2017-11-01 00:06:18 I have my Alpine laptop setup so it wont connect automagically to the internet unless i run udhcpc -i wlan0 2017-11-01 00:06:42 would that, in theory, keep attacker using the new wpa2 exploit on me 2017-11-01 00:38:47 what would be a reason for alpine to keep dropping my wifi connection? 2017-11-01 00:39:03 you have a marvell wifi card? 2017-11-01 00:39:25 It is not an issue with location. 2017-11-01 00:39:35 I am unsure what make my card is. 2017-11-01 00:39:54 I am pretty sure it is atheros 2017-11-01 00:39:59 lspci 2017-11-01 00:40:30 RTL8821AE 802.11ac 2017-11-01 00:40:37 realtek lol 2017-11-01 00:40:41 yup 2017-11-01 00:40:48 awful chips :( 2017-11-01 00:40:54 yup!! 2017-11-01 00:41:15 m4chm4n: attempt to remedy that if you can 2017-11-01 00:41:31 most computers can have the wifi card replaced if you try hard enough 2017-11-01 00:41:42 Any good options for internal? 2017-11-01 00:42:30 atheros or older ralink cards are good, haven't tried them since mediatek bought them 2017-11-01 00:42:33 some intels are alright too 2017-11-01 00:42:44 I have a spare intel wifilink 1000 2017-11-01 00:42:52 m4chm4n: tias? 2017-11-01 00:43:17 ?? 2017-11-01 00:43:25 Try It And See :) 2017-11-01 00:43:30 lol ok 2017-11-01 00:44:12 it will be the first time I opened this laptop too..... 2017-11-01 00:44:33 just be sure you are well awake so you are appropriately grounded :D 2017-11-01 00:44:45 guess I can throw some arctic silver on there too 2017-11-01 00:47:15 Downside to all this, I may end up losing BT with this swap 2017-11-01 00:48:57 i don't consider that a huge loss, i never end up needing bluetooth 2017-11-01 00:51:33 I use it for a speaker, but I found another intel that has BT and 5ghz band for 15 dollars 2017-11-01 00:53:43 The speakers on this lenovo ideapad are pure trash unfortunately..... 2017-11-01 01:03:17 if this intel does not work, is there a way to keep from dropping the signal? When I iwconfig after the wifi drops out, the card shows no essid attached to it. I know its not the card, becuase it does not happen with any other distro. 2017-11-01 01:03:43 i believe our intel wireless driver has some bugs 2017-11-01 01:03:54 but i havent been able to debug it because when i switched APs 2017-11-01 01:04:05 it quit doing it 2017-11-01 01:06:32 well shit.... 2017-11-01 01:07:22 oh well! Its only two terminal commands to fix the problem. 2017-11-01 01:08:11 But sometimes udhcpc doesnt like to work properly after it cuts out 2017-11-01 01:08:33 m4chm4n: cron? 2017-11-01 01:08:49 just have it check every 5 minutes for a stoned wifi card and reset it 2017-11-01 01:09:04 what about a stoned user!?!?!?! 2017-11-01 01:09:33 computers exist as a means to enact policy, it is a very bad idea to use computer logic when dealing with users though 2017-11-01 01:09:55 It was a bad joke.... 2017-11-01 01:10:15 i know 2017-11-01 01:11:06 .o0()0o. 2017-11-01 01:11:44 I will look into that cron idea though. sounds like it might work. 2017-11-01 01:14:13 I always see it doing things in my server logs. Never knew I could set em up to do small tasks like that. 2017-11-01 01:14:58 Dont really know much about linux yet..... But I am learning something new everyday. 2017-11-01 01:17:07 Took me a month to setup my alpine ssh server on an hp workstation. It works great but I am still having trouble with rsa keys.... 2017-11-01 01:37:39 m4chm4n: i suggest using the default settings if you can, they are known to work perfectly 2017-11-01 01:40:13 default settings for cron? 2017-11-01 01:53:10 m4chm4n: sshd 2017-11-01 01:53:29 im using dropbear 2017-11-01 03:41:02 im making a package and testing now, when i do abuild build it fails because of dependency missing , ive specified in APKBUILD but is not installed in my system yet, if i instal manually it will be in world (is glib) but i dont want to be in world, so there is some way to tell abuild to install missing depencencies before? 2017-11-01 03:41:44 -r flag 2017-11-01 03:44:59 is there a way to have the C preprocessor only run for a given file and not include headers for the entire world? 2017-11-01 03:45:14 in the output* 2017-11-01 03:45:52 ups i thought ok and how do i know if some dependency is only needed at compile time or also at running? 2017-11-01 03:46:11 it'll install things it needs when you run `abuild` 2017-11-01 03:48:51 but its me who has to specify depends and makedepends 2017-11-01 03:49:35 and i know i need curl glib openssl and gmp 2017-11-01 03:49:58 but dont know where to put each one as dependency or package-dev in makedepends 2017-11-01 03:50:39 i just read in install file these are depencencies 2017-11-01 03:51:26 and configure warns if they are not specified, but im just puting curl glib in depends then curl-dev glib-dev openssl-dev ... in makepdepends 2017-11-01 03:51:34 i dont know what im doing really 2017-11-01 03:57:53 almost works but.... i put a --disable-docs in configure and now install does not founds README etc... 2017-11-01 04:12:29 ok make install tries to install docs even if i specify --disable-docs in configure so if i have to include docs should i make the subpackage -doc? that does not means write another APKBUILD? 2017-11-01 04:13:04 or just specify the subpackage 2017-11-01 04:16:48 it keeps failing when installing doc 2017-11-01 04:17:17 it compiles 2017-11-01 04:17:26 just fails when installing docs 2017-11-01 04:17:37 can someone take my APKBUILD and fix? 2017-11-01 04:17:50 gist it? 2017-11-01 04:18:53 http://sprunge.us/QGgZ 2017-11-01 04:22:08 it doesn't work with libressl, i can't build it on my test machine 2017-11-01 04:22:59 i have libressl too 2017-11-01 04:23:08 but openssl-dev is just to compile 2017-11-01 04:23:39 watch this 2017-11-01 04:23:42 (13/26) Purging openssl-dev (1.0.2l-r0) 2017-11-01 04:23:44 (14/26) Purging zlib-dev (1.2.11-r1) 2017-11-01 04:23:46 (15/26) Purging libssl1.0 (1.0.2l-r0) 2017-11-01 04:24:02 it as been installed just for compiling the package even if libressl is in my system 2017-11-01 04:24:15 # cat /etc/apk/world| grep ssl 2017-11-01 04:24:17 libressl 2017-11-01 12:29:29 hi 2017-11-01 12:30:00 I'm getting into some trouble with OracleJDK java binaries I downloaded. 