory". Specially considering the machine has 1TB of RAM and the software is taking at most 10-20GB. 2017-06-22 14:42:25 there might be a spike when it crashes, but I've not managed to reproduce it out of the 'real' usage which takes hours to crash :) 2017-06-22 14:43:29 I'm sort of throwing the blame around and wondering if I could be hitting some odd limitation/bug of muslc since I can't really explain the error. 2017-06-22 14:44:17 I saw the MALLOC_WRAPPERS flag but don't yet understand what it does. C is also not my thing. 2017-06-22 14:46:15 strace 2017-06-22 14:46:35 Yeh It's not my forte either, but indeed, needing hours of runtime just to reproduce a bug seems like a bit of a problem 2017-06-22 14:47:35 oh its the tcsh thing 2017-06-22 14:47:40 I doubt the problem is actually out of memory, I think the call fails because its doing the wrong thing 2017-06-22 14:47:42 ncl: tried that.. good luck reading through a TB long log. 2017-06-22 14:47:56 ncl: no it's a different piece of software 2017-06-22 14:48:15 though my guess a similar problem 2017-06-22 14:49:23 Unode: usually the last bits of the strace are the important parts 2017-06-22 14:49:39 or run it in gdb and backtrace 2017-06-22 14:50:46 ncl: that's what I'm trying at the moment but I'm guessing all the way. I don't know how to use these tools properly. 2017-06-22 14:51:10 and strace, how do you handle the logging when the software has to run for hours? 2017-06-22 14:51:17 Unode: does it crash, or does it properly exit, though with a failure condition? 2017-06-22 14:51:36 it prints what I pasted above and exits with code 1 2017-06-22 14:51:54 so it's not like the system kills it. 2017-06-22 14:52:04 i'm just now seeing your first lines 2017-06-22 14:52:22 let me repaste then 2017-06-22 14:52:23 dna sequencing, neat 2017-06-22 14:52:35 Hi all, I've been experiencing a "out of memory" problem with a software compiled on alpine in conditions where such error shouldn't happen. I was hoping someone with better knowledge of musl libc could provide some feedback on this. I've compiled https://github.com/lh3/bwa statically on alpine and now run it on a Centos 6.6 machine. While running the software has a more or less constant 10GB memory 2017-06-22 14:52:36 footprint. It runs for a few hours and then fails with http://dpaste.com/2E33SS4 2017-06-22 14:52:49 the wrappers don't do much 2017-06-22 14:52:55 they just check for NULL return 2017-06-22 14:52:59 so i don't think that's it 2017-06-22 14:53:02 I think the wrappers are there just to catch the errors. 2017-06-22 14:53:12 It seems realloc returns NULL 2017-06-22 14:53:41 which is why I thought the issue would be at lower level, i.e. something with how it interacts with libc and performs the actual allocation. 2017-06-22 14:54:10 highly influenced by https://stackoverflow.com/a/4925670 2017-06-22 14:58:21 "kroundup32" for a (likely 64 bit) size_t seems odd, but you said it worked on centos so maybe not 2017-06-22 14:58:29 both running the same version, right? 2017-06-22 14:59:15 or is that rounding to 32s 2017-06-22 14:59:37 same version? 2017-06-22 15:00:08 of bwa 2017-06-22 15:01:09 I'm not sure I understand the question. I compiled bwa statically on alpine and then scp'ed it to a centos machine. The binary is the same. 2017-06-22 15:01:10 Shiz: also it's improperly reusing the pointer in realloc assigment + as an arg, which may be it? 2017-06-22 15:01:20 Unode: oh 2017-06-22 15:01:37 so it only fails when ran in alpine, or? 2017-06-22 15:01:41 the only two calls to that function are https://github.com/lh3/bwa/blob/5961611c358e480110793bbf241523a3cfac049b/bwamem.c 2017-06-22 15:01:54 could you explain what the bwa setup that actually works is like 2017-06-22 15:02:08 Also I wonder about this: "file bwa -> bwa: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked, stripped" but "ldd bwa" -> "statically linked" 2017-06-22 15:02:53 that's no reason for concern 2017-06-22 15:03:05 static PIE binaries are semantically ELF shared objects 2017-06-22 15:03:17 they're not any less static :) 2017-06-22 15:04:33 Shiz: ok, so the compiled binary does 'sometimes' work. The binary is from: https://gitlab.com/ngless/ngless/builds/artifacts/master/raw/bin/ngless?job=build-and-test . bwa is embedded and is unpacked on execution. 2017-06-22 15:04:52 The compilation recipe is: https://github.com/luispedro/ngless/blob/master/.gitlab-ci.yml 2017-06-22 15:05:24 Shiz: e.g. the shared/dynamic part implies the resultant binary itself can be *used* as a shared object, not that it *isnt* using statically linked objects itself, right? 2017-06-22 15:05:33 https://github.com/luispedro/ngless/blob/master/Makefile#L147 to be precise 2017-06-22 15:06:02 I've always wondered about that myself, aswell 2017-06-22 15:06:23 it means the ELF object type is shared object 2017-06-22 15:06:31 which has some implications on what kind of relocations it can do, for instance 2017-06-22 15:06:44 originally PIE was only a thing for dynamically linked binaries and shared objects 2017-06-22 15:06:48 static PIE got hacked in later 2017-06-22 15:06:57 meaning static PIE binaries are shared objects too, strictly 2017-06-22 15:06:59 :P 2017-06-22 15:07:15 that does not make it any less confusing :) 2017-06-22 15:07:16 but yes, shared object type does not mean it necessarily has DT_NEEDED entries (dynamic dependencies) 2017-06-22 15:07:26 but ok, thats starting to make sense 2017-06-22 15:07:42 You lost me :P 2017-06-22 15:07:51 i don't want to give too much of a reductive definition since it's a complex subject without knowing the ELF standard 2017-06-22 15:07:56 Unode: point being, it's no reason for concern 2017-06-22 15:08:00 ok 2017-06-22 15:08:02 shared object is perfectly fine for static binaries 2017-06-22 15:08:13 it's an artifact of using position-independent code 2017-06-22 15:08:28 it doesn't mean it has dynamic dependencies :) 2017-06-22 15:08:43 What sort of confused me was the "dynamically linked, stripped" vs "statically linked". 2017-06-22 15:08:58 right 2017-06-22 15:09:05 that's because file(1) is just wrong 2017-06-22 15:09:07 :P 2017-06-22 15:09:13 figured as much :D 2017-06-22 15:09:21 wrong magic bits 2017-06-22 15:09:49 so, back to the issue. The binary does work. 2017-06-22 15:09:56 And it doesn't always fail. 2017-06-22 15:10:44 which makes it even more fun to debug 2017-06-22 15:10:56 although it has to be said 2017-06-22 15:11:03 centos ldd will not work on musl binaries 2017-06-22 15:11:05 :P 2017-06-22 15:11:12 it will always report 'statically linked' 2017-06-22 15:11:27 So both are wrong. great! 2017-06-22 15:11:42 it's because glibc ldd is very broken 2017-06-22 15:11:52 the best way to determine dynamic-ness of a binary is to use a tool like readelf 2017-06-22 15:12:31 ok but I can at least trust that it's static since it works across systems. :) 2017-06-22 15:12:43 right 2017-06-22 15:12:46 for future reference: 2017-06-22 15:12:55 readelf -l | grep NEEDED 2017-06-22 15:13:01 sorry 2017-06-22 15:13:04 readelf -l | grep INTERP 2017-06-22 15:13:09 readelf -d | grep NEEDED 2017-06-22 15:13:15 if neither has output, your binary is static 2017-06-22 15:13:29 I see 2017-06-22 15:13:39 (strictly even just the first should be fine) 2017-06-22 15:13:53 (the latter check is also needed on dynamic libraries) 2017-06-22 15:14:00 regardless 2017-06-22 15:15:35 ok seems like now it's taking longer to fail... 2017-06-22 15:15:53 I'm starting to think it's just one of those odd ones :/ 2017-06-22 15:15:57 Moving to a smaller machine :P 2017-06-22 15:18:15 Could there be differences in say struct sizes/memory alignment across alpine (where it is compiled) and centos (where it is deployed)? 2017-06-22 15:18:37 it's a static binary 2017-06-22 15:18:41 there's no differences to be had 2017-06-22 15:18:54 the only variable would be the kernel (and environment vars/rootfs) 2017-06-22 15:21:36 I have a slight problem with the way syslinux is packaged 2017-06-22 15:21:58 it populates /boot even when that's unnecessary 2017-06-22 15:22:35 ah, so say except for when you pass structs to ioctls and such, there should never be alignment issues so long as it is statically compiled? 2017-06-22 15:22:46 I also have a problem with abuild. it no longer builds anything :) 2017-06-22 15:23:32 <_ikke_> TBB: What does it do then? 2017-06-22 15:23:35 elegast: yes 2017-06-22 15:23:47 it does nothing 2017-06-22 15:23:53 except for "all failed" 2017-06-22 15:23:59 been like that for a while now 2017-06-22 15:24:19 that sounds like a set -e thing 2017-06-22 15:24:25 TBB: try adding set -x to the top of the script 2017-06-22 15:24:30 Shiz: huh. interesting. Sorry for all the noob questions btw 2017-06-22 15:24:32 and pastebin the last 20 or so lines it outputs before all failed 2017-06-22 15:24:52 the script being abuild itself here 2017-06-22 15:26:48 apk-abuild add basically fails to install build time deps, returns 1 and runs "die all failed" 2017-06-22 15:27:34 that's a one line description of what running the script with set -x reveals 2017-06-22 15:30:20 and why does abuild-apk fail? :P 2017-06-22 15:30:40 I have no idea, but installing those dependencies manually works using the regular apk 2017-06-22 15:33:18 abuild-apk seems to have an extra local repository added under the user account's home/packages which doesn't exist 2017-06-22 15:34:16 ... but I suppose it would create that for the resulting package anyway 2017-06-22 15:34:36 right 2017-06-22 15:40:50 could this be something as simple as installing the deps failing because there's one pending error in apk operations? 2017-06-22 15:41:14 (shadow-doc conflicting with coreutils-doc and util-linux-doc) 2017-06-22 15:42:30 I think it is because it affects the return code of dependency install attempts 2017-06-22 15:44:41 that may be 2017-06-22 15:45:48 aaaaaaaand since package "docs" is installed I can't just remove shadow-doc or the other doc packages to resolve the conflict :) 2017-06-22 15:46:21 and this forcing me to do apk del docs would be a ridiculous solution... 2017-06-22 15:46:56 ah fuck it, this is why I wrote my own build system, to stop building packages on workstations and isolate the process to clearly restricted chroots 2017-06-22 15:47:04 alpine has buildlab... 2017-06-22 15:47:06 :) 2017-06-22 15:47:08 time to go that way 2017-06-22 15:47:13 which is exactly the point 2017-06-22 15:48:10 buildlab -C -p somecrootdit -m nl.alpinelinux.org -v edge -a x86_64 2017-06-22 15:48:12 :P 2017-06-22 15:48:19 s/somecrootdit/somechrootdir/ 2017-06-22 15:48:33 and then 2017-06-22 15:48:44 buildlab -p somechrootdir -b /some/APKBUILD 2017-06-22 15:49:58 I'll go with what I know and control myself 2017-06-22 16:09:20 what is this buildlab you speak of? and where does one find it? 2017-06-22 16:14:07 oh it's part of abuild 2017-06-22 16:44:00 what, except installing openrc, should i install in the minimal-rootfs to get it booting? 2017-06-22 16:47:30 kernel, bootloader 2017-06-22 16:48:18 Shiz: that works fine, but doesn't seem to start any getty / shell 2017-06-22 16:50:05 do you have an /etc/inittab? 2017-06-22 16:51:21 ah, found it, forgot to change the inittab for xen 2017-06-22 16:52:04 had to put /dev/hvc0 instead of /dev/tty0 2017-06-22 16:59:41 ah, and add the hvc0 to securetty 2017-06-22 17:04:15 :P 2017-06-22 17:43:50 hi ! i have a question about alpine linux in docker container 2017-06-22 17:44:43 i changed the docker container to openrc but always if i start alpine i cannot start dbus service 2017-06-22 17:44:54 i get this error 2017-06-22 17:44:57 /var/run/dbus/system_bus_socket": Address in use 2017-06-22 17:45:15 if i delete the file i can start dbus 2017-06-22 17:46:08 but after start again the docker container it can alo not start dbus 2017-06-22 17:55:08 don't use init systems in docker containers, they're not meant for that 2017-06-22 18:00:13 ok can i start without openrc dbus ? 2017-06-22 18:01:36 you'd probably want to start it in either your docker entrypoint script or by using something like supervisord 2017-06-22 18:07:00 ok thanks ! only one question i want install gns3 in alpine and i dont now after install server and gui i cannot start the gns3-server it seems a python module is missing 2017-06-22 18:08:12 i get this error if i want start gns3 server typing>=3.5.3.0' distribution was not found and is required by gns3-server 2017-06-22 18:08:50 which package belongs this typing module 2017-06-23 00:07:07 now, what is the mechanism for managing out-of-tree kernel modules in Alpine, or does one exist? 2017-06-23 00:12:59 I suppose I figured the answer to that one out; ahh, the magic of asking questions on IRC 2017-06-23 00:23:04 TBB: chek the zfs apkbuild for examples 2017-06-23 00:23:05 :P 2017-06-23 00:24:10 thank you for the pointer :) 2017-06-23 00:28:43 there's a caveat that we're looking at redoing the way we do kernel packages right now 2017-06-23 00:28:46 kaniini knows more 2017-06-23 05:09:01 just abou to install alpine linux as dual boot with arch 2017-06-23 05:10:01 already have efi boot disk with grub installed & os-prober on sda1 2017-06-23 08:05:55 very strange every attempt to make a bootable image has failed. 2017-06-23 08:06:17 I've made 100's of live usb or dvds 2017-06-23 08:06:28 tried with dd command 2017-06-23 08:06:46 sudo dd bs=4M if=/run/media/lee/storage/OtherOs/alpine-extended-3.6.2-x86_64.iso of=/dev/sdc status=progress && sync 2017-06-23 08:06:51 k3b 2017-06-23 08:06:54 etcher 2017-06-23 08:07:13 tried versions 2017-06-23 08:07:36 mystified1234, copy the contents to usb and install the syslinux bootloader. 2017-06-23 08:08:37 k 2017-06-23 08:08:53 but whys this happening 2017-06-23 08:09:20 our releases should be hybrid, but i didnt test it lately. 2017-06-23 08:09:32 k & thx 2017-06-23 09:39:45 would it be possible to make Xorg (X11) display only ONE color channel? (not all red, green, blue > only one of them, say, only colors within blue channel) 2017-06-23 09:41:02 maybe you can abuse one of those color correction APIs 2017-06-23 09:41:49 I am thinking of a similar effect possible within Gimp (disabling all but one color channels > keeping just blue) 2017-06-23 09:43:45 (or in any editor, even ImageMagick) to extract only one of color channels. In X, though, it would probably require either muting other two, or as you mentioned, Bun. 2017-06-23 09:44:20 if Xorg had some kind of rendering hook it should be easy to do at least 2017-06-23 09:48:53 there are utilities to invert colors e.g. 2017-06-23 09:50:09 it seems that uses the color correction stuff :) 2017-06-23 09:50:11 https://github.com/zoltanp/xrandr-invert-colors/blob/master/xrandr-invert-colors.c 2017-06-23 09:50:55 you can try that and maybe set two of the channels to zero (or whatever makes it not render that color) 2017-06-23 09:52:13 guys in the process of installing as dual boot with arch. bootloaders are not my strength. 