2017-11-01 12:30:23 I'm no expert on alpine, so I just did what I normally do, extracted the package and checked the binaries were compatible 2017-11-01 12:30:30 like ldd on them or run them 2017-11-01 12:30:48 however alpine just refuses to recognize the binaries even as files 2017-11-01 12:30:58 what do you mean by that 2017-11-01 12:31:00 they are there, but you execute them and I get no such file 2017-11-01 12:31:00 pastebin the output? 2017-11-01 12:31:05 oh 2017-11-01 12:31:11 musl vs glibc 2017-11-01 12:31:12 because they're compiled for glibc likely 2017-11-01 12:31:18 iorhi: hi, alpine uses musl libc, while oraclejdk is built with glibc 2017-11-01 12:31:56 you should ask oracle to provide binaries for musl libc 2017-11-01 12:32:07 i know they worked on it earlier, but not sure how it ended 2017-11-01 12:45:07 ncopa, oh 2017-11-01 12:45:09 Ic 2017-11-01 12:45:11 ok 2017-11-01 13:21:44 I have a question about a flagged package for nodejs, from the page pkgs.alpinelinux.org it says it was flagged automatically based on notification from Anitya 2017-11-01 13:24:52 If 6.11.5-r0 is not included anymore which is the 6.X version that I can use? 2017-11-01 13:51:34 Is the package nodejs:6.11.5-r0 flagged wronly in the pkgs.alpinelinux.org ? 2017-11-01 14:26:36 my soekris box, where alpine is still unable to see the CF and the HDD, does see the attached usb disk 2017-11-01 14:28:54 and while trying this i found that the box, which is unable to boot from USB directly, now boots the CF (initramfs) and then continues from USB 2017-11-01 14:29:23 so i got alpine live system running - it's still unable to see the CF and HDD though 2017-11-01 14:29:45 i wonder what else I could try from there? 2017-11-01 14:43:24 qoh: maybe check kernel config 2017-11-01 14:43:43 i suspect something is missing in the kernel config 2017-11-01 14:48:15 all the CONFIG_PATA options are set to module, so should have full converage 2017-11-01 14:48:23 including CONFIG_PATA_PCMCIA which would cover CF cards 2017-11-01 14:48:29 perhaps a missing module 2017-11-01 14:58:24 scv: 'missing' as in 'not in alpine at all?' 2017-11-01 14:58:47 missing as in not loaded 2017-11-01 14:59:18 which model soekris is this 2017-11-01 14:59:36 scv: net4801 2017-11-01 15:06:21 the driver is indeed not being built 2017-11-01 15:06:53 CONFIG_PATA_SC1200 is the option needed and it isn't present in any alpine kconfig i've checked 2017-11-01 15:06:58 scv: which one would it be? (and how did you find out? been searching all about the interwebs :) 2017-11-01 15:07:02 so you will need to build a kernel and enable that 2017-11-01 15:07:23 checked the manual for the device, been a /long/ time since i've even seen one of those things :p 2017-11-01 15:07:43 the processor is SC1100, but the pata driver is SC1200? for both, i guess? 2017-11-01 15:07:47 yep 2017-11-01 15:07:55 SC1200 is the support chip 2017-11-01 15:08:14 iirc 2017-11-01 15:08:34 and you just knew that i guess? hehe. 'cause there's no mention of SC1200 in the manual ... 2017-11-01 15:09:10 was that dropped just in alpine or generally in linux kernel ? 2017-11-01 15:09:20 it mentions SC1100 and checking through all the ata options there are natgeo/amd driver options 2017-11-01 15:09:38 it seems to be disabled in a lot of distros from googling around, likely due to age 2017-11-01 15:09:48 not many people are using IDE controllers anymore 2017-11-01 15:11:23 compiling a new kernel is kind of daunting ... especially since i would need to cross compile (or wait a year for it to build on the box itself :/ 2017-11-01 15:12:13 lemme see if i can crank a kernel out for you 2017-11-01 15:12:13 sec 2017-11-01 15:41:21 scv: are you compiling one? 2017-11-01 15:42:20 yeah 2017-11-01 15:42:37 had to go handle something else briefly sorry :p 2017-11-01 15:43:16 nice :) 2017-11-01 15:43:39 chances to have it activated back in alpine are 0 i guess? 2017-11-01 15:43:58 couldn't answer that unfortunately 2017-11-01 15:44:06 ncopa might know, or kaniini perhaps 2017-11-01 15:45:26 depends on what the "cost" is 2017-11-01 15:45:46 if it is a simple module, and thats it, then it shouldnt be too problematic 2017-11-01 15:46:00 specially since this will only be x86 and not the other arches (i suppose) 2017-11-01 15:49:07 ncopa: what would be required to verify that 'cost'? 2017-11-01 15:49:31 can it be built as module? 2017-11-01 15:49:32 no other arches 2017-11-01 15:49:33 and yes 2017-11-01 15:49:51 it seems to be the only PATA module disabled, which concerns me it might have some strange interaction with other hardware or some specific reason for being disabled 2017-11-01 15:49:54 if not it will increase size of kernel for everybody, regardless if they have the hardware or not 2017-11-01 15:50:44 if it needs to be enabled on all arches, then it requires more work (higher 'cost') 2017-11-01 15:51:03 it only applies to a single cpu variant, the natsemi/amd geode gx1 2017-11-01 15:56:57 *ding* 2017-11-01 15:57:07 qoh: http://git.brinstar.org/brinstar/x86/linux-net4801-4.9.59-r0.apk 2017-11-01 16:02:32 scv: downloaded. nice & shiney! thanks a dozen 2017-11-01 16:02:42 np 2017-11-01 16:03:41 how could we find out if there's a specific reason? ... if it's a module, as long as nobody loads it there shouldn't be issues, i guess? 2017-11-01 16:04:42 hmm now how can i get it from the apk into the CF card/initramfs? 