2017-06-23 09:52:34 dev/sda1 is efi boot 512mb 2017-06-23 09:52:51 sda3 is ext arch partion 2017-06-23 09:53:50 sda4 is alpine that i've just installed with the dual boot instruction. I'm needing to configure grub 2. to complete the install. 2017-06-23 09:54:01 not sure what to do next ! 2017-06-23 09:55:33 o/ 2017-06-23 09:55:35 did you run update-grub2? did it do something useful? 2017-06-23 09:56:23 if you do it from Arch you can at least use os-prober which should find things automatically, don't think that's available on Alpine 2017-06-23 09:56:40 sda2 swap 2017-06-23 09:57:15 so reboot into arch then update the grub. 2017-06-23 09:57:36 what command to use is it update grub 2017-06-23 09:57:53 https://wiki.archlinux.org/index.php/GRUB 2017-06-23 09:57:57 os-prober -o something like that 2017-06-23 09:58:22 grub-mkconfig -o /boot/grub/grub.cfg for Arch 2017-06-23 09:58:22 thx 2017-06-23 09:58:46 beautiful. hope to be back soon configuring xfce 2017-06-23 09:58:58 thanks ! 2017-06-23 09:59:10 :) 2017-06-23 09:59:16 good luck (especially with EFI) 2017-06-23 10:02:43 ok back in arch 2017-06-23 10:09:09 rip 2017-06-23 10:09:23 not looking good 2017-06-23 10:09:46 so you installed os-prober and ran grub-mkconfig? it should print out all the things it found 2017-06-23 10:10:59 but really, it shouldn't be more difficult than taking the config example on the Alpine wiki and changing some of the parameters for each of your OSes 2017-06-23 10:11:21 grub was successful, but when i rebooted & clicked alpine from the grub menu, it just went blac with no indication od starting the os. The screen went blk with a slight glow, a bit like when the wrong gpu driver is useds 2017-06-23 10:12:06 I'm just going back into bios reactivate in bios . legacy 2017-06-23 10:12:22 it could be that the kernel commandline it created is missing some stuff, dunno 2017-06-23 10:12:35 try removing quiet from the kernel command line 2017-06-23 10:14:04 how to that ? 2017-06-23 10:14:23 you can press "e" while in grub to edit the active commandline, or edit the config itself 2017-06-23 10:15:01 im in edit of grub for alpine 2017-06-23 10:19:30 there is no "quietin alpine advanced optionsthe grub 2017-06-23 10:21:12 try GRUB_GFXPAYLOAD_LINUX=text in grub.cfg, maybe 2017-06-23 10:23:53 im going to go over to grub, be back in a little while 2017-06-23 10:24:59 actually I'm going to take break and watch some rugby league on the box for a while.. 2017-06-23 10:25:21 thanks for your help see you in a hour or so ! 2017-06-23 10:25:44 aussie time 8:30pm 2017-06-23 11:24:10 <_spOOn_> hey guys, can anyone point me to something that explains package lifecycle in alpine? 2017-06-23 11:24:29 <_spOOn_> for example, how do packages move from edge to main? 2017-06-23 11:24:40 <_spOOn_> ie, is it just time based? 2017-06-23 11:24:46 <_spOOn_> or is there somethign further? 2017-06-23 11:26:04 I do not know of any such article 2017-06-23 11:26:33 But that would be a good addition, provided it really does not exist yet 2017-06-23 11:27:18 i don't think it's actually documented somewhere, but I can explain it 2017-06-23 11:27:26 <_spOOn_> Shiz: perfect, thanks 2017-06-23 11:27:30 _spOOn_: packages enter alpine in the testing repo, which is edge-only 2017-06-23 11:27:41 <_spOOn_> just didnt want to waste your time if it was written somewhere 2017-06-23 11:27:41 <_spOOn_> :) 2017-06-23 11:27:50 after a bit of usage, any alpine user who has verified it working can open a request to have it moved to the community repo 2017-06-23 11:27:59 it will then be in community, and in the next alpine release made 2017-06-23 11:28:33 if a package is deemed essential or important to the core alpine system, it is moved to the main repo, where it is maintained by alpine core devs 2017-06-23 11:29:07 alpine releases (like 3.6.0) snapshot the edge main and community repos at the time they are made, and branch off of it 2017-06-23 11:30:05 <_spOOn_> great, thanks, that makes sense 2017-06-23 11:30:14 <_spOOn_> just one further question if I may :) 2017-06-23 11:30:23 in other words, if you want to make sure a package is in the next alpine release, let us know that it works and it can be moved to community :) 2017-06-23 11:30:27 go ahead 2017-06-23 11:30:44 <_spOOn_> how does one request to have something moved from edge to community 2017-06-23 11:31:21 <_spOOn_> sorry, from testing -> community 2017-06-23 11:31:35 open a PR against aports that moves the APKBUILD 2017-06-23 11:31:37 :) 2017-06-23 11:31:52 <_spOOn_> okies, cool, thanks for the clarification 2017-06-23 11:33:50 Thanks for the explanation. While I have you handy, can lbu be used to read and write from/to a disk ? Scenario is the following: alpine instances will be hosted on xen, using diskless or data mode. I would to save instance-specific configuration (lbu) on a write-toggleable read-only media. As far as my understanding goes, I cannot mount a floppy or usb in such a way in xen, which leaves me with a 2017-06-23 11:33:52 VDisk. And AFAIK, the setup script does not offer this option ? I suppose this could be hacked somehow 2017-06-23 11:34:09 lemme answer that question when i get back 2017-06-23 11:34:14 have to travel for a bit real quick 2017-06-23 11:34:26 Sure, tty in a bit 2017-06-23 11:48:30 back 2017-06-23 11:48:43 you can simply tell it a path to save to 2017-06-23 11:49:17 set LBU_BACKUPDIR 2017-06-23 11:49:23 to the path to save the apkovl to 2017-06-23 12:02:04 oh 2017-06-23 12:02:15 Easy, I like that 2017-06-23 12:02:20 Thank you ! 2017-06-23 12:03:35 np 2017-06-23 13:50:14 need help with dual boot Arch & alpine, Grubs not my forte. Usually, I run grub-update or os-prober knconfig commands & that usually fixes. the problem. 0 2017-06-23 13:50:15 On this pc I have efi part sda1 fat 32 boor.esp 512 ram 2017-06-23 13:50:15 On this pc I have efi part sda1 fat 32 boot.esp 512 ram 2017-06-23 13:50:15 sda2 swap 2017-06-23 13:50:15 sda3 ext4 arch my default os 2017-06-23 13:50:17 sda4 ext just installed alpine-linux 2017-06-23 13:50:19 from arch I ran update-grub mkconfig os-prober. Still did not fix the issue. 2017-06-23 13:50:21 alpine is recognised by grub, with default & advanced grub options. 2017-06-23 13:50:23 when I click alpine, nothing happens, No attempt to boot. no text on screen. 2017-06-23 13:50:25 it's like booting linux with the wrong gpu. blk screen with a slight glow. 2017-06-23 13:50:27 when I hir edit grub at alpines grub option . There is nothing stating "quiet" 2017-06-23 13:50:29 If you guys could assist it would be appreciated 2017-06-23 13:51:02 have you tried GRUB_GFXPAYLOAD_LINUX=text in grub.cfg (instead of whichever value it is set to currently) 2017-06-23 13:51:40 Hi bun 2017-06-23 13:51:51 sorry no just come back 2017-06-23 13:52:14 i'm not sure what to do there 2017-06-23 13:52:40 vi /boot/grub/grub.cfg find GRUB_GFXPAYLOAD_LINUX=... and set it to =text ? 2017-06-23 13:58:17 teenage daughter issue. sorry must reboot. 2017-06-23 13:58:20 brb 2017-06-23 13:58:54 still booting at what point do iadd this 2017-06-23 13:59:05 at the point where it's booted and where you can edit the file? 2017-06-23 14:07:37 bun: trying to bpaste my grub, but to big for bpaste.net 2017-06-23 14:08:15 <_ikke_> whatabout tpaste.us ? 2017-06-23 14:09:23 https://bpaste.net/show/941277c0617f 2017-06-23 14:09:36 this is the part related just to alpine 2017-06-23 14:15:02 http://tpaste.us/K6WY 2017-06-23 14:16:09 oh, right, it translates it to set gfxpayload=... 2017-06-23 14:16:33 so where it says =keep change it to =text 2017-06-23 14:19:14 also I don't see any initrd lines for alpine 2017-06-23 14:19:24 that's probably not good 2017-06-23 14:19:56 that =keep, is related only to arch 2017-06-23 14:20:17 there are 4 of them 2017-06-23 14:20:30 hm, I'm not sure what the default behavior is 2017-06-23 14:21:02 start by adding appropriate initrd lines to alpine entries I guess 2017-06-23 14:21:43 so still change keep to text or not 2017-06-23 14:22:02 changing arch things will not matter 2017-06-23 14:22:17 so on all4 of them 2017-06-23 14:22:24 so on all 4 of them 2017-06-23 14:22:32 you can add the set command to alpine if you want 2017-06-23 14:22:49 I got no idea what that means 2017-06-23 14:23:23 you have alpine entries that don't have "set gfxpayload=.." 2017-06-23 14:23:37 I got no idea what that meens 2017-06-23 14:23:44 "meens: 2017-06-23 14:24:29 you said so yourself, the alpine menuentry items do not have a line set gfxpayload=keep 2017-06-23 14:25:00 but your editor has the magic ability to add such lines (!) 2017-06-23 14:25:20 i in nano now 2017-06-23 14:25:48 more importantly you can see there is no "initrd /boot/xxxxx" lines 2017-06-23 14:25:58 (compare to arch) 2017-06-23 14:28:24 changed keep to text 2017-06-23 14:30:35 somehow I don't think we're on the same page here 2017-06-23 14:30:44 Anyone have any luck changing the cpu governor with cpufreq-set? It keeps telling me to make sure the governor module is started, but I dont see it in the modules folder. 2017-06-23 14:38:57 bun they are missing 3 lines 2017-06-23 14:39:01 linux /boot/vmlinuz-linux root=UUID=cab1f283-589d-4475-957b-7429b844cfce rw quiet 2017-06-23 14:39:02 echo 'Loading initial ramdisk ...' 2017-06-23 14:39:02 initrd /boot/initramfs-linux-fallback.img 2017-06-23 14:39:40 the echo is irrelevant 2017-06-23 14:40:10 copy the initrd line, change the file name to point at the alpine initrd 2017-06-23 14:41:04 it's called something like initramfs-hardened in your case I think, verify it yourself 2017-06-23 14:45:42 initrd /boot/initramfs-hardened.img 2017-06-23 14:46:00 done that to both grub entries for alpine 2017-06-23 14:46:14 should i now save & reboot 2017-06-23 14:46:24 no it's not called .img 2017-06-23 14:46:38 ok 2017-06-23 14:46:51 initramfs-vanilla (or your variant) 2017-06-23 14:46:53 no .img 2017-06-23 14:47:05 ls /boot will tell you 2017-06-23 14:48:30 efi grub initramfs-linux-fallback.img initramfs-linux.img vmlinuz-linux 2017-06-23 14:48:55 but thats arch not alpine 2017-06-23 14:49:06 alpine's /boot yes :) 2017-06-23 14:49:50 no all of this taken from arch boot folder 2017-06-23 14:51:38 that's where your grub config is yes 2017-06-23 14:52:24 but before loading the files for alpine there is a "set root" command which makes grub look at alpine's partition 2017-06-23 14:52:45 because that is where the kernel and initramfs files are 2017-06-23 14:53:30 so mnt /dev/sda4 /mnt && ls /mnt/boot 2017-06-23 14:53:48 that will tell you about alpine's /boot 2017-06-23 14:54:02 this grub.cfg from alpine 2017-06-23 14:54:06 http://tpaste.us/oZEr 2017-06-23 14:54:29 ignore it 2017-06-23 14:54:34 ok 2017-06-23 14:54:44 it is not used 2017-06-23 14:55:01 yes 2017-06-23 14:55:11 make sense 2017-06-23 14:56:03 so what do in need to change re .img 2017-06-23 14:56:20 did you find out the filename yet? 2017-06-23 14:56:54 initramfs-vanilla 2017-06-23 14:57:22 ok so put /boot/initramfs-vanilla and save 2017-06-23 14:58:40 and maybe show us the full file before you reboot :) 2017-06-23 14:58:54 the arch grub.cfg that is 2017-06-23 15:01:30 ok about to save 2017-06-23 15:02:46 http://tpaste.us/jeRm 2017-06-23 15:03:38 that's not the full file :P 2017-06-23 15:03:44 my teenage daughter is mad at me because I'm keeping her from sleeping. Hearing the keyboard. 2017-06-23 15:04:10 ok.hang 5 2017-06-23 15:04:14 type louder? ;p 2017-06-23 15:04:22 :) 2017-06-23 15:04:28 tell her she can stay up late, and buy a silent keyboard 2017-06-23 15:04:37 "oh im sorry" *mash mash mash mash* lol 2017-06-23 15:06:27 http://tpaste.us/BeKb 2017-06-23 15:06:41 thx again BUN 2017-06-23 15:11:26 Bun: intriams Vanila not found! Irebooted 2017-06-23 15:12:52 Have to go daughter having a Tanty, witha capital T. 2017-06-23 15:13:14 see you a bit later... perhaps ! 2017-06-23 15:13:21 Againa big Thx. 2017-06-23 15:24:03 I blame scv for this! ;) 2017-06-23 17:11:34 do the build servers not want abuild tests to use /tmp for making files/directories during tests? 2017-06-23 17:48:47 Wow, getting xenserver to play fine with diskless alpine install was laborious. How do you create a wiki entry ? 2017-06-23 17:52:10 AlexIncogito: not sure, I didn't see a way where the general public is allowed to add to the wiki, the iso creation guide has lots of problems 2017-06-23 17:52:54 you can't just create an account on the wiki and edit pages? 2017-06-23 17:53:28 Bun: will try, haven't had a chance 2017-06-23 18:01:52 I don't think they have write priv, do they ? Let me know if you get a chance to try okeuday_bak, got mailbox issues atm 2017-06-23 18:03:34 https://wiki.alpinelinux.org/wiki/Special:ListGroupRights 2017-06-23 18:06:12 anyone should be able to edit the wiki, i think 2017-06-23 18:06:16 cc AlexIncogito 2017-06-23 18:06:29 okeuday_bak: anything specifically failing re: /tmp ? 2017-06-23 18:07:18 Shiz: I can't see the test output, since the makefile isn't outputting it to the log, that should be the cause of the test failure though since everything else is isolated 2017-06-23 18:07:44 Shiz: it would either be creating a directory in /tmp, and/or creating files in the /tmp path, since both are done 2017-06-23 18:08:02 maybe it's not respecting the TMPDIR env var 2017-06-23 18:08:35 Shiz: yeah, the test isn't attempting to use the TMPDIR env var, it assumes /tmp exists and can be used 2017-06-23 18:09:38 Shiz: fantastic 2017-06-23 18:09:55 Shiz: so, for now, I just disabled that test, since postgresql/mysql has also been problematic, Travis has these enabled, but they need to be configured for the tests to work 2017-06-23 18:33:05 alpine has dpkg, this means I'll be able to install Debian packages? 2017-06-23 18:33:43 AlexIncogito: yeah, was able to modify the wiki 2017-06-23 18:33:50 What happens when one of them wants to pull in glibc? 2017-06-23 18:34:09 what you can do is install packages in .deb format 2017-06-23 18:34:17 straight debian packages won't work though 2017-06-23 18:34:27 they have different dependencies and are compiled against a different base system 2017-06-23 18:34:29 namely, debian 2017-06-23 18:34:45 (as for why we package dpkg, I have absolutely no idea -- it's most likely some legacy thing) 2017-06-23 18:34:56 any way to 'convert' .deb to .apk? 2017-06-23 18:35:08 none whatsoever 2017-06-23 18:35:09 Why doesn't musl install its libc under a different name? Then a dual-libc system would be possible. Right? 