2017-11-01 16:07:27 you said you had it booting on usb right? how are you creating the usb 2017-11-01 16:08:23 that was just the recent iso dd'd to the stick 2017-11-01 16:11:08 so i suppose: install in the life environment ... and then somehow get it installed during along 'setup-alpine' 2017-11-01 16:12:50 bit chicken and egg though, i doubt the module will load 2017-11-01 16:14:17 sec 2017-11-01 16:23:55 qoh: http://git.brinstar.org/iso/alpine-net4801-171101-x86.iso 2017-11-01 16:23:58 give that a shot 2017-11-01 16:27:38 hm 2017-11-01 16:28:33 well, that'll get you booted on usb with the right kernel at least 2017-11-01 16:28:45 it still installs vanilla if you just run setup-alpine 2017-11-01 16:29:24 should get you on the right track though 2017-11-01 16:32:13 i'll check. it only has 82M. is that good still? i had used the extended which is around 275MB 2017-11-01 16:35:11 just recalled that I did the initial installation to CF on another machine. so that shouldn't be the issue 2017-11-01 16:43:48 if you do the install from another machine then you can just load the package there 2017-11-01 16:47:03 at it :) 2017-11-01 16:58:47 booted into the CF on another machine. still warns: ERROR: You tried to add a non-repository package to system, but it would be lost on next reboot. Enable package caching (apk cache --help) or use --force if you know what you are doing. 2017-11-01 17:01:32 i suppose that 'setup-bootable' uses what is currently installed in the iso for the bootable to be set up? 2017-11-01 17:02:02 otherwise the hint to install syslinux before 'setup-bootable' as per https://wiki.alpinelinux.org/wiki/Create_a_Bootable_Compact_Flash\#Automated wouldn't make much sense (?) 2017-11-01 17:08:06 so i ran 'setup-apkcache'. still loads of erros regarding /lib/... being on read-only FS ... ignore or remount,rw? 2017-11-01 17:09:06 ah remount rw is not an option. iso fs ... 2017-11-01 17:14:32 running out of ideas. i'm stuck. 2017-11-01 17:53:32 can somebody help me with an apkbuild 2017-11-01 17:53:49 ? 2017-11-01 17:54:27 aGaTHoS: what kind of help do you need? 2017-11-01 17:55:17 im writting an apkbuild but dont works, and dont know how to solve 2017-11-01 17:55:40 i know the probem but dont know what i have to do as i dont fully understand apkbuild functions etc.. 2017-11-01 17:55:52 its just at doc install... 2017-11-01 17:55:59 the rest compiles and works 2017-11-01 17:56:07 what's the error message emitted by abuild? 2017-11-01 17:56:23 /usr/bin/install -c -m 644 './docs/megadf.1' '/home/h/devel/aports/testing/megatools/pkg/megatools/usr/share/man/man1/docs/megaget.1 docs/megals.1 docs/megamkdir.1 docs/megaput.1 docs/megareg.1 docs/megarm.1 docs/megacopy.1 megacopy.1 megacopy.1' 2017-11-01 17:56:26 install: can't create '/home/h/devel/aports/testing/megatools/pkg/megatools/usr/share/man/man1/docs/megaget.1 docs/megals.1 docs/megamkdir.1 docs/megaput.1 docs/megareg.1 docs/megarm.1 docs/megacopy.1 megacopy.1 megacopy.1': No such file or directory 2017-11-01 17:56:32 it seems install is not doing what it should 2017-11-01 17:56:49 takes first param a file then the rest of files are taken as the name for the target 2017-11-01 17:57:16 dont know where to changue this and how as moreover ive specified a subpackage for -doc 2017-11-01 17:57:57 this is the apkbuild file 2017-11-01 17:58:06 http://sprunge.us/QGgZ 2017-11-01 17:58:38 ive made it for megatools 2017-11-01 17:59:37 maybe i have to adjust also depends but atm i just need to becompiled and installed cleanly 2017-11-01 18:01:16 well now i see this is not really the first error, is the last, but the other is the same for README LICENSE ... 2017-11-01 18:01:21 another set of doc files 2017-11-01 18:02:22 same error it seems it tryies to install allfiles at time, well i dont have many experience using install command 2017-11-01 18:04:24 but dont know how to solve as if i write by hand the code to install each doc dile i cant use make install to install the rest of the package... 2017-11-01 18:04:56 and if i have to modify the makefile or something create a patch etc... is to much for my experience with this 2017-11-01 18:04:59 I was looking into using alpine as base for my mailserver. dovecot was not compiled with --with-pam which I'd like to use though. what is the prefered way to continue with this? change the APKBUILD and install my own version of dovecot? can anyone point me in the right direction? 2017-11-01 19:53:42 ok, so i want my system to be read-only (since it's a CF ... and for security reasons) ... but how do i get the custom kernel onto such a system... 2017-11-01 21:21:09 is this a known issue? 2017-11-01 21:21:11 (1/2) Reinstalling util-linux-bash-completion (2.31-r0) 2017-11-01 21:21:11 ERROR: util-linux-bash-completion-2.31-r0: trying to overwrite usr/share/bash-completion/completions/rfkill owned by bash-completion-2.7-r1. 2017-11-02 00:26:19 where can I find my sshd logs? 2017-11-02 00:28:05 ah, /var/log/messages, duh 2017-11-02 01:48:37 Hi all, I just hit https://github.com/tianon/docker-bash/issues/4 when running some tests on alpine:edge (docker). Can any of you reproduce this on a non-docker installation of Alpine? 2017-11-02 01:49:11 Wondering if this is a docker oddity or something affecting alpine as a whole. 2017-11-02 01:49:44 alpine:edge - bash-4.4 only. Works fine in alpine:latest 2017-11-02 02:27:12 someone help with thi please http://sprunge.us/QGgZ 2017-11-02 02:27:25 i dont know why make install is not working properly 2017-11-02 03:18:02 Unode: seems like a regression in bash 4.4 ;/ 2017-11-02 12:23:36 Hi 2017-11-02 12:23:50 no pptp client for Alpine? 2017-11-02 12:24:46 $ apk search pptpclient 2017-11-02 12:24:46 pptpclient-doc-1.9.0-r2 2017-11-02 12:24:46 pptpclient-1.9.0-r2 2017-11-02 12:25:54 ACTION sent a long message: cypa[m]_2017-11-02_12:25:53.txt  2017-11-02 12:26:06 e.g. not found pptpclient 2017-11-02 12:27:00 I'm confused 2017-11-02 12:27:25 ouh ... testing!!! 2017-11-02 12:27:25 looks like its in edge/testing 2017-11-02 12:27:38 can you please verify that it works, then give me feedback 2017-11-02 12:27:47 and i'll move it so it gets included in v3.7 2017-11-02 12:28:03 OK minute 2017-11-02 12:28:26 does it need an init.d script? 2017-11-02 12:29:23 not in my case, I guess 2017-11-02 12:44:56 seems pptpclient works fine ... 2017-11-02 12:45:11 nice 2017-11-02 12:45:25 and default config is in a shape that can be shipped? 2017-11-02 12:47:00 mmm ... I use manual config from here https://wiki.archlinux.org/index.php/PPTP_Client 2017-11-02 12:47:52 so I've reviewed config manually, and I've installed pptpd first so default /etc/ppp/options.pptp is from pptpd package actually 2017-11-02 12:48:14 ok good 2017-11-02 12:48:21 ACTION sent a long message: cypa[m]_2017-11-02_12:48:20.txt  2017-11-02 12:49:07 do you need the pptpd and pptpclient installed at the same time? 2017-11-02 12:49:10 is that common? 2017-11-02 12:51:36 no, I do not need pptpd now 2017-11-02 12:57:01 i cleaned up the apkbuild a bit 2017-11-02 12:57:20 can you please try that 1.9.0-r3 still works? I think it does, but just to doublecheck 2017-11-02 12:57:36 it should be available in 5mins or so 2017-11-02 12:57:42 if its ok, then will i move it 2017-11-02 12:57:57 i claimed maintainership of it too 2017-11-02 13:04:29 Hello, I have the 'tun' module loaded and the device /dev/net/tun with 777 permissions, but vpnc (edge-testing) still does not want to load it with 'can\'t initialise tunnel interface: No such file...' - I also tried installing openvpn to get it to create the tun device but it still does not work, can someone help please? 2017-11-02 13:05:03 Also attempted strace, it seems to try all /dev/tun### and give up 2017-11-02 13:05:24 Is it perhaps a linux-hardened thing? should I try vanilla? 2017-11-02 13:13:32 I suppose the main question is how do I create a /dev/tun0 device? I do see the 'ip tunnel add tun0 mode ...' but that seems to require me to know the ip routes that would be negotiated by the vpn itself 2017-11-02 13:20:26 anything in dmesg when you try to run vpnc? 2017-11-02 13:21:06 kahiru: No, nor in /var/log/messages 2017-11-02 13:21:49 It it something to do with the /etc/vpnc/vpnc-script? it seems that runs after the connection is made however, to deal with routing and ip addresses 2017-11-02 13:23:56 I'm not sure, I never used vpnc. but I assume it should work in a smiliar fashion as openvpn or tinc does 2017-11-02 13:27:49 o/ 2017-11-02 13:29:26 yo ScrumpyJack 2017-11-02 13:30:52 god i miss Alpine at work 2017-11-02 13:31:02 what do you work with there? 2017-11-02 13:50:53 Show how do I get openvpn to create /dev/tun0? The command `openvpn --mktun --dev tun0` only creates an entry in `ip taptun list`, not the device file 2017-11-02 13:53:18 kahiru: They gave me a mac with osx on it 2017-11-02 13:54:09 de_worde: are you setting something up for the first time or join a VPN? 2017-11-02 13:55:22 ScrumpyJack: Sorry about that, got disconnected as I sent a reply 2017-11-02 13:56:02 ScrumpyJack: Yes, using vpnc to connect to an old cisco VPN 2017-11-02 13:56:32 I have the credentials/group password, worked on ubuntu with vpnc, just cannot get it to work on alpine (lost the ubuntu install) 2017-11-02 13:57:16 But its failing with 'can't initialise tunnel interface: No such file or driectory ... strace -f does not reveal anything helpful either, just thats its trying all /dev/tun### 2017-11-02 14:05:14 i was looking on the gentoo wiki and alpine wiki but cant seem to find documentation on how to create an openrc file 2017-11-02 14:05:23 i was looking for syntax and stuff 2017-11-02 14:07:54 alazare619: you mean creating a /etc/init.d/ script or using rc-update to add/remove services to a runlevel? 2017-11-02 14:08:16 creating an openrc script 2017-11-02 14:08:33 i looked at a few others in /etc/init.d/ but i wanted to know what all was supported 2017-11-02 14:09:08 alazare619: Like https://wiki.alpinelinux.org/wiki/Writing_Init_Scripts ? 2017-11-02 14:09:23 didnt see that one 2017-11-02 14:09:24 alazare619: Details at https://wiki.gentoo.org/wiki/Handbook:X86/Working/Initscripts 2017-11-02 14:09:30 i was googling openrc syntax 2017-11-02 14:09:36 and openrc documentation 2017-11-02 14:09:43 guess i should of typed init -_- 2017-11-02 14:09:46 openrc is the manager, init scripts are their own thing =] 2017-11-02 14:11:04 like i wanted to modify the docker on (or create a custom one) 2017-11-02 14:11:07 that depends on NFS 2017-11-02 14:11:31 and then create an openrc init file for each docker container to auto start them on boot depends on docker 2017-11-02 14:12:12 alazare619: the LXC init.d script has a good example of that, it has like lxc.containername, so you would want a custom docker one with docker.containername 2017-11-02 14:12:46 But docker scripts have their own init don't they? would it not just go into that? 2017-11-02 14:12:57 not really an "init syste" 2017-11-02 14:13:13 but they support restart: always 2017-11-02 14:14:32 which is more of a fire once then watch if it fails but i would rather that be possibly handled by openrc or maybe something else not sure 2017-11-02 14:20:04 ScrumpyJack: oh well, but hey, at least its not windows 2017-11-02 14:21:21 it's not so much the OS, it's that I can figure out how to replace OSX on this macbook 2017-11-02 14:21:29 so i'm stuck with it 2017-11-02 14:22:19 oh 2017-11-02 14:31:37 it just refuses to boot from usb 2017-11-02 14:32:14 ScrumpyJack: Is that the latest 2017 one? I have not tried with mine yet, I did have a rEFInd usb working with the 2016 model 2017-11-02 14:32:40 I just kind of gave up because of various features such as lid close, gpu, etc. 2017-11-02 14:32:55 and god knows how to get the touchbar working on alpine =D 2017-11-02 15:03:53 Anyone else know how to get a '/dev/tun#' device created? The /dev/net/tun does not seem to work for VPNC, various ip commands do not create the device file 2017-11-02 15:13:37 can you create a tun device by hand? 2017-11-02 15:14:40 Do you mean outside of tunctl? 2017-11-02 15:14:43 or..? 2017-11-02 15:16:46 ScrumpyJack: AG: tunctl does not create the /dev/tun# device either 2017-11-02 15:16:54 ACTION steps back 2017-11-02 15:17:04 It says it was successful 2017-11-02 15:17:18 but I assume it requires C io calls to /dev/net/tun for that to work 2017-11-02 15:19:53 I suppose I could ssh to it, grab the strace -f output, sanatise it and put it on pastebin.. 2017-11-02 15:20:24 just run tunctl 2017-11-02 15:21:11 ScrumpyJack: Set 'tap0' persistent and owned by uid 0 2017-11-02 15:21:48 de_worde_: do you have tun module loaded? 