2017-06-23 18:35:23 or well 2017-06-23 18:35:25 only those debs that are non-arch dependent, and those that are linked against musl 2017-06-23 18:35:31 no way to convert a Debian .deb to something that will work on Alpine 2017-06-23 18:35:36 :P 2017-06-23 18:35:49 Maxpm: we are not interested in a dual-libc system 2017-06-23 18:35:52 yeah, whatever software would have ot have been built for alpine or something really close to it, which debian is not 2017-06-23 18:36:06 and also, if that were the case, it would make no sense to package in .deb anyway 2017-06-23 18:37:41 I mean, in general, for any distro, why don't libcs "identify" themselves in their filenames? They're just libraries. If I reimplemented libpng, I wouldn't call it libpng. 2017-06-23 18:38:05 uhm 2017-06-23 18:38:12 libc is a standard API 2017-06-23 18:38:26 if it wanted to be a special snowflake and have its own name, very standard build things like -lc would not work anymore 2017-06-23 18:41:08 if you reimplemented libpng with an identical api, you'd want to call it libpng if it was a drop-in replacement 2017-06-23 18:41:14 unless you want to modify every package out there using libpng 2017-06-23 18:42:05 I don't think things like -lc should be standard. Have the gcc driver pass -lglibc. Make it explicit. libc may be a standard API, but as we've seen, in practice, things often don't just depend on the abstract API - they depend on a specific implementation and its extensions. And for binary packages, compatibility is even more hopeless. 2017-06-23 18:42:28 well, you can fight that one out with posix 2017-06-23 18:42:31 we abide by posix 2017-06-23 18:44:13 Standards compliance should improve portability, not complicate it. 2017-06-23 18:44:46 a standard lib name with a standard api does improve portability 2017-06-23 18:46:07 A standard API, certainly. 2017-06-23 18:46:56 POSIX is that standard API 2017-06-23 18:47:07 A standard lib name implies a false portability when the various lib implementations are not drop-in replacements for each other, which is the case for libc. 2017-06-23 18:47:18 they are drop-in replacements if you abide by POSIX. 2017-06-23 18:47:51 yes, even on a binary level 2017-06-23 18:48:04 "you" being what, application developers? 2017-06-23 18:48:30 yes 2017-06-23 18:48:40 and also the libc, of course 2017-06-23 18:48:44 which doesn't mean restricted to POSIX 2017-06-23 18:48:49 but conforming to all its requirements 2017-06-23 18:49:04 for the libc 2017-06-23 18:49:10 Sure. 2017-06-23 18:50:37 Let me be clear about what I'd like to see: 2017-06-23 18:52:11 Have packages glibc and musl-libc both satisfy the dependency requirement posix-libc. Applications that want glibc can ask for it explicitly, and applications that just want POSIX can default to musl. This would require the libcs not to clobber each other on the filesystem. 2017-06-23 18:54:12 But we've settled into this pattern of linking -lc regardless of what we expect that to be. 2017-06-23 18:56:40 we're not gonna package glibc 2017-06-23 18:57:28 I'm asking why no other distros have tried this. 2017-06-23 18:58:08 no demand for it 2017-06-23 18:58:12 you're welcome to try 2017-06-23 18:59:45 Story of my life. :) 2017-06-23 19:00:02 well, here's one for you 2017-06-23 19:00:29 what if you have an application that needs two libraries 2017-06-23 19:00:35 put all of the non-POSIX glibc thing in its own library and go from there :^) 2017-06-23 19:00:39 one that is linked against glibc in your distro, and one that's linked against musl 2017-06-23 19:00:42 now what? 2017-06-23 19:00:47 you can't mix libcs in a process 2017-06-23 19:00:48 Oof, good question. 2017-06-23 21:21:02 Bun: that is basically what gcompat is about 2017-06-23 21:21:47 Bun: http://pkgs.alpinelinux.org/package/edge/testing/x86_64/gcompat 2017-06-23 21:22:15 it features a customized ELF loader that injects the glibc extensions into musl 2017-06-23 21:22:42 sounds terribly frightening and cool 2017-06-23 22:33:23 this confuses me: https://pkgs.alpinelinux.org/contents?branch=edge&name=elfutils&arch=x86_64&repo=main 2017-06-23 22:34:46 https://git.alpinelinux.org/cgit/aports/tree/main/elfutils/APKBUILD?h=3.6-stable#n91 2017-06-23 22:34:49 doesn't confuse me 2017-06-23 22:34:50 :P 2017-06-23 22:35:32 Ah, I see what you mean. 2017-06-23 22:36:25 i wonder what the resaon for it is 2017-06-23 23:10:23 hey guys need some help with grub on a dual boot efi, 2017-06-23 23:10:48 Bun helped me about 7hrs 2017-06-23 23:12:03 you got the initramfs filename wrong at least 2017-06-23 23:14:18 ah 2017-06-23 23:14:36 ive modified & screwed things up this morning 2017-06-23 23:15:12 it sucks that os-prober didn't Just Work™ 2017-06-23 23:15:16 maybe not 2017-06-23 23:15:30 hang 5 ill reboot 2017-06-23 23:16:03 no. 2017-06-23 23:16:07 there is an error 2017-06-23 23:16:11 Syntax error at line 183 2017-06-23 23:17:29 in nao how to show line numbers 2017-06-23 23:17:35 in nano how to show line numbers 2017-06-23 23:17:44 not 2017-06-23 23:18:19 nano -c apparently 2017-06-23 23:18:51 ctrl-_ is go to line#ls 2017-06-23 23:19:11 windowing problem >.> 2017-06-23 23:19:22 thx 2017-06-23 23:20:47 no that makes terminal smaller 2017-06-23 23:21:54 head -183 $file | tail -n 1 ;-) 2017-06-23 23:24:44 syntax error near unexpected token `)' 2017-06-23 23:25:18 head -183 $/boot/grub/grub.cfg | tail -n 1 ;-) 2017-06-23 23:25:40 I don't think the shell accepts smileys 2017-06-23 23:25:51 and get rid of the $ 2017-06-23 23:25:53 :0 2017-06-23 23:25:58 ::) 2017-06-23 23:26:05 just realived 2017-06-23 23:26:09 just realized 2017-06-23 23:26:11 is there anyone who could assist me in changing my cpu governor. I have the required packages. when I run cpufreq-set -r -g performance it says something about loading a module, but the module that needs loaded isnt where it says it is. 2017-06-23 23:26:16 is it time to learn vi yet 2017-06-23 23:26:38 cannot open '$/boot/grub/grub.cfg' for reading 2017-06-23 23:26:45 $ 2017-06-23 23:26:51 k 2017-06-23 23:27:18 # 2017-06-23 23:27:34 m4chm4n: https://pkgs.alpinelinux.org/contents?file=cpufreq*ko&path=&name=&branch=v3.6&repo=&arch= ? 2017-06-23 23:27:34 nothing happening 2017-06-23 23:27:57 mystified: the error is probably somewhere else, can't you just upload the whole config again? 2017-06-23 23:28:29 thanks bun 2017-06-23 23:29:01 ok so what do ido now 2017-06-23 23:29:09 aaaaaaahhhh 2017-06-23 23:29:24 no hair 2017-06-23 23:29:29 :) 2017-06-23 23:29:39 mystified: how about you give me ssh access and I'll fix it in 5 minutes ;P 2017-06-23 23:29:46 BTW a big thx for offering your help 2017-06-23 23:29:55 how to 2017-06-23 23:30:23 if I edit the config you pasted earlier, do you have a method of getting it on the system? 2017-06-23 23:30:44 i can install teamviewer 2017-06-23 23:31:06 or another method if you prefer 2017-06-23 23:31:20 chrome desktop 2017-06-23 23:31:28 never used i 2017-06-23 23:31:32 ^ 2017-06-23 23:31:32 never used it 2017-06-23 23:32:03 what would you like me to install & or configure 2017-06-23 23:32:27 team viewer i used a few yrs ago 2017-06-23 23:32:29 I'd like you to answer the question I asked 2017-06-23 23:32:42 ok 2017-06-23 23:32:56 yes 2017-06-23 23:32:59 thanks bun/ 2017-06-23 23:33:13 paste in to alpine or into arch 2017-06-23 23:33:30 arch /boot/grub/grub.cfg 2017-06-23 23:33:35 yes 2017-06-23 23:33:40 http://sprunge.us/AWTQ it should be this 2017-06-23 23:33:43 im in there at this moment 2017-06-23 23:34:57 so paste this int grub.cfg 2017-06-23 23:35:15 Those are the modules I need. I dont have them, at least I cant find them. How do I download them and where should I save them? 2017-06-23 23:35:16 yes replace grub.cfg with this 2017-06-23 23:35:26 m4chm4n: which kernel do you have installed? 2017-06-23 23:35:52 nevermind I found them. 2017-06-23 23:36:20 how do I go about using them to set my cpu to performance 2017-06-23 23:36:35 did you modprobe them? 2017-06-23 23:37:17 there should be a file like /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor 2017-06-23 23:37:54 but regardless cpufreq-set should work 2017-06-23 23:38:34 ok so I should modprobe the userspace.ko then run the set command? 2017-06-23 23:38:58 sure, modprobe cpufreq_userspace 2017-06-23 23:39:11 if it's intel, you should use performance, which iirc, lets the the CPU handle itself. 2017-06-23 23:39:38 k rebooting 2017-06-23 23:42:37 sorry had to login as root 2017-06-23 23:43:44 that usually helps when setting cpufreq things 2017-06-23 23:44:26 bun: 2017-06-23 23:44:39 thx what ever you did it works 2017-06-23 23:44:43 i'm in 2017-06-23 23:44:49 on my laptop now 2017-06-23 23:44:51 now never touch it again 2017-06-23 23:45:12 haha ^^ 2017-06-23 23:45:38 :) 2017-06-23 23:45:47 (next time you run grub-mkconfig on arch it will break again, I don't think there's any reason to do that again) 2017-06-23 23:46:04 ill save you cfg script on another hdd 2017-06-23 23:46:12 and use it as a ref point 2017-06-23 23:46:42 will never us grub mkconfig 2017-06-23 23:48:30 setup ok now im in alpine ill go configure it 2017-06-23 23:48:42 just out of curiosity. 2017-06-23 23:49:41 the text shown at startup is exactlt the same as the installer "setup-alpine" 2017-06-23 23:49:54 so do I need to move the userspace.ko module to the governor file? 2017-06-23 23:50:11 m4chm4n: huh? you don't need to move ko 2017-06-23 23:50:13 files 2017-06-23 23:50:43 modprobe cpufreq_userspace; cpufreq-set -g userspace 2017-06-23 23:50:57 (not sure why you wanna use userspace but ok) 2017-06-23 23:51:06 when I modprobe it it says its not in lib/modules/4.9.blah blah 2017-06-23 23:51:17 did you upgrade your kernel without rebooting 2017-06-23 23:51:53 the version reported in "uname -a" should have a matching directory in /lib/modules 2017-06-23 23:51:54 no 2017-06-23 23:52:32 it does and the cpufreq modules are in a different location than that 2017-06-23 23:52:47 different how 2017-06-23 23:54:00 cpufreq modules = /lib/modules/4.9.32-0-hardened/kernel/drivers/cpufreq/ 2017-06-23 23:54:21 kernel modules = /lib/modules/4.9.32-0-hardened/ 2017-06-23 23:54:40 that's fine though 2017-06-23 23:54:44 m4chz0n3:~# modprobe cpufreq_userspace.ko 2017-06-23 23:54:45 modprobe: FATAL: Module cpufreq_userspace.ko not found in directory /lib/modules/4.9.32-0-hardened 2017-06-23 23:54:45 m4chz0n3:~# modprobe cpufreq_userspace 2017-06-23 23:54:45 m4chz0n3:~# cpufreq-set -r -f 2GHz 2017-06-23 23:54:45 Error setting new values. Common errors: 2017-06-23 23:54:45 - Do you have proper administration rights? (super-user?) 2017-06-23 23:54:47 - Is the governor you requested available and modprobed? 2017-06-23 23:54:49 - Trying to set an invalid policy? 2017-06-23 23:54:52 - Trying to set a specific frequency, but userspace governor is not available, 2017-06-23 23:54:53 for example because of hardware which cannot be set to a specific frequency 2017-06-23 23:54:55 or because the userspace governor isn't loaded? 2017-06-23 23:54:57 m4chz0n3:~# uname -a 2017-06-23 23:54:59 Linux m4chz0n3 4.9.32-0-hardened #1-Alpine SMP Fri Jun 16 12:20:58 GMT 2017 x86_64 GNU/Linux 2017-06-23 23:55:01 m4chz0n3:~# modprobe cpufreq_userspace 2017-06-23 23:55:03 m4chz0n3:~# cpufreq-set -r -f 2GHz 2017-06-23 23:55:05 Error setting new values. Common errors: 2017-06-23 23:55:07 - Do you have proper administration rights? (super-user?) 2017-06-23 23:55:08 the modprobe didn't complain so that's not the issue 2017-06-23 23:55:09 - Is the governor you requested available and modprobed? 2017-06-23 23:55:11 - Trying to set an invalid policy? 2017-06-23 23:55:15 - Trying to set a specific frequency, but userspace governor is not available, 2017-06-23 23:55:17 for example because of hardware which cannot be set to a specific frequency 2017-06-23 23:55:19 or because the userspace governor isn't loaded? 2017-06-23 23:55:21 m4chz0n3:~# 2017-06-23 23:55:23 m4chz0n3:~# 2017-06-23 23:55:36 you probably want to use a pastebin to avoid spaming everyone 2017-06-23 23:55:54 did you do cpufreq-set -g userspace 2017-06-23 23:57:30 it says the same thing as above 2017-06-23 23:57:51 well that's mighty unhelpful 2017-06-23 23:58:07 okeuday_bak umm you are not even active soo be quiet 2017-06-23 23:58:08 do: cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors 2017-06-23 23:58:39 hi there 2017-06-23 23:58:50 you use musl for security reasons only? 2017-06-23 23:59:15 it's also a lot smaller than glibc :P 2017-06-23 23:59:33 we use musl because it fits our three goals 2017-06-23 23:59:35 small, simple, secure 2017-06-23 23:59:40 (and additionally, correct) 2017-06-24 00:00:11 I already know what governors there are. and when I switch to performance it shows no errors but also doesnt change the settings 2017-06-24 00:00:44 well nevermind I spoke too soon 2017-06-24 00:00:57 m4chm4n: don't you need a cpufrequtils service running, or did you do that? 2017-06-24 00:01:17 you don't need a service 2017-06-24 00:01:29 k, wasn't sure, know you do elsewhere 2017-06-24 00:01:59 at least not to get it to work, I'm sure there could be a service for dynamically changing the policy 2017-06-24 00:02:06 and yes I have that installed and have been using th utils to check what speed I am at. 2017-06-24 00:14:06 well that is figured out. I wonder what module lets me set specific freq. 2017-06-24 00:19:17 m4chm4n: you might want to check to make sure it isn't running hot, since even if you are able to set it high, it may automatically set it low, which would be annoying 2017-06-24 00:19:49 yea I have proper cooling for my laptop 2017-06-24 00:20:05 definitely not my first rodeo 2017-06-24 00:20:23 well with alpine it is but not computers 2017-06-24 00:21:00 but thank you for stating the obvious 2017-06-24 00:21:29 wasn't attempting to, I have a laptop that can overheat, it is the summer here, so it is somewhat on my mind :-) 2017-06-24 00:24:10 well you should get a dual fan dock like I have. it keeps my temps low 2017-06-24 00:25:39 m4chm4n: I have it on a wire shelf which normally is enough, but it isn't decent for loadtesting or anything too major 2017-06-24 00:26:27 yea definitely get a fan dock. well worth the money. 2017-06-24 00:28:34 speaking of temps, I need to get lm_sensors so I can keep an eye on my temps 2017-06-24 00:28:42 m4chm4n: in the cpufrequtils config you should be able to set the governor along with the max/min speed set to the same thing, and then restart, if that doesn't work, something else is likely changing it I would assume, though I haven't tried this on alpine 2017-06-24 00:29:23 when I try to change the freq and not the governor I get an error like the ones in my spam 2017-06-24 00:29:30 on mine, I use GOVERNOR="performance" MAX_SPEED="2201000" MIN_SPEED="2201000" 2017-06-24 00:30:43 m4chm4n: I have only tried to stick with the performance one, trying to maintain the max speed 2017-06-24 00:32:04 m4chm4n: I also had to disable "loadcpufreq" but not sure if that is the same on alpine 2017-06-24 01:03:01 Hey guys for gpu driver install is intel gpu included in xorg-base & does it configure it at the same time 2017-06-24 01:03:22 you will want to add it 2017-06-24 01:03:28 or do in need to that by installing xf86-video-intel 2017-06-24 01:03:33 xf86-video-intel yes 2017-06-24 01:03:33 apk add xf86-video-intel 2017-06-24 01:03:46 thx 2017-06-24 01:04:05 setup-xorg does some kind of autodetection these days i think 2017-06-24 01:04:25 anyone try out the collectd cgp? 2017-06-24 01:05:18 the wiki states to add video-intel and input-synaptics separately 2017-06-24 01:40:52 Guts up running with a base xfce4 2017-06-24 01:41:04 Guys up running with a base xfce4 2017-06-24 01:41:21 is xfce4-goodies available in binaries 2017-06-24 01:51:13 i dont think so myst 2017-06-24 01:52:13 no goodies 2017-06-24 01:52:25 just a few plugins 2017-06-24 02:31:29 hi, i'm trying to build some custom package on raspberry pi but it runs out of space. I can't seem to find which filesystem is configured to do the builds and how I can increase the size. I setup swapspace but that didn't help 2017-06-24 02:33:25 sorry it's the / filesystem that's mounted as tmpfs that ran out, is there any way I can increase its size with swap? 2017-06-24 02:34:44 swap is literally the opposite of tmpfs 2017-06-24 02:35:14 One sec. 2017-06-24 02:37:28 budric[m], try: mount -o remount,size=NEWSIZE / 2017-06-24 02:46:14 Maxpm: thank you! 2017-06-24 08:35:56 ACTION waves 2017-06-24 08:37:39 is there a way to tell alpine-virt-3.6.2-x86_64.iso to automatically use DHCP while launching through qemu? 2017-06-24 09:04:05 <^ingo^^^> is there an easy way to set/manipulate the size of the swap-partition during the setup-alpine process? I do a setup-alpine with sys disksetup. And if i give the VM 4 GB memory he create the same swap-size. This is in this case a waste of resources in my case. I would like to use absolut mimimal(512MB maybe ) swap, or zero swap. 2017-06-24 09:22:54 ^ingo^^^: you could use setup-disk and specify swapsize with -s 2017-06-24 09:23:02 instead of using setup-alpine 2017-06-24 09:48:44 <^ingo^^^> kahiru: so first run setup-disk and than setup-alpine? 2017-06-24 09:49:52 ^ingo^^^: setup-alpine runs a bunch of setup-* scripts. so you should probably run things like setup-network, setup-apkrepos and so on and setup-disk last 2017-06-24 09:50:04 no need for setup-alpine afterwards 2017-06-24 09:50:05 ? 2017-06-24 09:57:49 <^ingo^^^> hm, not the easy way. ok, give it a try. 2017-06-24 10:54:11 <^ingo^^^> ok found better solution before call setup-alpine make an "export SWAP_SIZE=0" 2017-06-24 12:25:04 Is https://en.wikibooks.org/wiki/Grsecurity still considered relevant with regard to getting up to speed with grsec-enabled distro administration ? Anything else I should read for alpine-specific security best practices ? With context of setting up server apps (mail, ...) 2017-06-24 12:25:04 lol 2017-06-24 12:39:28 yes 2017-06-24 12:39:42 although we just package gradm so you shouldn't need to compile it 2017-06-24 12:47:51 Fantastic. Thanks 2017-06-24 15:42:17 hello is there an alternative for hwinfo in alpine? I am trying to find out the available video modes for my tty without having x, so no xrandr 2017-06-24 15:49:04 how about /sys/class/graphics/fb0/modes ? 2017-06-24 15:57:54 "/sys/class/graphics" is empty 2017-06-24 15:58:13 should I have installed alpine-desktop or something along those lines? 2017-06-24 15:58:41 I'm on my first alpine virtualbox install and want to change my tty size 2017-06-24 16:00:25 have you tried just passing vga=792 or something as kernel commandline? 2017-06-24 16:01:46 or if you use grub with GRUB_GFXMODE=auto or something 2017-06-24 16:01:51 maybe. 2017-06-24 16:03:29 i think I am using syslinux 2017-06-24 16:03:45 if you switch to grub you can play with GRUB_GFXMODE at least 2017-06-24 16:03:47 but i can't seem to find syslinux.cfg 2017-06-24 16:03:57 it's called extlinux.cfg probably 2017-06-24 16:04:03 .conf 2017-06-24 16:04:30 extlinux.conf 2017-06-24 16:04:32 found it 2017-06-24 16:05:47 i got an undefined video mode number 2017-06-24 16:05:58 but it gave me the option to list the available ones 2017-06-24 16:06:03 was that the goal? 2017-06-24 16:06:06 :D 2017-06-24 16:08:04 oh 792 was 1024x768 2017-06-24 16:22:39 whatever works, right ;P 2017-06-24 16:23:46 i wonder maybe i don't get any video modes because it might be something not quite right with the guest additions? 2017-06-24 16:25:23 i installed the virtualbox-additions-grsec package and added vboxpci,vboxdrv, and vboxnetflt to /etc/modules 2017-06-24 16:25:30 https://www.virtualbox.org/manual/ch09.html#idm7674 2017-06-24 16:25:33 but on another wiki page 2017-06-24 16:27:04 hmm so basically i have to add the video modes to the virtual machine 2017-06-24 16:27:09 never did this before 2017-06-24 16:27:18 oh well 2017-06-24 16:28:39 also https://superuser.com/questions/31669/change-resolution-from-command-line-in-virtual-box 2017-06-24 16:29:48 and if all else fails, you can always just SSH into the system instead 2017-06-24 16:33:04 i think i will just go back to ssh for the time being 2017-06-24 16:33:57 thank you for taking the time :) 2017-06-24 16:40:54 Hi ! is it possible to install 32bit libssl on a 64bit alpine install ? 2017-06-24 16:43:09 not really, the packages don't have an arch-specific path for the libs 2017-06-24 16:43:23 I'm guessing a 32-bit chroot would be the easiest 2017-06-24 16:45:45 e.g. apk --update-cache --arch x86 --root /mychroot --repository http://nl.alpinelinux.org/alpine/v3.6/main --initdb add alpine-base 2017-06-24 16:48:31 hmm i need only libssl1.0-dev:i386 for a programm 2017-06-24 16:49:09 sure, and its dependencies, and so on 2017-06-24 16:49:31 in arch : pacman -S lib32-openssl lib32-gcc-libs 2017-06-24 16:49:45 alpine doesn't work that way 2017-06-24 16:53:36 ok with the line above i install alpine-base in this /mychroot directory right ? 2017-06-24 16:54:32 indeed 2017-06-24 16:56:05 ok and then i can link to the 32bit libraries right ? 2017-06-24 16:57:55 that's an option, e.g. LD_LIBRARY_PATH, or chroot /mychroot /bin/ash and run your program from the 2017-06-24 16:58:09 from there* 2017-06-24 17:00:52 (the chroot is more work I guess, to get networking to work and so on; there's documentation on the wiki if you wanna go that route) 2017-06-24 17:04:12 ERROR: http://nl.alpinelinux.org/alpine/edge/main: UNTRUSTED signature 2017-06-24 17:04:59 that error i get if i paste that line ... 2017-06-24 17:06:00 ok i dont if its better to install a new lxc container with 32bit 2017-06-24 17:29:10 https://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot#Install_the_alpine_base_installation_onto_the_chroot fyi 2017-06-24 17:42:45 I based my commandline on a situationw here I had already installed the keys prior, you can just add --allow-untrusted to work around it, also you can add "libssl1.0" after alpine-base for a complete one-liner 2017-06-24 17:45:11 hey all, so I'm trying to add my user to a group which is allowed sudo, but 'usermod' isn't there. Any other commands to do this? 2017-06-24 17:45:58 sudo adduser user group 2017-06-24 17:46:30 ah... that was easy :). thanks a bunch 2017-06-24 17:48:02 guys, I'm trying to figure out boot via grub on Alpine. Any proper howto on this? 2017-06-24 17:48:47 howto on wiki useless 2017-06-24 17:50:07 check out https://wiki.alpinelinux.org/wiki/Create_A_VirtualBox_Guest_with_Grub_and_XFS -- it's got a lot of unrelated stuff, but the basic steps are there 2017-06-24 17:50:21 I installed grub2 package but couldn't figure out build a proper grub.cfg 2017-06-24 17:50:39 Bun: thanks..checking 2017-06-24 17:51:10 just be sure to replace the xfs stuff with e.g. ext2 2017-06-24 17:54:30 which howto did you use? 2017-06-24 17:54:56 i hae now installed 32bit alpine and now i get the same error 32bit support not available 2017-06-24 17:55:12 this drives me crazy 2017-06-24 17:55:33 i want to use gns3 with iou umages 2017-06-24 17:56:08 oh I dunno if the kernel lets you run 32 bit binaries 2017-06-24 17:56:16 that would complicate things 2017-06-24 17:56:32 it does, in a chroot. 2017-06-24 17:56:45 but multilib is not supported on alpine at this time 2017-06-24 17:57:59 the host system is arch 2017-06-24 17:59:02 oh, you were already in a container? 2017-06-24 17:59:02 32bit alpine linux should use the 32bit libraries right ? 2017-06-24 17:59:23 yes lxc container 2017-06-24 17:59:39 Bun: I'm not a newbie but I see that something off with grub2 on Alpine. I read this but nothing speacial https://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting 2017-06-24 18:00:57 Bun: except update-grub2 command doesn't exist which must be a "grub-mkconfig -o /boot/grub/grub/cfg" 2017-06-24 18:01:21 *grub.cfg 2017-06-24 18:01:27 yeah it seems a bit incomplete 2017-06-24 18:02:29 ACTION is going to test new grub.cfg 2017-06-24 18:14:06 so what software do you guys use to listen to music on alpine? 2017-06-24 18:16:57 turns out apk del syslinux; apk add grub grub-bios;grub-install /dev/vda;grub-mkconfig -o /boot/grub/grub.cfg is sufficient, wonder if that's worth a wiki page 2017-06-24 19:45:09 Bun: yes please, i had wondered about this in the past 2017-06-24 21:57:49 i'm having problems finding the package that provides the runuser command. anyone know how to find it? 2017-06-24 21:58:17 https://pkgs.alpinelinux.org/contents?file=runuser&path=&name=&branch=&repo=&arch= 2017-06-24 21:58:44 hrmm, so none provide it. i kind of expected it to be in util-linux 2017-06-24 21:58:48 wonder why it's not 2017-06-24 22:00:42 is it a PAM thing? 2017-06-24 22:00:53 no 2017-06-24 22:01:08 it's a smaller command line tool for doing sudo like things, just just for changing effective UID 2017-06-24 22:01:59 https://git.kernel.org/pub/scm/utils/util-linux/util-linux.git/tree/login-utils/runuser.c so does it do anything su doesn't? ;p 2017-06-24 22:02:18 the "effective" part, I guess 2017-06-24 22:02:41 maybe poke ncopa about it 2017-06-24 22:02:44 i guess I can make my stuff work with su 2017-06-24 22:30:58 hiya guys 2017-06-24 22:31:36 just wanted to ask -- how does alpine-hardened differ from grsec? 2017-06-24 22:35:50 <_ikke_> jord: iirc, currently not much 2017-06-24 22:38:36 _ikke_, fair enough, I'm a gentoo hardened user (musl) myself but I've been eyeing alpine, good stuff! 2017-06-24 22:39:30 been pinching a few patches from you guys since a small bit of stuff doesn't compile on musl anyways, cheers 2017-06-24 22:41:49 Speaking of grsec, the newest thread in oss-sec is quite clash-iesque 2017-06-24 22:42:05 <_ikke_> aye 2017-06-24 22:42:20 As usual, it does not seem like anything good will happen out of it... politics... 2017-06-24 22:45:38 I'm currently setting up my home server services. Does it make any sense to spend time configuring grsec+pax on alpine instances, given recent event ? 2017-06-24 22:46:16 (and with future updates in mind) 2017-06-24 23:41:34 hey guys, successfully installed Alp, now configured xfce4, now for fine tunning xfce4 & installing my fav apps. 2017-06-24 23:42:08 In pkgs is there xfce4-goodies & tor-browser 2017-06-24 23:42:33 I've looked in and did not locate 2017-06-24 23:43:00 or they in git & have not configured yet 2017-06-24 23:45:05 jord: basically for now it is just which is rebased :) 2017-06-24 23:45:25 we are observing the situation and may do something else in future but at the same time maybe not 2017-06-24 23:56:44 fryguy: considered su-exec? 2017-06-24 23:59:56 Does a kernel need to be aware of the medium it's being booted from? e.g. what happens if I compile a kernel without USB support, put it on a USB flash drive and use the BIOS menu to boot it? 2017-06-25 00:01:16 yes, it does 2017-06-25 00:01:20 or well 2017-06-25 00:01:29 depends if you need it to use a rootfs post-initramfs 2017-06-25 00:01:46 for alpine's initramfs, it wouldn't necessarily no 2017-06-25 00:02:05 as the bootloader takes care of loading the initramfs into memory and passing it to the kernel appropriately 2017-06-25 00:02:14 and alpine's initramfs can just bootstrap the system into ram 2017-06-25 00:02:34 Right. 2017-06-25 00:04:35 so, it only needs to be aware of its boot medium if you also have a rootfs on the same medium 2017-06-25 00:04:37 :P 2017-06-25 00:11:21 kaniini, ah well, at least we have 4.9 LTS till 2019 :) 2017-06-25 00:11:33 Thanks for explaining, Shiz. 2017-06-25 00:12:08 nw 2017-06-25 00:12:37 jord: we can likely rebase to next kernel LTS if necessary 2017-06-25 00:12:44 it is a pain, but it is possible 2017-06-25 00:12:49 we have done it before 2017-06-25 00:15:17 i'm not confident that it would be fully functional, though 2017-06-25 00:58:56 I'm running alpine-extended-3.5.2-x86_64 via VMWare Player using a physical disk. I just setup alpine, and installed alpine-desktop with xfce4. Enabled `dbus`, and ran `startx`. Everything wen't fine, I even see the desktop, but I can't move the mouse or anything. Any guidance? 2017-06-25 01:01:49 i don't think simply installing alpine-desktop gives you the right xorg driver packages 2017-06-25 01:01:52 two things: 2017-06-25 01:02:02 1) 3.5.2 is outdated by now as 3.6.2 is out :P 2017-06-25 01:02:12 2) setup-xorg should install the needed xorg drivers 2017-06-25 01:12:27 Upgraded to 3.6.2, updated/upgraded packages. Ran `setup-xorg-base` and it's running mostly fine now (missing icons?). 2017-06-25 01:12:29 Thanks for the help! 2017-06-25 01:13:46 hmm 2017-06-25 01:16:13 i'm not quite sure on the icons, it ma need a package for that but i wouldn't know which :P 2017-06-25 01:17:38 are any specific icons missing or icons in general? 2017-06-25 01:20:59 Doesn't seem to be any real method to the madness. Some are there, some are not. https://a.safe.moe/Atffz.png 2017-06-25 01:24:12 hmm 2017-06-25 01:24:53 hicolor-icon-theme perhaps? 2017-06-25 01:26:25 and maybe gnome-icon-theme 2017-06-25 01:36:05 Yup, that was it. Chagned icons and they're all there now. Thanks for your help! 2017-06-25 07:53:33 hi ! i want to install telnet on alpine linux but i can find which package should i install ? 2017-06-25 08:17:40 in arch is telnet included in inetutils 2017-06-25 08:18:20 but on alpine there is not a inetutils and no telnet ? 2017-06-25 08:18:52 what do you need telnet for? use nc if you want a simple socket client 2017-06-25 08:23:44 Check BusyBox extras 2017-06-25 08:29:59 I'm not sure i got a response earlier, trying my luck again: 2017-06-25 08:30:02 I'm currently setting up my home server services. Does it make any sense to spend time configuring grsec+pax on alpine instances, given recent event ? 2017-06-25 08:30:06 I'm not sure i got a response earlier, trying my luck again: 2017-06-25 08:30:29 (and with future updates in mind) 2017-06-25 08:34:51 Hello, I just upgraded my edge alpine, and rxvt-unicode does not work since then :/ 2017-06-25 08:34:59 Should I open a bug report? 