2017-11-02 15:21:57 przemoc: Yes, lsmod shows it 2017-11-02 15:21:59 neat, so you see if with ifconfig for example and you can play with it? 2017-11-02 15:22:31 ifconfig tap0 192.168.3.54 ? 2017-11-02 15:22:43 ScrumpyJack: Yes I did get as far as making tun0 appear in ifconfig, but still no entry in /dev/tun0 for it and vpnc is not working 2017-11-02 15:22:58 I was under the impression the vpn client would set it up with the correct ip 2017-11-02 15:23:05 tun/tap dev will not appear in /dev/ 2017-11-02 15:23:20 Yeah, I have also tried installing and starting udev 2017-11-02 15:24:03 This is just a memory install, so I scrap it completely and started it again a few times with different configurations 2017-11-02 15:24:45 /dev/net/tun is the way to create them via ioctl() 2017-11-02 15:24:58 ScrumpyJack: I now have a tap0 and tun0 with ifconfig 2017-11-02 15:27:05 przemoc: Only ioctl in strace a few times is `ioctl(1, TIOGWINS2, 0x...) = -1 ENOTTY (Not a tty) 2017-11-02 15:27:53 Perhaps vpnc 0.5.3 does not support ioctl for the new tun module and only works with the older kernel... 2017-11-02 15:27:55 kaniini: that's the odd part. bash 4.4 works fine here (not alpine). Could this be a weird result from compiling bash with muslc? 2017-11-02 15:27:57 i'm not sure about vpnc, but you'll probably need configure it to use your tun device. i see it has a config file and a script file 2017-11-02 15:29:11 ScrumpyJack: I'll give those settings a go, but it seems to only support having a physical file /dev/tun#, is there a way to get that functionality back instead of using ioctl on /dev/net/tun? 2017-11-02 15:36:14 in its strace it really tries various /dev/tun# files? 2017-11-02 15:38:01 przemoc: Yes, from /dev/tun0 to /dev/tun254 2017-11-02 15:38:19 The vpnc-script does check for /dev/net/tun and that part of it seems fine 2017-11-02 15:41:14 /dev/tunN thing is a BSD-thing IIRC. I'm not sure if anything like that was ever in Linux 2017-11-02 15:41:25 Using the ifname, for tap0 I get 'error: arbitrary naming tunnel interface is not supported in this version' and tun0 is 'can\'t initialise tunnel interface: no such file or directory' 2017-11-02 15:49:51 kaniini: https://bugs.alpinelinux.org/issues/8098 2017-11-02 15:50:09 I wonder if the issue is just that the wrong version of bash was packaged. 2017-11-02 15:50:17 4.4.0 instead of 4.4.12 2017-11-02 15:50:33 but the package still shows up as 4.4.12-r0 in apk 2017-11-02 15:51:17 Is anyone else running a yubikey with CCID enabled? I'd like a sanity check on what I'm running into. I've installed - I believe - everything I need. However, pcscd fails to initialize the yubikey. The error is... 2017-11-02 15:51:32 RFInitializeReader() OPen Port 0x2000001 failed (...port here) 2017-11-02 15:51:37 I've added the udev rules. 2017-11-02 16:34:17 de_worde_: sorry can't help more. sounds like a vpnc issue, as you can happily create the interfaces yourself. 2017-11-02 16:39:02 if you have the time and energy you could look at https://github.com/ndpgroup/vpnc/blob/b1243d29e0c00312ead038b04a2cf5e2fa31d740/sysdep.c#L94 to figure out what's going on 2017-11-02 16:54:32 ScrumpyJack: Maybe IFF_TUN is not defined: https://github.com/ndpgroup/vpnc/blob/b1243d29e0c00312ead038b04a2cf5e2fa31d740/sysdep.c#L433 ... The include is linux/if_tun.h 2017-11-02 16:54:55 going to install the headers and check 2017-11-02 16:56:29 omg its been commented out: https://git.alpinelinux.org/cgit/aports/tree/testing/vpnc/working.patch 2017-11-02 16:59:05 Shiz: Hey, thanks for moving vpnc from unmaintained, but why was the if_tun.h commented out? 2017-11-02 17:41:47 Heh, include just needed to be moved... 2017-11-02 17:53:01 So I fixed the vpnc package, just modified the existing patch and updated the checksums, what is the quickest way to submit it? 2017-11-02 17:53:34 email a patch to the mailing list 2017-11-02 17:55:13 https://wiki.alpinelinux.org/wiki/Creating_patches 2017-11-02 17:57:07 Hi. Just noticed, that there is no dkms in alpine. What is the right way of keeping 3rd paty modules in alpine? 2017-11-02 18:00:31 <_ikke_> de_worde_: Or submitting a PR in github 2017-11-02 22:39:55 . 2017-11-03 04:51:27 5~can anybody help me with this? http://sprunge.us/chVI 2017-11-03 07:15:17 whats about it? 2017-11-03 07:21:31 https://rclone.org/ 2017-11-03 07:21:58 maybe rclone supports mega.nz already? 2017-11-03 07:23:51 oops! no rclone package for Alpine? 2017-11-03 07:24:31 <_ikke_> cypa[m]: Nobody packaged it yet apaprently 2017-11-03 08:07:16 ncopa: care to explain why alpine dovecot is built without --with-pam? what was the reason behind that? 2017-11-03 08:46:24 morning 2017-11-03 08:46:51 bl1nk: the reasoning is: dont add support for features that nobody has asked for, to keep things simple and small 2017-11-03 08:47:05 ... til someone asks for it 2017-11-03 08:47:15 that way we avoid enabling features nobody uses 2017-11-03 08:47:37 so... I can kindly ask to enable it? :) 2017-11-03 08:47:45 morning by the way 2017-11-03 08:47:56 linking with pam will pull in pam as dependency for everybody, even for those who dont use it 2017-11-03 08:48:22 bl1nk: can you please file an issue on bugs.a.o? 2017-11-03 08:48:43 sure! 2017-11-03 08:48:53 its mostly for history, so i can refer to the issue in the commit message 2017-11-03 08:49:26 so in the future, if i wonder why it was enabled, i have the reason 2017-11-03 09:04:21 ncopa: done, thank you :) 2017-11-03 09:15:09 hey, does full disk encryption with grub without separated /boot partition will work on Alpine? 