2017-06-25 08:35:03 Same for i3status 2017-06-25 08:35:18 Youbi: sounds like you need to look for more info. 2017-06-25 08:35:42 jushur: well, rxvt-unicode tells me that it misses the good version of a perl lib. 2017-06-25 08:35:53 netcat is nice but i see only the output and cannot type on remote device 2017-06-25 08:36:14 And i3status yells a strange error while parsing the config (although I didn’t touch the config before updating) 2017-06-25 08:54:12 Got disconnected, in case anyone replied 2017-06-25 09:46:51 Hi! im still struggling with installing kodi on rpi3 on alpine linux. at least i got it compiled now. but playback doesnt work. it complains about "mmal: mmal_component_create_core: could not find component 'vc.ril.video_decode'". i could really use some help here ;-) 2017-06-25 10:26:10 Youbi: ah looks like it needs a rebuild then 2017-06-25 10:26:33 AlexIncogito: yes 2017-06-25 10:29:02 Shiz: Does it imply any of the following ? a) Alpine will somehow foreward port grsec ? b) It makes sense to stick with an older grsec-enabled version when new updates arrive c) Something else ? 2017-06-25 10:37:55 Shiz: i3status or rxvt-unicode? 2017-06-25 10:54:44 hi, is alpine xen version need to run as dom0 and the virtual verison need to run as a vm? 2017-06-25 10:58:49 I created https://wiki.alpinelinux.org/wiki/Bootloaders since some people were wondering about it 2017-06-25 11:00:36 Bun: "install yslinux manually" 2017-06-25 11:00:49 +s 2017-06-25 11:00:51 typing is hard 2017-06-25 11:00:57 :) 2017-06-25 12:19:54 Youbi: rxvt-unicode 2017-06-25 12:50:23 Rollos: I believe you are correct. It comes with xen files in boot/. Also, tutorials on running alpine as domU refer to virtual alpine iso version. I would venture and say xen alpine iso version isn't necessary per say, as you can use apk to install "xen" and "xen-hypervisor" packages. The xen iso would be a convenience package I assume 2017-06-25 12:51:46 But this is mostly speculation, you should wait for confirmation from the local authorities, or another reliable source 2017-06-25 12:52:26 Crap, he left... 2017-06-25 12:52:31 Hi, is any wiki administrator online? 2017-06-25 12:53:11 I think shiz is off and on IRC, probably busy and answering quick or critical questions 2017-06-25 12:53:23 Perhaps you should just state your question 2017-06-25 12:55:08 well, I have a copule... first: I try to add an extra line to one of the wikipage, but wiki waid, it was restricted, since I'm new (that's true btw) and what I tried to do is harmful, suggested to contact admins or sg like that. just... there were no link, where can I find them. 2017-06-25 12:55:59 You're at the right place, you can also try the forums, but it is probably better here 2017-06-25 12:56:07 Although forum leave a track for future reference 2017-06-25 12:57:47 second: can they extend that warning message with some hint, where and whom to look for? another direct wikipage would be enough. I had to dig around a tons of page and a tons of links... (ok, maybe not, but.. if someone is new, and try to help, maybe he easily become frustrated, if you ask extra careness, but doesn't give extra info what to do) 2017-06-25 12:57:49 (Or stackoverflow for thaat matter) 2017-06-25 12:58:45 deferring that one to clandmeter since i don't have wiki admin access 2017-06-25 12:58:46 so, the closest what I could found was the contrib page, and I thought the irc is the fastest way to find sy 2017-06-25 13:00:27 @Shiz: When you've got time, would be a huge help if I had a better grasp of what to expect in regard to grsec & alpine -- see my question about 2hours ago. Thanks a lot ! 2017-06-25 13:00:29 AlexIncogito, I found the solution for my problem. I just wanted to help, to extend the wikipage with that 2017-06-25 13:00:53 AlexIncogito: the 3.6.x branch will still use grsec and is supported for two years 2017-06-25 13:00:54 Shiz: clandmeter? is that sy's nick? 2017-06-25 13:01:18 I'm not sure who sy is supposed to be, but it's not clandmeter 2017-06-25 13:01:19 :P 2017-06-25 13:01:45 @Shiz: Excellent news, thanks ! 2017-06-25 13:01:52 sy == somebody 2017-06-25 13:01:53 AlexIncogito: the long and short of it is that we're still considering our options for 3.7, but unless something changes I personally consider it unlikely we'll be using grsec in 3.7 2017-06-25 13:02:10 yes, clandmeter is a person 2017-06-25 13:03:12 Yeah given the direction the grsec situation is taking, and recent (today actually) official positions, it seem like there is very few hope for public use. And integration in future version would probably be painful and divert a lot of your time 2017-06-25 13:03:37 time is not even the major problem, it's that forward-porting grsec to new kernel major versions is extremely hard to get right 2017-06-25 13:03:46 because of all the areas in the kernel it touches 2017-06-25 13:04:03 I don't think we'd be able to pull it off 100% right 2017-06-25 13:04:23 True, actually spengler make an excellent job of pointing the errors in all mainstreaming efforts... heh 2017-06-25 13:04:35 spender* 2017-06-26 01:24:48 <_spOOn_> hmm, just downloaded a new iso, and was going through the install... 2017-06-26 01:24:54 <_spOOn_> If I select auto-detect mirror 2017-06-26 01:25:02 <_spOOn_> I get a connection reset from all hosts 2017-06-26 01:25:13 <_spOOn_> If I select a mirror, it works fine... 2017-06-26 01:25:21 <_spOOn_> strange... 2017-06-26 09:31:29 Hello, do we have any tools for kernel management? 2017-06-26 09:31:49 Something along the lines gentoo has 2017-06-26 09:41:29 Hi. What's the usual way to check which users are logged into a system, given that musl doesn't support utmp/lastlog? 2017-06-26 10:32:18 Invader_Bork: ? what tools would we need 2017-06-26 10:32:25 pl: you can check /dev/pts 2017-06-26 10:32:31 ownership 2017-06-26 11:02:46 something like eselect kernel list 2017-06-26 11:03:14 Shiz: something to list the available kernels, set the default one, delete the ones i want to delete to 2017-06-26 11:03:30 slots would be great, especially because updating the kernel deletes the running kernel's /lib/modules 2017-06-26 11:27:14 hi, i get an error if i want to install alpine with setup-alpine. During partition erase i get the following error: Failed to add partition: Invalid argument 2017-06-26 11:29:12 find no solution in the web.. 2017-06-26 11:30:10 Bun: I got make grub menu appears with copy-paste grub.cfg fom link. 2017-06-26 11:30:40 terra: https://wiki.alpinelinux.org/wiki/Bootloaders I tested it myself, this worked for me 2017-06-26 11:31:34 Janneck: https://forum.alpinelinux.org/forum/installation/setup-alpine-failed-add-partition-invalid-argument and you saw this? 2017-06-26 11:33:33 Bun: thank you i have not seen, will try it =) 2017-06-26 11:33:38 grub-mkconfig -o /boot/grub/grub.cfg didn't work for me. Btw, initramfs can't mount ext4 root partition beause initramfs doesn't autoload ext4 module even if ext4 module included.. this is weird 2017-06-26 11:36:58 initramfs puts me emergency shell after boot failure. I can mount ext4 root partition manually after loading ext4 module 2017-06-26 11:39:53 initramfs generated by 'mkinitcpio' from Arch was working flawless but I have no access to Arch right now. 2017-06-26 11:40:58 btw I'm using Liquorix distribution kernel 2017-06-26 11:44:16 Thanks Bun it works, but i have a second issue.. my internet connection don't work during the installation and i dont know why.. 2017-06-26 11:47:29 terra: can you maybe share the grub.cfg that was generated? if you're using "other" kernels then I guess a bit of custom work is expected anyway 2017-06-26 11:47:52 Janneck: which part doesn't work? is it wired, wireless? 2017-06-26 11:49:08 Bun: maybe.. probably Alpine looks for vmlinuz and initramfs like simplistic file names for generating menu entries. 2017-06-26 11:50:29 Bun: wired... initialize: eth0, manual configuration: no, DNS? Why DNS? i need DHCP... i only know my gateway... 2017-06-26 11:50:30 I'm guessing it doesn't generate a useful kernel cmdline either 2017-06-26 11:51:30 Bun: my file names were differ but I tried to add menu entries manually but it gave me parsing error then I just ended up with simplistic copy-paste grub.cfg 2017-06-26 11:52:24 Now, I have a different problem regarding initramfs and ext4 root partition mentioned above 2017-06-26 11:52:41 is it the alpine initramfs? 2017-06-26 11:52:47 yes 2017-06-26 11:52:56 and you pass modules=ext4 as cmdline? 2017-06-26 11:53:05 I mean generated bt Alpine mkinitfs 2017-06-26 11:53:36 e.g. for an alpine kernel: linux /vmlinuz-virthardened root=/dev/vda3 ro modules=sd-mod,usb-storage,ext4 quiet rootfstype=ext4 2017-06-26 11:53:58 of course..As I noticed earlier, ext4 modules included in initramfs 2017-06-26 11:54:13 Janneck: what is it asking about DNS exactly? it might just be asking for the DNS "search domain", which you can ignore 2017-06-26 11:54:35 oh no auto probe? then it is my fault. 2017-06-26 11:54:44 apparently not 2017-06-26 11:55:18 Bun: thank you for commenting out 2017-06-26 11:56:05 Janneck: in fact, if you just answer all networking related questions with their defaults it should work in your case 2017-06-26 11:56:18 Bun: during the installation i can inizialize eth0, afther that the setup aks me for a dns nameserver and thats all... but i have no access to the alpine server... 2017-06-26 11:57:04 and if i manually config /etc/network/interfaces with auto eth0, iface eth0 inet dhcp it dosn't work again... 2017-06-26 11:57:45 well that should work, you restarted the networking service or rebooted, of course? 2017-06-26 11:58:03 you'll want to inspect /var/log/messages and/or dmesg for any warnings 2017-06-26 11:58:16 no reboot because i booted from a usb stick... 2017-06-26 11:59:35 ACTION is going to try new grub.cfg and initramfs 2017-06-26 12:00:12 so service networking restart or somesuch 2017-06-26 12:02:18 uname -a: Linux alpine 4.11.0-6.1-liquorix-amd64 #1 ZEN SMP PREEMPT liquorix 4.11-13 (2017-06-21) x86_64 GNU/Linux :) 2017-06-26 12:03:48 a desktop kernel, eh? feeling adventurous are you 2017-06-26 12:04:50 Bun: you're right as well 2017-06-26 12:07:42 ah Bun now it works thanks =) 2017-06-26 12:15:26 hello 2017-06-26 12:15:48 I have a problem since few days with the alpine setup on rapsberry pi 2017-06-26 12:16:23 I do a setup-disk and now it won't boot. I think it's related to the upgrade of apk-tools. Do you have some report about that ? 2017-06-26 12:17:12 it's for a alpine 3.5 setup on a rpi 3, only with setup-disk mode after the reboot, I have a black screen, the configuration files are right and the partitions clean 2017-06-26 12:28:27 boris: have you acces on serial tty console? 2017-06-26 12:29:19 no no cable here ... i have already checked 2017-06-26 12:35:20 boris can you try latest 3.6 2017-06-26 12:35:56 and setup-disk is afaik not supported on rpi 2017-06-26 12:36:33 I installed alpine on a rpi a few days ago 2017-06-26 12:36:58 following https://wiki.alpinelinux.org/wiki/Raspberry_Pi and in particular https://wiki.alpinelinux.org/wiki/Raspberry_Pi#Traditional_disk-based_.28sys.29_installation 2017-06-26 12:37:29 it worked fine 2017-06-26 12:38:25 yes that should work fine 2017-06-26 12:38:53 yeah but it works since 2/3 days ... 2017-06-26 12:39:48 i think something is wrong on the latest upgrade of apk-tools 2017-06-26 12:40:02 I don't talk about 3.6 but 3.5 should works 2017-06-26 12:46:30 trying the 3.6 now with my provisioning script 2017-06-26 12:46:49 hopefully I have a backup of a 3.5.2 sdcard which boots 2017-06-26 12:47:10 so I can diff the problem on the new apk tools scripts 2017-06-26 12:48:33 ok same problem black screen with 3.6 at the boot 2017-06-26 12:48:44 but the green LED blinks 2017-06-26 13:26:36 /msg NickServ identify whowhatwhere1! 2017-06-26 13:28:57 setup-keymap changes keymap not permanently. how can i set this permanently? 2017-06-26 13:30:02 lbu commit ? 2017-06-26 13:33:56 Invader_Bork: there's not really any kernel selection to be had in alpine, so i'm a bit confused by your question 2017-06-26 13:34:02 except for flavour 2017-06-26 13:35:40 flovor as in vanilla or hardened ? 2017-06-26 13:36:08 Shiz: I was thinking of that and different versions, like having current and lts 2017-06-26 13:36:22 or custom 2017-06-26 13:36:32 and activate which one 2017-06-26 13:36:40 but I guess I can do that via the bootloader 2017-06-26 13:37:19 we don't ship different kernel versions 2017-06-26 13:37:36 oh I see 2017-06-26 13:37:59 what are you using? lts 2017-06-26 13:38:00 ? 2017-06-26 13:38:14 or arbitrary versions that are deemed stable 2017-06-26 13:38:18 ? 2017-06-26 13:44:27 Invader_Bork: looks like LTS 2017-06-26 13:47:30 Invader_Bork: latest lts at the time the release gets cut, afaik 2017-06-26 14:10:32 anyone got a good tutorial on writing packages, whether for arch or alpine? 2017-06-26 14:10:44 I'm finding the wiki good for reference, but unclear at times 2017-06-26 14:13:33 Hi, anyone used libsodium in alpine PHP images? 2017-06-26 14:16:20 please can anyone help me? how can i change keyboard layout permanently in console? 2017-06-26 14:21:57 Janneck: /etc/conf.d/keymaps and loadkeys(1) is your friend 2017-06-26 14:23:11 yes i add keymaps="ch" in /etc/conf.d/keymaps and reboot but no changes... 2017-06-26 14:25:07 mkinitfs regenerates your initrd, takes the keymap via /etc/mkinitfs/features.d/keymap.files 2017-06-26 14:32:27 lxGzx53q034r thanks but dont work... in keymap.files i add /etc/keymap/ch-de_nodeadkeys.bmap.gz 2017-06-26 14:39:12 It can not be that it is so difficult to change the keyboard layout. I've never had such problems in a distribution. ;-) 2017-06-26 14:49:28 less /etc/mkinitfs/features.d/keymap.files 2017-06-26 14:49:36 eh. wrong window 2017-06-26 14:50:08 Janneck: does /etc/keymap/ch-de_nodeadkeys.bmap.gz exist? 2017-06-26 14:50:19 and did you regenerate your initrd? 2017-06-26 14:51:03 ls 2017-06-26 14:51:15 lxGzx53q034r i can not find any .bmap.gz on the system... 2017-06-26 14:52:17 but after i run setup-keymap it add ch-de_nodeadkeys.bmap.gz in the /etc/keymaps and the kayboard layout works perfectly but after reboot its the old layout... 2017-06-26 14:53:06 you need to install kbd-misc 2017-06-26 14:53:29 oh. maybe not. hmmm. 2017-06-26 14:53:44 theres no bmap.gz only map.gz-s 2017-06-26 14:56:15 i reached the limits of my knowledge, i have no clue where the referenced path in setup-keymap /usr/share/bkeymaps should come from, and why it even works for you 2017-06-26 14:56:49 maybe you sh -x setup-keymap and figure out where the keymap comes from, and copy it from there to you regenerate your initrd? 2017-06-26 14:57:27 s/ to you / to your /etc and / 2017-06-26 15:08:37 dont work.. 2017-06-26 15:08:50 thanks for you help 2017-06-26 16:08:40 it seems that v3.5/3.6 is completely broken for rpi in disk mode. Good news to revert to raspbian 2017-06-26 16:09:36 no kernel log on serial ... 