2017-11-03 09:22:00 You can probably make it work if you install a bootloader that supports it, but I don't think extlinux (the default bootloader) can read LUKS volumes 2017-11-03 09:22:24 Why do you want to do it? Adding extra complexity to the boot process normally just creates problems for yourself later, and nothing sensitive should be in /boot anyway 2017-11-03 09:23:53 grub support it and just thought somebody did it without some weirdness steps 2017-11-03 09:24:11 grub support in alpine is not very polished 2017-11-03 09:24:25 well, dont wanna create any scripts to check if somethnig or somebody modificated images in /boot 2017-11-03 09:24:34 oh 2017-11-03 09:24:39 you are pretty much on your own there 2017-11-03 09:24:44 with FDE you will need to enter your password twice 2017-11-03 09:25:14 MY_R: would be nice if you could experiment with it, and share your experiences 2017-11-03 09:25:15 will avoid that by crypttab file 2017-11-03 09:25:49 our current setup scripts does not support it 2017-11-03 09:25:53 found some paste... so the only weird step there is recompile grub... but it is old 2017-11-03 09:25:59 https://gist.github.com/joncfoo/9181cdac0adc4f5459ae 2017-11-03 09:26:17 might be it works with the current provided grub 2017-11-03 09:26:24 checked now and grub is in fine version and package include lvm2 so should works fine 2017-11-03 09:26:48 im interested to hear if it works! 2017-11-03 09:27:11 would also be great if you could write a wiki article or similar with your findings 2017-11-03 09:27:15 or even a blog article 2017-11-03 09:27:28 so it can be googled 2017-11-03 09:27:54 ye, first have to check if that will work at all :) 2017-11-03 09:27:55 something like this? https://wiki.alpinelinux.org/wiki/Create_A_VirtualBox_Guest_with_Grub_and_XFS 2017-11-03 09:30:28 that is already prety good base, just add grub steps https://wiki.alpinelinux.org/wiki/LVM_on_LUKS 2017-11-03 11:18:47 Does anyone have experience installing alpine on zfs root? 2017-11-03 11:20:38 i have 2017-11-03 11:20:44 its somewhat tricky 2017-11-03 11:20:56 due to boot media does not include the zfs module 2017-11-03 11:21:17 so the "simplest" thing to do is generate your own boot media with zfs enabled 2017-11-03 11:29:36 well, I'm booting from UEFI vfat partition, so it is not a problem. I just don't understand how to make initramfs-init mount zfs on root 2017-11-03 11:30:32 As I understand, it have to be mounted on /sysroot before switch, but if I specify zfs set mountpoint=/sysroot tank/ROOT/default it does not work 2017-11-03 11:31:34 In the end it all goes from how I just don't understand how nlplug-findfs works 2017-11-03 12:11:01 With ZFS legacy mounts it finally works. But it looks "wrong" somehow. 2017-11-03 12:53:32 bl1nk: what do you use PAM for with dovecot? 2017-11-03 12:53:56 bl1nk: https://github.com/alpinelinux/aports/commit/10b458891b729fee807ff3860398019a1b18c5af#commitcomment-25376732 2017-11-03 12:53:59 I use it as my passdb 2017-11-03 12:54:09 what do you use as backend? 2017-11-03 12:54:14 the /etc/passwd? 2017-11-03 12:55:30 yes 2017-11-03 12:55:36 you dont need pam for that 2017-11-03 12:56:05 there's the passwd driver too 2017-11-03 12:56:21 yes 2017-11-03 12:56:23 dovecot wiki encourages to use pam for that though 2017-11-03 12:56:44 If you'd rather not build dovecot with pam I can try to use the passwd driver 2017-11-03 12:56:53 or build it on my own 2017-11-03 12:57:36 i tend to agree with jirutka here. the benefit with pam is small, while cost is relatively high 2017-11-03 12:58:47 i'm fine with not having pam if you'd rather want less deps for dovecot 2017-11-03 13:02:36 well, LVM on single LUKS partition with grub works, but ye need to figure out how to avoid type password twice :) 2017-11-03 13:15:42 ncopa: can pam not be moved to its own package like it's done with mariadb/sqlite/ldap etc? 2017-11-03 13:22:28 I mean the pam driver 2017-11-03 13:30:52 bl1nk: thats the problem. it cannot be built as separate module (i think) 2017-11-03 13:30:55 but i will check that 2017-11-03 14:24:58 so that GRUB encryption thing, it's completely separate from the OS encryption? 2017-11-03 14:26:16 if so, and if you trust the encryption GRUB provides you with, then you can keep a keyfile on the GRUB encrypted container, partition or whatever it is, and no more two passwords 2017-11-03 14:35:25 ah okay, so it's LUKS 2017-11-03 14:38:03 bl1nk: i checkd. its not possible to build pam support in a plugin 2017-11-03 14:38:10 thats why we didnt do it in the first place 2017-11-03 14:38:43 bl1nk: i am also 100% sure it is possible to do the auth without pam, because I use that myself 2017-11-03 14:38:57 or, wait.. 2017-11-03 14:39:00 no i dont use it 2017-11-03 14:39:32 i think i use dovecot passwd file 2017-11-03 14:39:43 i will revert it. sorry 2017-11-03 15:35:38 ncopa: thanks for looking into it 2017-11-03 15:36:17 bl1nk: im pretty sure you don't need PAM 2017-11-03 15:36:40 you only need to add dovecot user to the group "shadow" 2017-11-03 15:37:18 and in your dovecot config enable passwd and point it to /etc/shadow and userdb to /etc/passwd 2017-11-03 15:37:21 TBB: well, mkinitcpio or dracut got some hooks which support key files, in mkinitfs I see nlplug-findfs which expect passphrase only or dunno :\ 2017-11-03 15:38:03 ncopa: I am thinking about moving to virtual users now. Not sure if that's less complicated than creating system users for everyone. 2017-11-03 15:38:43 depends if the users need to log in to the server, eg with ssh 2017-11-03 15:38:48 no never 2017-11-03 15:38:50 that's why 2017-11-03 15:39:05 then I'd definitively would recommend virtual users 2017-11-03 15:39:09 it was just convenient for me to just create a system user and not think about it 2017-11-03 15:39:31 how does users change password? 2017-11-03 15:39:59 it's all family 2017-11-03 15:40:10 so you do it for them 2017-11-03 15:40:20 they enter it on my keyboard 2017-11-03 15:40:31 but that happened once 2017-11-03 15:41:52 there is a tool to generate the virtual users passwords 2017-11-03 15:42:09 which is close to as simple as system user 2017-11-03 15:42:19 in your case i'ts probably easier with system suer 2017-11-03 15:42:23 user* 2017-11-03 15:44:31 I will find a way 2017-11-03 16:22:45 cypa[m]: ive tried to package megatools but dont work, dont know how to solve, it fails in its make install phase, think Makefile needs a patch .. 2017-11-03 16:26:54 does http://mega.nz works with rclone? 2017-11-03 16:27:13 which protocol does http://mega.nz use? 2017-11-03 16:28:38 im watching the website and seems not 2017-11-03 16:31:52 i dont think it uses a standard protocol 2017-11-03 16:33:10 any protocol description? 2017-11-03 16:37:04 i cant access the website 2017-11-03 16:37:13 but its supposed there is the api etc.. 2017-11-03 16:38:16 I prefer hacking API to rclone and packaging rclone for Alpine 2017-11-03 16:40:09 for me is ok as i also use drive and box which i see are supported by rclone 2017-11-03 16:43:32 but still need megadl for just downloading from mega links out of the browser 2017-11-03 16:43:56 megadl is one of the binaries of megatools 2017-11-03 16:44:39 which im using ... as in the process of making the package it compiles just installed libssl1.0 as dependency, so i just copy in /usr/local/bin 2017-11-03 16:44:53 because as the package fails when installing docs... 2017-11-03 16:45:19 why do you want package it after all? 2017-11-03 16:45:28 copy it manually for now 2017-11-03 16:46:15 thats why i did but its a pity, its a patch, its an exception, it should be done the right way 2017-11-03 16:48:34 well i suppose i can rewrite the package function not use make install at all and copy each file "manually" from the function 2017-11-03 16:48:41 that is what you say? 2017-11-03 16:50:18 do you distribute these tools to many users? 2017-11-03 16:50:23 and then a question comes to me, if i specify a doc subpackage the function doc is gonna do it right once i have all files installed in the fake root? 2017-11-03 16:51:36 no its just for me and i suppose im gonna send the apkbuild when finished to a ticket asking to include the package in the aports 2017-11-03 16:53:10 I'm sceptic if you really need a package for you personally ... maybt if you want to practice packaging? 2017-11-03 16:54:26 what do you think should i do? 2017-11-03 16:54:52 using the abuild system is good ... it downloads and install makedepends for me an clean after 2017-11-03 16:55:27 if i just download what im missing compile and install in /usr/local that is what i think you think i should do it will be harder 2017-11-03 16:55:32 when dealing with dependencies 2017-11-03 16:56:42 also ishould kepts sources 2017-11-03 16:56:49 to do a meke uninstall 2017-11-03 16:56:58 if i dont want trash in /usr/local 2017-11-03 16:57:12 if i make a local repo 2017-11-03 16:57:24 all goes better 2017-11-03 16:57:40 have you compiled the tools successfully? 2017-11-03 16:57:48 yes 2017-11-03 16:57:59 it only fails when installing docs 2017-11-03 16:58:28 watch this: /usr/bin/install -c -m 644 ./LICENSE NEWS TODO README INSTALL HACKING '/home/h/devel/aports/testing/megatools/pkg/megatools/usr/share/doc/megatools' 2017-11-03 16:58:30 install: can't stat './LICENSE NEWS TODO README INSTALL HACKING': No such file or directory 2017-11-03 16:58:32 make[2]: *** [Makefile:796: install-docDATA] Error 1 2017-11-03 16:58:47 it takes a string with files as a single file name 2017-11-03 16:59:00 something is wrong with the Makefile or the install command 2017-11-03 17:16:50 ok im gonna try to use the install-sh script by hand on each file, $pkgdir is /.../pkg/ or /.../pkg/$pkgname ? 2017-11-03 17:24:01 the install command works fine, so it seems like a makefile issue 2017-11-03 17:47:48 ok i dont understand really how it works this and ive read the wiki... now it worked but dont know where the package is... 2017-11-03 17:48:46 http://sprunge.us/QXha 2017-11-03 17:49:00 this is the APKBUILD 2017-11-03 17:49:26 not sure if im useing properly the pkgdir variable 2017-11-03 17:49:43 i compiled wwith abuild -r and now 2017-11-03 17:50:03 i have nothing aparently in the aports dir but the APKBUILD 2017-11-03 17:50:08 where is my package? 2017-11-03 17:50:15 in ~/packages 2017-11-03 17:52:43 ok is here /var/local/apks/testing/x86_64 this dir ive configured i dont remember where and now have a testing dir distinct that the old one jut X86_64 where i tried to make a loca repo 2017-11-03 17:54:26 ok is in /etc/abuild.conf and i have a ANTAINER thats why yhe varuabe in the apkbuild was not set 2017-11-03 17:55:40 cant apk add megatools 2017-11-03 17:56:05 ive made a apk update; apk upgrade 2017-11-03 17:57:06 for what is wrong 2017-11-03 17:57:12 it can work like this 2017-11-03 17:58:12 should i add a repo /var/local/apk/testing? 2017-11-03 17:58:21 i have already /var/local/apk/ 2017-11-03 17:58:36 or should i move megatools to the root local repo directory 2017-11-03 17:59:48 how should i work if i want to make local packages? not creating the new package in aports/testing but in another completely diferent dir? 2017-11-03 18:01:24 can abuild work on any directory or it needs something in the aports tree other than apkbuild's? 2017-11-03 20:08:34 Hi, i am new to alpine and wanted to know if there is openjdk oder oracle java version for arm architecture ? Maybe someone has some infos regardings this topic 2017-11-03 20:10:21 <_ikke_> https://pkgs.alpinelinux.org/packages?name=*jdk*&branch=v3.6&repo=&arch=armhf&maintainer= 2017-11-03 20:10:39 <_ikke_> that's armhf 2017-11-03 20:10:54 <_ikke_> aarch64 also has the same packages 2017-11-03 20:11:43 ok sorry and thank you, i was some kind of blind 2017-11-03 20:12:39 <_ikke_> np 2017-11-03 20:18:27 anyone here working on openwrt ui on alpine linux? 2017-11-03 20:20:15 like i wouldnt mind getting luci working on alpine linux 2017-11-03 20:39:23 alazare619: i will work on it when and if i get alpine ported to mips 2017-11-03 20:39:37 Hi, I was looking for a pointer of systemd support for Alpine 2017-11-03 20:39:59 danieli: id love to see an x86 port of luci w/ alpine 2017-11-03 20:40:01 <_ikke_> gugurohu: alpine is not planning on supporting systemd 2017-11-03 20:40:20 that would put alpine+luci in line to compete with the likes of the Bsd Devil itself PfSense 2017-11-03 20:40:46 we'll see, i'm more interested in alpine on mips myself 2017-11-03 20:40:55 isnt alpine just lede? 2017-11-03 20:41:01 which lede was openwrt 2017-11-03 20:41:09 and openwrt was mips 2017-11-03 20:41:19 so in a round about way it should already? 