2017-06-26 16:30:11 musl (1.1.15-r6 -> 1.1.15-r7) & linux-rpi2 (4.4.52-r0 -> 4.4.52-r1) 2017-06-26 16:30:19 so it's the last kernel for rpi 2017-06-26 16:30:27 it breaks :p 2017-06-26 16:30:39 nothing related to apk-tools 2017-06-26 20:09:17 hey, i'm currently using alpine as docker base image (alpine:latest). I do a quick apk upgrade --no-cache && apk add --no-cache boost. Boost installs i got no error message but my program cannot include /usr/include/boost/date_time/posix_time/posix_time_types.hpp (No such file or directory) 2017-06-26 20:09:54 Anyone has expierenc with this or maybe there is a known issue? 2017-06-26 20:11:54 https://pkgs.alpinelinux.org/contents?file=posix_time_types.hpp&path=&name=&branch=v3.6&repo=&arch= 2017-06-26 20:12:29 oh, thank you a lot! 2017-06-26 20:49:26 configure: error: C compiler cannot create executables -> ideas? 2017-06-26 21:06:43 config.log 2017-06-26 22:54:18 hey, first time getting started, booted iso from usb stick, now its asking for login? 2017-06-26 22:54:24 ok i'm in with "root" 2017-06-26 23:02:17 after setup 2017-06-26 23:02:41 i do "setup-xorg-base" to install the windowing system? 2017-06-26 23:09:50 then howto xinit? 2017-06-26 23:36:51 ok got it, apk install xfce4 lxdm and start via rc-service thnx, looks clean and fast, nice work all :) 2017-06-27 00:23:38 alpine allows PVH mode in xen, correct ? 2017-06-27 00:28:17 or more semantically correct "as a xen guest" 2017-06-27 00:28:29 (dom0 or domU) 2017-06-27 00:34:31 AlexIncogito: should, ask royger when he's around 2017-06-27 00:39:00 darkfaded: Will do, thank you. Meantime, if anyone has experimented wih this feel free to jump in 2017-06-27 01:59:34 i can't seem to understand mdev/devtmpfs behaviour during the initrd stage of boot. i have 2 partitions, one is /boot and the other is luks encrypted, and of course i need to decrypt the root partition during boot. i can do that with the cryptroot option fine, but i want to have the ability to decrypt over ssh 2017-06-27 02:00:20 i tried using https://github.com/mk-f/alpine-initramfs-dropbear and i tried writing an initrd from scratch, and i'm hitting similar issues on both 2017-06-27 02:02:16 both attempts i can't see /dev/sda* being created automatically, and i need to manually mknod these devices in order to use them. i can of course manually create these in the init script, but why shouldn't these appear automatically? why do they work fine with cryptroot but not when i introduce dropbear into the mix? 2017-06-27 05:40:44 Is there anything like FAI (http://fai-project.org/fai-guide/) for Alpine? 2017-06-27 07:26:29 royger: Hello. I understand you are an authority on the subject of xen. I was inquiring earlier whether pvh mode was supported by alpine guests in xen, and was hinted toward asking you 2017-06-27 07:30:39 AlexIncogito: hm, you will need Xen >= 4.8 and Linux 4.11 (IIRC) in order to use PVH. 2017-06-27 07:31:50 Ah, no good then if that's the case, as Alpine 3.6.2 is on kernel 4.9.32 2017-06-27 07:32:13 It's simply a matter of adding pvh=1 in xen config according to xen wiki, correct ? 2017-06-27 07:35:02 AlexIncogito: for various reasons ATM the only way to create a PVH guest (provided that you have the right version) is by adding "builder=hvm; device_model_version=none" to your config file. 2017-06-27 07:35:37 but yes, pvh=1 should be the right way to do it, it's just not working with current versions 2017-06-27 07:36:48 That's strange, isn't pvh supposed to be PV ? I was under the impression that build=hvm ran guests in HVM mode 2017-06-27 07:37:46 AlexIncogito: no, PVH is basically a HVM guest without a QEMU instance 2017-06-27 07:38:46 search for PVHv2 2017-06-27 07:38:50 http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.11-Xen-Changes you were correct, it introduces PVHlite 2017-06-27 07:39:50 I was basing my understanding on this : http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.11-Xen-Changes 2017-06-27 07:39:54 Perhaps this changed in PVHv2 2017-06-27 07:40:02 Or typo in wiki 2017-06-27 07:40:40 PVHv1 was more like a PV guest, PVHv2 is certainly a HVM guest. 2017-06-27 07:40:54 Okay, makes sense now 2017-06-27 07:41:12 PVHv1 was basically a PV guests running with hardware assisted paging, PVHv2 is basically a HVM guest with less emulated devices 2017-06-27 07:41:31 yes, the wiki needs a very good cleanup 2017-06-27 07:41:39 I shall find time to do that 2017-06-27 07:42:54 Do you believe something like this would make sense: setup latest version of alpine guests in HVM with PVHVM drivers, and switch to pvh mode when it is upgraded to 4.11+ kernel ? Is that possible at all ? 2017-06-27 07:43:19 AlexIncogito: yes, that would be my recomendation 2017-06-27 07:44:01 Fantastic, I have a better understanding now. That was very helpful thank you sir 2017-06-27 07:44:04 performance wise PVHv2 should be equal to PVHVM, is just that you no longer need a QEMU instance running in Dom0, and the surface of attack is smaller 2017-06-27 09:07:22 anyone know of a way to unconditionally forward packets from one interface to another? so far I've come up with eBPF, PF_RING, iptables, and maybe PF_PACKET if the interface has forwarding disabled 2017-06-27 09:08:07 nftables? :P 2017-06-27 09:08:29 and ebtables 2017-06-27 09:11:19 I'm just kind of surprised there is no way to specify a source device based route 2017-06-27 09:59:22 Bun: One reason might be that routing in general works on dest ips, so decisions not based on that will break some assumptions and perhaps validation steps done elsewhere in the ip stack. I do see the idea at times, but that could be an explanation to why its hard to find examples and simple implementations 2017-06-27 10:01:40 I understand, but everything should be software defined networking these days :^) 2017-06-27 10:29:11 Hi. Let's say that I want to run Alpine in diskless mode, and I 'dd' the iso to an usb stick. Now I get a partition that is smaller than the usb stick, of course. I'd like to expand it to cover all the usb stick memory, in order to have more space for lbu commits, apk cache, and so on. Is this process documented somewhere? 2017-06-27 10:33:31 just make a regular install to the usb stick 2017-06-27 10:34:29 actually, no 2017-06-27 10:35:12 TBB, well, but let's say I'm fine with diskless mode. Mostly because I'm using a random usb stick, and I want to limit writes 2017-06-27 10:35:52 https://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB it's discussed here, no? with setup-bootable 2017-06-27 10:37:13 Bun, so setup-bootable does grow the partition/filesystem up the device size? 2017-06-27 10:37:51 (and, I don't have another Alpine system to start from, but that's another issue) 2017-06-27 10:38:06 I think you just make a full size partition and then copy the necessary stuff 2017-06-27 10:38:18 I don't like that method at all, dd'ing the cdrom to the stick already creates a partition table 2017-06-27 10:38:30 oh? in that case why not create a second partition 2017-06-27 10:39:29 I'd personally partition manually, install from chroot and then use a separate data partition 2017-06-27 10:39:32 also, if you make both lvm, you could easily ask the dd'd lvm to stretch into the newly created lvm part 2017-06-27 10:40:10 but that said, I don't do installs for use with lbu 2017-06-27 10:40:49 you don't need another alpine system anyway, if the USB ISO thing runs fully in memory; you could reformat the USB drive then (of course be wary of breaking it) 2017-06-27 10:41:42 I'll do a few tests, thanks 2017-06-27 14:12:24 since 'echo -e $(cat )' eats all newlines, is there a workaround? 2017-06-27 14:15:10 oh nvm. forgot to put " around $(). 2017-06-27 14:44:19 Any idea on why the crda package is still in "edge"? https://pkgs.alpinelinux.org/packages?name=crda&branch=&repo=&arch=&maintainer= 2017-06-27 14:44:40 it's needed in order to make an access point with alpine 2017-06-27 17:11:31 bah, one of those days again, can't seem to install because my brand new root partition "doesn't exist or access denied" after partitioning with sgdisk 2017-06-27 17:27:28 and no amount of syncing and partprobing changes that. ah well; I wanted an excuse for leaving work early today, seems I got one :D 2017-06-27 20:37:02 latest firefox in edge fails to load pages for me and has segfaulting subprocesses 2017-06-27 20:37:07 MOZ_DISABLE_CONTENT_SANDBOX=1 fixes it 2017-06-27 20:37:58 it seems something (seccomp filter?) makes musl's clone syscall fail and then an mThread pointer in some class remains uninitialized which causes nullptr derefs 2017-06-27 20:38:40 (happens in child process so the main gui seems to work, and in some cases loaded pages for me despite child processes are crashing in the background) 2017-06-28 07:40:02 hey guys still configuring my DE xfce4 is there a pkg like xfce4-goodies 2017-06-28 07:40:40 mystified: is there something specific you're looking for? 2017-06-28 07:41:55 no.. It's just xfce have a pkg called goodies. I installed the whisker plugin, it's usally in goodies or extras 2017-06-28 07:42:10 btw, thx kahiru: 2017-06-28 07:42:45 what about sandboxed-tor-browser 2017-06-28 07:42:59 or tor-browser inthe binaries 2017-06-28 07:49:21 where is Xorg.conf 2017-06-28 09:13:52 I'm having trouble mounting an nfs disk in 3.5.2; nfsv4 appears in lsmod however this fails: mount -t nfs -o nfsvers=4 192.168.1.15:/vol /mnt/vol 2017-06-28 09:14:27 the same call works on a centos7 box but on alpine the syscall returns EPROTNOTSUPPORT 2017-06-28 09:15:03 EPROTONOSUPPORT* 2017-06-28 09:17:29 the syscall for reference: mount("192.168.1.15:/vol", "/mnt/vol", "nfs", 0, "vers=4,addr=192.168.1.15,clientaddr=192.168.1.7") 2017-06-28 09:17:57 come to think of it, there's a chance the nfs server has a preference for pNFS 2017-06-28 09:26:07 fuck you kaniin 2017-06-28 09:26:09 fuck you kaniini 2017-06-28 09:26:11 MH0815 dfs hrmlgon-z Ganwell sigmoidbigfoot ahrs sparklyballs hiro kunev bghost_ nightmared afics xsteadfastx blahdodo fireglow nidan_ mortis304 silviof leprechau greguu fcolista jackmcbarn Tsutsukakushi sigjuice ageis duncan^ tdtrask Klowner mitchty TemptorSent Kooda[b] tg rafalcpp Unode systo 2017-06-28 09:26:11 k0nsl peterrus chris| robru reda Madgod skee mitemite tweek_ Kero vectr0n|cloud Janhouse smlb johansglock hrk jomat thum bfritz atomi nmeum lonix skjoldr tru_tru rymax99 dikaio so stateless book` codex xmux tw j0ni ScrumpyJack Kruge TBB highvoltage fragtastic shodan45 chamunks blueness XgF telling 2017-06-28 09:26:11 rsal ghavil alias___ koollman jvoisin nikolaos1 nbastin jpo feuerteufel|afk darkfaded eau bison Tritlo asonge algitbot miggyb TomJepp Graypup_ parazyd Antarez bougyman geekblogtv ericnoan AGiardini[m] triple Diftraku psychi[m] jnettlet leepa_ jyc_ mepholic EnTeQuAk leah2 ganlub mauli vectr0n 2017-06-28 09:26:11 hanez Amplificator ahills stevenroose frew ses1984 Wayward_One attractr_ mnemoc scv odc radhus jaustinpage_ firstlast arch3y James_T Metacity qrwteyrutiyoup mbentley Lord nsz Spydemon aemadrid moul tobias1 nepochal manacit eripa is_null esainane lxGzx53qO34r YoursTruly urzds __number5__ 2017-06-28 09:26:11 Adran Reeckz victorbjelkholm iamthemcmaster letoram serverascode Lloyd robtec_ snappy stwa Orochimarufan Keridos nshp Bun Elizabeth tugrik fish_ dhole[m] Kooda budric[m] itsfemme[m] copumpkin hl kaniini programmerq qman__ vifino Xe ThomasRuiz z1mme JStoker Peasant65 clandmeter ephemer0l doppo 2017-06-28 09:26:11 trfl dlewen nwmcsween Death_Syn pmcg_ mattaitchison khanku FergusL ng_ Nycatelos iron_houzi _spOOn_ zoidbergwill krainboltgreene kolev Fishfish0001 fhuntisch Somasis cydrobolt devth MuffinMedic hadret coredumb Bureaucat jlyo igitoor chapui_s dfgg madknight InAnimaTe vaelen plow ovf metalrain 2017-06-28 09:26:11 yo61 jailbox TML littleli grepory saintdev colstrom dbarrett nixdork thebadshepperd shansen trn flyinprogrammer ryanlelek Chloe wycats squaremo feepo _ikke_ Shiz DLange xming_ bontibon SirCmpwn marcelius_ mosez systmkor_ robogoat tmh1999_ john51_ usil_ mwbrown StarWarsFan BlackIkeEagle 2017-06-28 09:26:12 okeuday_bak IAMB3NW hackerhercules ddollar Marc1n przemoc NIN101 kahiru_ th__ agb norii_ Mr_H_ julius___ felixsanz_ lazybear ChrisRut andor2007 CcxWrk pickfire ncopa orbea IcePic pl gattuso nixfloyd saarg orbiter mmlb avih user632 oida czart__ xentec zopsi ncl sigtrm switchy dasher00 sagebind 2017-06-28 09:26:12 Invader_Bork minimalism epony cyteen ogny untoreh royger trebuchet_ t0mmy blackwind_123 gregman_ goberle 2017-06-28 09:26:15 fukcing niggers 2017-06-28 09:26:16 i hope 2017-06-28 09:26:17 you all die 2017-06-28 09:26:33 Thanks 2017-06-28 09:26:38 t-thanks I guess 2017-06-28 09:26:42 we probably will 2017-06-28 09:26:44 :) 2017-06-28 09:26:47 eventually 2017-06-28 09:26:54 fireglow: speak for yourself, I plan to live forever 2017-06-28 09:27:00 TomJepp ;) 2017-06-28 09:27:12 share your secret TomJepp 2017-06-28 09:27:22 I love you too 2017-06-28 09:27:23 I'll run my brain on alpine ;) 2017-06-28 09:27:41 wow someone is a bit bitter 2017-06-28 09:28:42 damn irc ransomware, messages are all encrypted... 2017-06-28 09:29:01 i can't wait to never die 2017-06-28 09:29:02 where can i send my money? 2017-06-28 09:29:12 to me 2017-06-28 09:29:15 clandmeter: camp 5, just ouside Pyongyang 2017-06-28 09:29:19 no, the nigerian prince sends you money 2017-06-28 09:29:23 you got it all wrong 2017-06-28 09:29:41 /dcc send Lord allyourmoney.btc 2017-06-28 09:30:22 why does alpine not have a big security alert that it is vulnerable to the latest randomsware outbreak? 2017-06-28 09:30:26 it has wine after all 2017-06-28 09:32:29 ! 2017-06-28 09:32:44 no wonder opal is mad 2017-06-28 09:33:26 and for platforms without wine, you could run alpine in qemu and wine on that and then ransomware. shock horror! 2017-06-28 09:33:32 it would be a strange combination of capable and insane to run SMBv1 under wine. 2017-06-28 09:38:07 what's wrong with this guy? 2017-06-28 09:38:16 is he a spammer 2017-06-28 09:38:19 or just snapped 2017-06-28 09:38:20 ? 2017-06-28 09:38:24 <_ikke_> no clue 2017-06-28 09:38:37 spammer for sure. 2017-06-28 09:40:28 running smb1, and being unpatched and THEN getting ransomware is very close to "paying the stupid tax". 2017-06-28 09:41:08 apart from this last one seemingly not giving you your files back even if you pay, from what the twitters claim. 2017-06-28 09:41:50 well, the email address you're supposed to contact was blocked right away 2017-06-28 09:42:01 so people pay first and then email, and then... well... 2017-06-28 09:50:06 fucking wat 2017-06-28 10:35:31 hi - i have a rtl8188eus wireless adapter - it seems that alpine's linux-hardened or linux-vanilla don't include the module for it even though it appears to be enabled in the config in the aports tree 2017-06-28 11:17:06 neakitten, which module is it? 2017-06-28 11:21:03 clandmeter: r8188eu 2017-06-28 11:21:29 seems its only enabled for rpi? 2017-06-28 11:21:46 ... hm 2017-06-28 11:21:46 /lib/modules/4.9.33-0-rpi2/kernel/drivers/staging/rtl8188eu/r8188eu.ko 2017-06-28 11:22:00 any reason why it wouldn't be enabled on x86_64? 2017-06-28 11:22:44 because its new? 2017-06-28 11:23:06 what's new? 2017-06-28 11:23:12 x86_64? or the module 2017-06-28 11:23:23 the module has been there since at least 3.x 2017-06-28 11:23:27 linux 3.x that is 2017-06-28 11:24:17 is my best option here just abuilding linux-hardened with it enabled? 2017-06-28 11:24:37 can you add a ticket to our tracker and assign it to ncopa? 2017-06-28 11:24:46 sure 2017-06-28 11:25:00 he will probably add it on next bump 2017-06-28 11:25:14 which will be end of this week i guess (if there is a new kernel) 2017-06-28 11:27:13 cool, i'll add a ticket now 2017-06-28 11:31:16 and done 2017-06-28 11:32:00 #7455 2017-06-28 11:33:13 in the meantime i'm just abuilding it so i can get a package that has it so i can continue setting up my new desktop c: 2017-06-28 15:30:36 Hey, what´s the advantage of chrony over busybox ? 2017-06-28 15:30:56 Why would I want to use chrony ? 2017-06-28 15:32:01 chrony can also be a server I believe 2017-06-28 15:32:35 but maybe bb can do that as well? 2017-06-28 22:58:39 BusyBox and Chrony aren't really the same thing 2017-06-29 00:06:37 mortis304 pretty sure the question was referring to bb's ntp implementation 2017-06-29 01:05:41 how hard is it to port apps over to alpine? 2017-06-29 01:12:18 m4chm4n: most of the time you have to just recompile 2017-06-29 01:14:42 Oh, excuse me. I am new to that aspect of things. 2017-06-29 01:15:06 I guess I have some lernin to do. 2017-06-29 02:54:59 I'm trying to install packages using apk during a Docker image build, but get permission errors on only *some* of the packages that are being installed: https://paste.gnome.org/prxdbb9j4/xxqwls 2017-06-29 02:55:10 Whether I include "USER root" or not 2017-06-29 03:31:32 Orbixx: curious 2017-06-29 03:31:36 tried a different mirror? 2017-06-29 03:31:47 anything in dmesg? 2017-06-29 03:50:33 hey guys, been racking my brain the past day on getting wkhtmltopdf loaded up on an Alpine install. I'm to the point where I'm about to use a Debian image just for this microservice, really breaking our devops workflow and organization. Any tips? Has anyone succeeded in getting it to work? 2017-06-29 03:50:48 I've tried many various ways and end up getting seg faults continually 2017-06-29 03:58:13 Looks like a pain in the butt to build. 2017-06-29 03:58:29 Yes 2017-06-29 03:58:52 I haven't actually seen or heard of anyone being successful, aside from a couple docker containers 2017-06-29 03:59:15 I assume those are working, but not sure why it'd work in Docker and not baked 2017-06-29 04:52:18 I'm useing context to stop all goroutines if in any of gorutines something bad is happend. (those gorotuines also comunicate with channels). The problem is that some of this channels could be closed when a goroutine try to write too channel. Here is the code https://play.golang.org/p/feauesDin4 . How can I solve this problem. 2017-06-29 04:52:37 ups 2017-06-29 04:52:40 wrong chanel 2017-06-29 04:52:42 sory 2017-06-29 05:28:12 Shiz: Ran it from a different host system and it worked without a problem, although I still don't understand the cause 2017-06-29 07:35:21 Hi, guys 2017-06-29 07:35:27 I have a question 2017-06-29 07:35:33 How to pronounce alpine? 2017-06-29 08:18:38 roxma: Like the English word. 2017-06-29 09:34:43 roxma: how to pronounce alpine: https://www.youtube.com/watch?v=eMbxtcO15xg 2017-06-29 12:06:10 hello there, i have a question 2017-06-29 12:06:25 how can i list contents of local .apk package? 2017-06-29 12:06:46 apk info -L is not working with .apk files directly 2017-06-29 12:06:55 tar -tf foo.apk 2017-06-29 12:08:22 many thanks Shiz 2017-06-29 19:20:05 does anyone know of an example of a apkovl creation script creating a user and group, or is that a bad thing to pursue? 2017-06-29 19:47:10 it appears I can just make my own /etc/group, /etc/passwd, and /etc/shadow, and add the /home path to the apkovl, but say something if that seems wrong to you 2017-06-29 19:47:55 okeuday_bak: that seems like an appropriate use of apkovl tbh 2017-06-29 19:48:08 Xe: k, cool 2017-06-29 19:48:33 just be double-dog sure you got /etc/shadow correct though lol 2017-06-29 19:53:02 Xe: I am making it easy since I can have the passwds be blank here, just want to be able to have some user separation 2017-06-29 19:53:29 ah, okay 2017-06-29 20:45:20 Hello. I may have done something wrong, or something strange is happening. I have been configuring my xen hypervisor since yesterday, and enabled (rc-update add xendomains default) xendomains which aren't so by default. Now, I have made some changes to the script (directly in /etc/init.d/xendomains), perhaps is this bad practice, but it used to work as expected and started my domains at startup. Today, 2017-06-29 20:45:22 after rebooting for independant reasons, the script did not fire. Running rc-status xendomains gives a "xendomains does not exist". Running rc-update add xendomains default gives "xendomains already installed in runlevel 'default'; skipping". 2017-06-29 20:45:39 Am I missing something ? Could this be a bug ? 2017-06-29 20:47:05 Also the file is still present in /etc/init.d/xendomains, and does not seem corrupted or modified whatsoever 2017-06-29 21:11:44 AlexIncogito: that is weird 2017-06-29 21:11:50 and rc-update add xendomains ? 2017-06-29 21:11:58 (yes, again) 2017-06-29 21:13:35 rc-status works on a runlevel, not on an init script 2017-06-29 21:13:58 rc-service xendomains status 2017-06-29 21:18:27 Ahah, genius. Ok I apparently inadvertently added a wrong character in the script, which rendered it moot 2017-06-29 21:18:50 Thank you 2017-06-29 21:23:03 a good habit to have is to run the script with the shell -n switch after modifying it 2017-06-29 21:25:39 Excellent idea sir 2017-06-29 21:32:46 I've wiped out enough workstations to memorize that one :D 2017-06-29 21:36:51 <_ikke_> 1heh 2017-06-29 21:39:12 ACTION insists on #!...sh -u 2017-06-29 21:39:22 whatever shell, but -u should just stay in there 2017-06-29 21:39:39 i'm writing our style guides for $work atm 2017-06-29 21:39:45 hoping i can put them online too 2017-06-29 21:39:49 well I kinda code all of my shellscripts in "strict mode" anyway 2017-06-29 21:39:55 :)) 2017-06-29 21:40:02 -eu and with bash -o pipefail 2017-06-29 21:40:14 yeah 2017-06-29 21:40:42 it was a pain in the ass at first but I notice my scripts have hugely improved since that 2017-06-29 21:40:50 well -e i hate since a guy in pkgsrc just added it everywhere 2017-06-29 21:40:54 to improve things 2017-06-29 21:41:00 he fucking broke whole pkgsrc 2017-06-29 21:41:12 should've started by -n that idiot 2017-06-29 21:41:22 yeh, it's quite painful to add afterwards :) 2017-06-29 21:41:28 and i was in middle of porting to a diff platform 2017-06-29 21:41:35 yeah, underrstandably, and the code was bad 2017-06-29 21:41:44 but the arrogance to call it a good idea like that :) 2017-06-29 21:42:07 i don't think we'll generally do -e but for any critical part they gotta 2017-06-29 21:42:24 i just don't wanna have a situation like, uh, no vms started because the first failed 2017-06-29 21:42:27 etc 2017-06-29 21:43:00 I've talked several times about this installer I've done for work... it's not even that big really, but trying to patch it to be "strict" was so much work that I chose to start from scratch 2017-06-29 21:43:04 but i'm at the point where i wanna put our devs on fire so there's a lot of options to improve 2017-06-29 21:43:09 hehe 2017-06-29 21:43:16 and it was less work that way? 2017-06-29 21:43:39 first as an exercise ported it from bash to pure sh and it was quite neat but then I started missing some specific bash features and coded it form scratch - again 2017-06-29 21:43:45 :)) 2017-06-29 21:44:01 i thought about making something that replaces all binary calls with first /bin/true and then with /bin/false 2017-06-29 21:44:09 to see how hardened the scripts really are 2017-06-29 21:44:10 but it was still probably easier than trying to fix the old one 2017-06-29 21:44:23 less burnoutish for sure 2017-06-29 21:45:18 yeh, and when you really work with one specific toolset you keep advancing so fast that you'll end up improving all the code whenever you rewrite it 2017-06-29 21:45:26 good point 2017-06-29 21:45:41 i need to get over thqat bash people put functions in front of functions 2017-06-29 21:45:45 just to break portability 2017-06-29 21:45:50 or local inside a function 2017-06-29 21:45:55 as if it would ever not be local in there 2017-06-29 21:46:20 because others like it and i just gotta adjust if i ask them to go with a standard 2017-06-29 21:46:36 but i just think it's a waste of bytes :) 2017-06-29 21:47:13 i mean the word "function" before funcname() 2017-06-29 21:47:20 it has no meaning but it's there 2017-06-29 21:47:27 I've also said this several times before, but what I'd like in my shell is slightly better scoping, it would enable me to do objectlike programming without too much hassle 2017-06-29 21:48:21 but the most common counterargument to that is "if you need that, just use python" :) 2017-06-29 21:48:29 grab that book some day: 2017-06-29 21:48:30 https://www.amazon.com/Unix-Shell-Objects-Christopher-Jones/dp/0764570048 2017-06-29 21:48:39 i don't recommend following much in there 2017-06-29 21:48:49 but it is super entertaining and likeable 2017-06-29 21:49:24 oh, I've actually done OOP in bash, it's just a bit painful to write 2017-06-29 21:49:36 ...you can do oop in bash? 2017-06-29 21:49:39 i'd use python if python had similar culture of error handling 2017-06-29 21:49:42 that sounds satanic 2017-06-29 21:49:44 in ksh already :) 2017-06-29 21:49:58 just for fun, and to flex fingers 2017-06-29 21:50:11 or maybe, sometimes, to make something simpler on the outside 2017-06-29 21:50:23 I've got a strong background in TCL so I made an "advanced" array class to help script generation (I write a lot of script generators) 2017-06-29 21:50:32 oh wow 2017-06-29 21:50:45 code generation is far above my level 2017-06-29 21:51:16 like, you define an array with "array x" and then use it like "x add 'blah'" 2017-06-29 21:51:47 it was a nice exercise :) 2017-06-29 21:53:08 Xe: sure you can, you can do oop in most procedural languages to an extent, you just end up writing a bit more code 2017-06-29 21:53:11 btw, what would you guys recommend for style guide... markdown or rather working scripts with doxygen or pod or heredoc inside 2017-06-29 21:53:38 (markdown with snippets obviously) 2017-06-29 21:55:09 I don't really have an opinion on that, since most scripts I write for a living I'm the entire target audience of... 2017-06-29 21:55:34 hehe 2017-06-29 22:54:04 what actually writes resolv.conf with a dhcp setup, shouldn't udhcpd be doing that? and why would it refuse to do so on a custom iso? 2017-06-29 23:17:05 udhcpc I mean, it always complains about no space left on device, due to the way the iso is mounted 2017-06-29 23:42:15 So... I've been giving mythological names to my services. Hermes for IMAP, Mercury for SMTP, Cerberus for Auth (that was an easy one), Argus for networking monitoring & reporting, ... But I can't choose one for the file syncing service (syncthing). Ideas ? 2017-06-29 23:43:24 "Sync" looks sad now compared with the other 2017-06-29 23:44:26 not mythological but imo good enough: Equalizer 2017-06-29 23:45:43 AlexIncogito ^ 2017-06-29 23:49:32 The sementic does fit the purpose, but I will certainly obsess over the new-ag-y/antique discontinuity 2017-06-29 23:50:09 Now I picture transformers and hercules side by side 2017-06-29 23:50:21 A sad view realy 2017-06-29 23:53:34 maybe search in another pantheon? 2017-06-29 23:55:00 Southern europe has monopoly there, but I hear nordic gods are tough, so it might just work 2017-06-29 23:57:03 Loki would definetly work for the windows service 2017-06-29 23:57:12 But still new age/antique 2017-06-29 23:59:18 I'll make an exception for this one since it also goes way back in history 2017-06-30 00:05:25 hermes 2017-06-30 00:11:09 Already taken by imap 2017-06-30 00:11:27 Fides (fiducia) for certbot service. That works, one more down 2017-06-30 00:16:36 your Tor gateway... well... 2017-06-30 00:19:01 That one will feel right at home 2017-06-30 00:26:41 fyi: I'm getting `temporary error` on `apk add --no-cache $pkg` from 151.101.0.249 2017-06-30 00:27:06 wget'ing the address says `503 Backend is unhealthy` 2017-06-30 00:29:29 http://dl-cdn.alpinelinux.org/alpine/v3.6/main/x86_64/APKINDEX.tar.gz 2017-06-30 00:34:04 cdn is dying again 2017-06-30 00:34:25 try other mirriros like cz.alpine... or nl.alpine... 2017-06-30 00:46:35 A wise internet individual suggested Harmonia, which I adopted. Thank you for your participation 2017-06-30 12:19:46 Hi, in wich package do I find badblock? 2017-06-30 12:20:30 https://pkgs.alpinelinux.org/contents?file=badblocks&path=&name=&branch=v3.6&repo=&arch=x86_64 2017-06-30 12:23:45 Hello. I have an issue mounting a disk image. In xen : "file:/var/opt/apkovl/alpine-repo.img,sda,r", which only contains the .apkovl.tar.gz. ; in extra: "apkovl=sda:ext4:alpine-repo.apkovl.tar.gz". When I boot the guest, it does not load the apkovl, nor does /dev/sda is present 2017-06-30 12:28:33 thanks 2017-06-30 12:34:01 So this sda is an emulated scsi device. The guest is a HVM, since alpine comes with PVHVM drivers, it is supposed to detach the qemu device, and reattach it using PV drivers 2017-06-30 12:35:04 I am unsure why this apparently fails. I have tried with different images in case the issue lied in a corrupted image 2017-06-30 12:36:28 Attaching xvd* devices works properly, so this may be correlated with the scsi feature 2017-06-30 12:39:43 gosh.. got it 2017-06-30 12:40:03 the #alpine-linux effect strikes again 2017-06-30 12:40:10 sda is mounted as xvda in alpine linux 2017-06-30 12:40:12 ahahah 2017-06-30 12:40:25 and xvda was already tied to another device 2017-06-30 12:42:39 AlexIncogito: So you think, xvda is only once useble? 2017-06-30 12:43:14 I'm not a linux expert by any mean, but I assume there is a collision in the namin 2017-06-30 12:43:52 Actually the sd*->xvd* completely make sense, since the scsi emulated device is detached with profit to its PV counterpart 2017-06-30 12:44:02 And PV devices are always xvd* 2017-06-30 12:45:03 Right 2017-06-30 12:45:45 Surprisingly there were no error messages 2017-06-30 12:45:59 But I'm using alwas xvda for the first HD in each DomU 2017-06-30 12:48:14 And there we go, a diskless, read/only LBU alpine linux ! Great success 2017-06-30 12:48:18 *y* 2017-06-30 12:51:16 So you start out of an image and you have no more disks? 2017-06-30 12:53:29 Diskless was abusive, but I used it for simplicity. The guest has 3 block devices: alpine iso /dev/cdrom (ro), apkovl.tar.gz /dev/xvdb (ro), ext4 data device /dev/xvdc (rw) 2017-06-30 12:54:16 Using extra= kernel options, it loads the configuration from apkovl.tar.gz, and kicks off the cd 2017-06-30 12:54:54 Data (in this instance a mirror of alpine repo) are written to xvdc, and that is all it is used for 2017-06-30 12:58:02 But why are you do it that complcated? 2017-06-30 12:58:13 Yust to try? 2017-06-30 12:59:03 It mitigate risks of allowing an attacker persist his presence on the gues 2017-06-30 12:59:05 t 2017-06-30 12:59:23 Also, it makes it very to reproduce/deploy/backup an instance 2017-06-30 12:59:57 ...very easy... 2017-06-30 13:00:22 I type faster than my keyboard can handle, so annoying... 2017-06-30 13:01:10 Once upon a time, I solved the meaning of life, but still... keyboard didn't register... 2017-06-30 13:01:13 Are you using LVM on Your Dom0 2017-06-30 13:01:42 Yes 2017-06-30 13:03:44 Shiz: could you merge this? https://github.com/alpinelinux/aports/pull/1792 2017-06-30 13:10:11 AlexIncogito: I do it like that, it's yust a sample: 2017-06-30 13:10:23 https://pastebin.com/WDPeEstj 2017-06-30 13:13:31 What do you use the different volume groups for ? 2017-06-30 13:18:04 vg0 is yust for boot devices, I dont't backup it that often ... 2017-06-30 13:19:37 The rest will backuped weekly 2017-06-30 13:21:47 I also have seen, that it is used a single volume group for each DomU 2017-06-30 13:22:13 But I think that is over the top ;) 2017-06-30 13:58:49 cu all 2017-06-30 14:21:15 hi 2017-06-30 14:21:30 how can I build a docker image using alpine linux with dnsmasq? 2017-06-30 14:29:46 amosbird, is a pretty big question 2017-06-30 14:30:14 gyif 2017-06-30 14:30:17 anyway: 2017-06-30 14:30:18 https://hub.docker.com/r/andyshinn/dnsmasq/ 2017-06-30 14:30:23 this is a good start 2017-06-30 14:30:42 fthanks 2017-06-30 14:30:56 yw 2017-06-30 14:34:50 do I always use FROM alpine:edge ? 2017-06-30 14:42:23 hmm 2017-06-30 14:42:24 https://la.wentropy.com/a0PO 2017-06-30 15:26:09 hello world 2017-06-30 15:26:36 can I increase the UID_MAX for adduser? 2017-06-30 15:26:43 or can I install useradd? 2017-06-30 15:39:10 https://pkgs.alpinelinux.org/contents?file=useradd&path=&name=&branch=v3.6&repo=&arch=x86_64 2017-06-30 15:41:02 is UID_MAX really a problem? https://github.com/mozilla-b2g/busybox/blob/master/loginutils/adduser.c#L39 2017-06-30 15:42:35 hm that seems to be different code 2017-06-30 15:43:58 https://git.alpinelinux.org/cgit/aports/tree/main/busybox/busyboxconfig?h=3.6-stable#n497 seems to be this 2017-06-30 15:50:38 hi, does alpine provide bash by default? 2017-06-30 15:50:49 there is a bash package, the default shell is busybox ash 2017-06-30 15:50:52 can I run this script using ash https://la.wentropy.com/gpzN ? 2017-06-30 15:51:13 if you change /bin/bash to /bin/sh or /bin/ash probably 2017-06-30 15:51:24 ok 2017-06-30 16:12:49 Bun: yes. Because of our internal setup, I need the UID of a user to be higher than the UID_MAX currently set in busybox 2017-06-30 16:18:40 useradd is in the shadow package, if that works for you 2017-06-30 16:19:51 otherwise maybe bug ncopa or open a ticket 2017-06-30 16:25:26 the shadow package worked for me 2017-06-30 16:25:31 thank you very much 2017-06-30 16:54:46 amosbird: sed -i is not posix. But I think busybox provides it anyway, check for that. 2017-06-30 17:05:22 yeah, it works 2017-06-30 17:34:48 In Alpine specifically I am trying to run this cronjob but it is not running, I can get it to run on other Linux OS types but not Alpine (the script has executable permissions for all users and I've manually started crond so that when I run ps I see it running): */30 * * * * /usr/local/bin/my_script.sh 2017-06-30 17:35:20 I even have another cron that's not running either: * * * * * /bin/echo "test" > /tmp/test 2017-06-30 17:35:26 what is the sheebang ? 2017-06-30 17:35:46 in the script it's: #!/usr/bin/env bash 2017-06-30 17:35:58 well, do you have bash installed 2017-06-30 17:36:07 yes 2017-06-30 17:36:08 by default bash is not present 2017-06-30 17:36:24 it's there in /bin/bash 2017-06-30 17:36:49 and it's strange that even a simple echo command wouldn't write to /tmp 2017-06-30 17:37:01 which cron daemon are you using? 2017-06-30 17:37:22 I am running /usr/sbin/crond 2017-06-30 17:38:26 so busybox crond? 2017-06-30 17:38:28 or dcron 2017-06-30 17:38:30 could be either 2017-06-30 17:39:27 ah sorry, it's: BusyBox v1.25.1 (2016-10-26 16:15:20 GMT) multi-call binary 2017-06-30 17:52:46 Anyone have any ideas? It's so strange 2017-06-30 20:23:05 When using run from RAM configuration, installed apk are supposed to be reinstalled automatically when local backup has been updated with lbu commit, correct ? 2017-06-30 20:25:47 presuming /etc/apk/world is included in your lbu backup, which it should be I think, yes 2017-06-30 20:27:04 Hmm.. it is, after rebooting the apks are not reinstalled. But I need to check something, I've got a hunch 2017-06-30 20:41:41 https://justpaste.it/18g2q 2017-06-30 20:41:48 This is weird, I must have done something wrong 2017-06-30 20:42:14 The lbu setup is mostly working, setup-* are saved and so on 2017-06-30 20:42:30 But no apk 2017-06-30 20:43:49 hmm 2017-06-30 20:43:55 what happens when you run `apk fix` after boot 2017-06-30 20:44:01 Where can I check the logs of what is being installed at startup ? 2017-06-30 20:44:07 Let me try that 2017-06-30 20:44:10 your console :P 2017-06-30 20:44:48 apk fix seem to do the trick 2017-06-30 20:44:51 heh 2017-06-30 20:44:54 I'm using xl console 2017-06-30 20:45:00 I can't scroll that far :( 2017-06-30 20:45:17 hmm 2017-06-30 20:45:22 Should I just add a manual command ? Is that normal behavior ? 2017-06-30 20:45:25 it may be that you don't have ip connectivity early enough 2017-06-30 20:45:30 try adding ip=dhcp to your kernel command line 2017-06-30 20:45:33 (presuming you use dhcp) 2017-06-30 20:46:01 I do, also I do see the DHCP offer accepted in start logs 2017-06-30 20:46:04 I will try that 2017-06-30 20:46:34 Actually 2017-06-30 20:46:50 It seem a bit silly, this is a repository mirror 2017-06-30 20:47:12 Shouldn't I just point it to the local repository directory ? 2017-06-30 20:47:39 your local repo dir probably does not have all apks in the repo 2017-06-30 20:47:48 (presuming it's the same one that's distributed with our releases) 2017-06-30 20:47:58 "his repo was a few packages short" 2017-06-30 20:48:14 ... 2017-06-30 20:48:19 sorry, please continue 2017-06-30 20:48:21 Well I rsynced all the main and edge x86_64 apks 2017-06-30 20:48:53 that terminology is confusing considering main is a repo in edge 2017-06-30 20:48:57 Basically I followed the setup your own repository mirror with --include --exclude to only retain main/edge x86_64 2017-06-30 20:49:05 sorry 2017-06-30 20:49:10 I meant 2017-06-30 20:49:31 v3.6/*/x86_64 and edge/*/x86_64 2017-06-30 20:49:37 right 2017-06-30 20:49:39 my mistake 2017-06-30 20:49:49 then it would work under the presumption that the initramfs can find that repo dir :P 2017-06-30 20:50:17 Yep it's a scsi qemu, which is then mounted with PV driver to xvdb 2017-06-30 20:50:25 xvda* sorry 2017-06-30 20:50:47 So if I put the repository potion in kernel it would theoritically work ? 2017-06-30 20:50:55 option of course 2017-06-30 20:51:40 /dev/xvda isn't a directory but a block device 2017-06-30 20:51:48 it'd need to be mounted somewhere first, i don't think the initramfs does that 2017-06-30 20:52:04 what about https://wiki.alpinelinux.org/wiki/Local_APK_cache#To_manually_enable_Local_Cache_on_run-from-RAM_installs 2017-06-30 20:52:31 and symlink the cache to v3.6/main/x86_64 even 2017-06-30 20:53:37 alpine_repo=/var/www/localhost/htdocs/alpine/ 2017-06-30 20:54:18 Bun: I believe this was related to "prior v2.3" iirc 2017-06-30 20:54:45 @Shiz: The device is mounted in fstab 2017-06-30 20:55:20 right but you can still replace the symlink 2017-06-30 20:55:41 eh 2017-06-30 20:55:49 AlexIncogito: i think a better way to do all of this is to use setup-apkcache 2017-06-30 20:55:57 not the wiki above, and definitely not anything with alpine v2.3 2017-06-30 20:56:32 Sounds much cleaner, good idea 2017-06-30 20:59:41 Nay that did not do the trick 2017-06-30 20:59:52 I see it installing 28 modules at startup 2017-06-30 21:00:01 But it does not include lighttpd for some reasons 2017-06-30 21:00:20 Hmm 2017-06-30 21:00:40 Hold that thought /etc/apk/repositories does not contain the new directory 2017-06-30 21:01:23 Oh silly me 2017-06-30 21:01:25 This is cache no repo 2017-06-30 21:02:40 But then I should setup-apkrepos, not apk-cache, for this is a mirror of the repo, not a cache 2017-06-30 21:02:46 Let me try that 2017-06-30 21:04:03 seems better to just drop the local mirror entirely in my personal opinion 2017-06-30 21:04:11 and just rely on cache, since i think that solves what you'd want to solve 2017-06-30 21:08:56 Not really, I would like to feed this repository to my other alpine instances on the network 2017-06-30 21:09:16 Worst case I can use apk fix, but I'd like to get to the bottom of it 2017-06-30 21:09:19 hmm, rght 2017-06-30 21:09:38 'll let you know if I make progress 2017-06-30 21:10:06 are you planning to feed it over the network too? 2017-06-30 21:10:08 :P 2017-06-30 21:11:14 Heh I wish I had that kind of resource 2017-06-30 21:13:03 so how does that work, if I may ask? 2017-06-30 21:13:06 out of pure curiosity 2017-06-30 21:13:15 mounting the same virtual disk under many vms? 2017-06-30 21:13:43 Oh wait I thought you were asking if I wanted to put online a public mirror 2017-06-30 21:14:00 Yeah it will be inter-vm network, with isolated bridges 2017-06-30 21:14:20 The ":P" part got me wondering heh 2017-06-30 22:46:56 @Shiz: So indeed, it seem the issue is it cannot find the correct apk. I've been trying to figure out why it fails 2017-06-30 22:47:11 It goes like this 2017-06-30 22:47:21 * Mounting boot media: [ 1.025330] sr 1:0:0:0: [sr0] scsi3-mmc drive: 4x/4x cd/rw xa/form2 tray 2017-06-30 22:47:24 [ 1.025342] cdrom: Uniform CD-ROM driver Revision: 3.20 2017-06-30 22:47:26 [ 1.138868] EXT4-fs (xvda2): mounted filesystem with ordered data mode. Opts: (null) 2017-06-30 22:47:28 [ 1.313594] EXT4-fs (xvdb): mounted filesystem with ordered data mode. Opts: (null) 2017-06-30 22:47:30 ok. 2017-06-30 22:47:44 xvda is my /var device, which holds the repository mirror files 2017-06-30 22:47:49 so far everything seem fine 2017-06-30 22:48:03 Right after that, apkovl loads 2017-06-30 22:48:10 * Loading user settings from /media/xvdb/alpine-repo.apkovl.tar.gz: 0% 2017-06-30 22:48:12 (ok) 2017-06-30 22:48:54 And then it goes into the package install part 2017-06-30 22:49:10 Installing packages to root filesystem: WARNING: Ignoring /var/www/localhost/htdocs/alpine/v3.6/main/x86_64/APKINDEX.tar.gz: No such file or directory 2017-06-30 22:49:34 For some reason it fails to find this file, and thus use the local repository 2017-06-30 22:49:49 It then falls back to the next one, which is /media/cdrom, and holds only minimal apk 2017-06-30 22:50:16 And then online repo, but there is no network available at that step, just as you predicted 2017-06-30 22:50:39 This is strange, because xvda2 is correctly mounted, or supposedly so, and the APKINDEX.tar.gz exists 2017-06-30 22:51:41 When I run apk update/apk upgrade after logging in, apk does not complain about this repo and actually finds it has the same revision as the online repo 2017-06-30 22:52:08 I assume this suggests an issue with the state of how /var is mounted at that point in the boot process 2017-06-30 22:53:00 Shiz: is there anyway to get someone to merge a package into testing? it will be awhile before I am able to test it if this goes much longer, it has been there for 7 days, which seems somewhat typical 2017-06-30 22:53:58 okeuday_bak: by means of a PR 2017-06-30 22:54:14 Shiz: yeah, I mean the PR has been sitting there for 7 days 2017-06-30 22:55:09 Shiz: it seems like it requires using the separate git repo with a manual merge, since the master isn't really github, so that might be part of the delay I perceive 2017-06-30 22:55:16 got a link? 2017-06-30 22:55:30 Shiz: https://github.com/alpinelinux/aports/pull/1763 2017-06-30 22:56:04 holy moly that's a lot of build deps 2017-06-30 22:56:50 any reason for the -j1? 2017-06-30 22:57:02 Shiz: so that can spook people? just due to potential complexity? 2017-06-30 22:57:10 nah, just a remark 2017-06-30 22:57:35 Shiz: yeah -j1 is currently required due to java jar file creation and erlang compilation both doing copies in automake, it is something that will be fixed in the future 2017-06-30 22:57:54 might be useful to comment that in the apkbuild 2017-06-30 22:58:10 Shiz: k, sure 2017-06-30 22:59:48 i'm not a big fan of that initd script 2017-06-30 23:00:05 also, does it really need g++ at runtime? 2017-06-30 23:03:58 Shiz: unfortunately it can't get libstdc++ any other way, and the abuild has a warning about that, where it has to be there 2017-06-30 23:04:27 it shouldn't need libstdc++ explicitly specified, i think 2017-06-30 23:04:44 and libstdc++ is just its own package 2017-06-30 23:05:24 Shiz: the warning said it shouldn't be a build only dependency, I can look at a different c++ dependency, but so far, that is what the warning said to do 2017-06-30 23:05:59 Shiz: the initd script is using the included release script, since the way erlang releases are handled is a bit complex 2017-06-30 23:06:15 isn't there simply a binary that openrc can supervise? 2017-06-30 23:07:00 Shiz: no, since it is trying to use signals to kill stuff, not have a clean shutdown 2017-06-30 23:50:33 Shiz: https://github.com/CloudI/CloudI/blob/develop/src/rel/files/cloudi otherwise this stuff would get buried in an OS-specific file 2017-06-30 23:50:41 added the comment 2017-06-30 23:50:53 yuck. 2017-06-30 23:51:55 Shiz: more specifically, it is problematic how the pid file is created, just due to how erlang wants to spawn stuff 2017-06-30 23:52:16 pidfiles on their own are problematic are wrong 2017-06-30 23:52:48 i don't think this script is needed at all tbh 2017-06-30 23:53:24 Shiz: the initd one? that would force local usage for starting/stopping 2017-06-30 23:53:33 no, that script you linked just now 2017-06-30 23:54:40 Shiz: you would rather have that complexity in an OS-specific file? 2017-06-30 23:54:49 90% of the complexity there is unnecessary 2017-06-30 23:55:04 if not more 2017-06-30 23:55:15 Shiz: there are different ways of executing the erlang VM 2017-06-30 23:56:56 Shiz: you are welcome to try to reduce it to nothing, but it would be an exercise in futility