2017-11-03 20:41:30 alpine has no relationship with any of those projects 2017-11-03 20:42:52 musl connecting everyone! :P 2017-11-03 20:47:26 could of swore it was forked from lede (hence the vpn patches in the kernel) 2017-11-03 20:47:54 well it’s not 2017-11-03 20:48:46 Ok found it 2017-11-03 20:48:55 <_ikke_> Found what? 2017-11-03 20:49:03 Alpine began as a fork of hte leaf project which was in turn a fork of the linux router project 2017-11-03 20:49:09 so not directly openwrt/lede 2017-11-03 20:49:15 but similiar thats why i was mistaken 2017-11-03 20:52:52 <_ikke_> The mailing list post link on wikipedia is not loading 2017-11-03 20:56:24 so is that not correct that alpine is a fork of LRP/LEAF 2017-11-03 20:56:51 nope it’s not correct 2017-11-03 20:57:51 someone should fix that wikipedia article then 2017-11-03 20:58:16 https://en.wikipedia.org/wiki/Alpine_Linux 2017-11-03 21:07:00 Does the most recent Alpine Linux image has systemd installed on it 2017-11-03 21:07:25 nope 2017-11-03 21:11:35 gugurohu: alpine linux uses openrc 2017-11-03 21:11:44 it will not be using systemd any time soon (thank god) 2017-11-03 21:12:03 gugurohu: https://wiki.alpinelinux.org/wiki/Writing_Init_Scripts 2017-11-03 21:12:17 thanks for the link 2017-11-03 21:23:36 Also this page does not state why openrc is better than systemd.. 2017-11-03 21:24:53 or Alpine will not use systemd..any pointers on history/reasoning behind supporting one vs other 2017-11-03 21:25:09 <_ikke_> gugurohu: That topic is a hornets nest 2017-11-03 21:25:37 hmm..no small writeup..or any discussion pointer.. 2017-11-03 21:26:02 <_ikke_> gugurohu: that topic usually causes pages and pages of discussion 2017-11-03 21:28:49 k, so other qs: how much is openrc adopted..looks like a recent addition 2017-11-03 21:29:15 <_ikke_> openrc has been there a while iirc 2017-11-03 21:31:17 :) 2017-11-03 21:37:48 thought Alpine is minimalistic and simple distribution so dont have to add anything more why why why :P 2017-11-03 21:39:20 we do not use systemd because we do not want to use systemd 2017-11-03 21:39:25 that is all there is to say on it 2017-11-03 21:39:36 if you want to use systemd, i am sure you can find 48895283989238923893289321 other distributions that have it 2017-11-03 21:40:17 to clarify, we do not care either way about systemd, we just do not wish to use it ourselves because it is much larger than openrc 2017-11-03 21:40:42 <33333333333 for Alpine 2017-11-03 21:40:52 Sweden represent 2017-11-03 21:41:10 (yeah a bit drunk and shouldn't be here) 2017-11-03 22:22:55 really. 2017-11-03 22:47:09 what is amd-ucode for? 2017-11-03 22:49:00 microcode 2017-11-03 22:50:30 yes but what does this microcode? 2017-11-03 22:51:22 patching your cpu 2017-11-03 22:51:55 seriously? 2017-11-03 22:52:13 yes 2017-11-03 22:52:53 if i install amd-ucode from the non-free repository the kernel is gonna replace the microcode for that firmware? 2017-11-03 22:53:18 check by self after and before 2017-11-03 22:53:21 grep microcode /proc/cpuinfo 2017-11-03 22:53:37 nothing 2017-11-03 22:54:00 dmesg |grep microcode 2017-11-03 22:54:19 nothing but i dont installed 2017-11-03 22:54:25 do i need this? 2017-11-03 22:54:48 wont hurt to have it but who knows... 2017-11-03 22:55:26 but this modifies forever the cpu microcode or is just loaded ? 2017-11-03 22:59:30 just patching it early 2017-11-03 23:00:02 so can undo by remove it and regenerate ram disk 2017-11-03 23:03:23 i dont trust to install this who know which maliciou things does that blob 2017-11-03 23:07:43 so you can just remove your cpu from socket cus it is blob too :) 2017-11-03 23:08:16 new cpus for sure 2017-11-03 23:08:19 yeah your CPU already runs a blob, why would not updating an outdated blob to a new blob be better? 2017-11-03 23:08:35 they for sure have backdoors 2017-11-03 23:08:41 mine is quite old 2017-11-03 23:08:49 sure it only does what it should 2017-11-03 23:11:10 aGaTHoS: but you need to upgrade so you can post screenshots of the bios to social media 2017-11-03 23:11:22 /s 2017-11-03 23:13:41 social media i dont have this 2017-11-03 23:13:45 im out of society 2017-11-03 23:14:26 I have only social with no media 2017-11-03 23:14:30 $REPODEST is the var i should set to avoid abuild put my package in a community subdir as is in the aports if i want to not create that subdir in my repository directory? 2017-11-03 23:16:06 no i see that var is set in /etc/abuild.conf 2017-11-03 23:16:27 how do i make a local reposiitory??? 2017-11-03 23:17:51 i mean ive downloaded aports the make a package in aports/community/megatools then abuild and the package is in /var/local/apks/community/x86_64/... in /etc/apk/repositories i have added /var/local/apks 2017-11-03 23:18:31 abuild has indexed and signed but still apk after update upgrade does not finds when i try to add 2017-11-03 23:19:22 what im missing 2017-11-03 23:21:25 ok i cant invent a branch right my local repo has to end also with comunity etc...? 2017-11-04 00:36:09 in https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package when in the comment for git commit says put ...$pkgname etc... means literally the variable or the contents? 2017-11-04 00:40:30 the contents 2017-11-04 04:38:43 when specifying a patch in an aport it has to be in the root directory of the aport with the APKBUILD file or inside src/ 2017-11-04 11:57:56 whats the diference between edge and 3.6 repos 2017-11-04 11:58:14 should i use both or just one ? 2017-11-04 11:58:36 just one 2017-11-04 11:58:38 don't mix them 2017-11-04 11:59:09 3.6 is a stable release, edge is more testing 2017-11-04 11:59:13 thats what i thought 2017-11-04 11:59:30 im installing inside virtualbox 2017-11-04 12:00:49 use only one, mixing will screw your system up 2017-11-04 12:46:52 hello all, I cannot find usermod/moduser 2017-11-04 13:16:22 yb: what are you trying to do? 2017-11-04 16:53:33 Hello. I'm trying to compile a custom Rasbperry Pi kernel (as I want a 64-bit kernel). Is there anywhere I can find instructions for making the initram and modloop images? 2017-11-04 16:54:28 or is this an #alpine-devel question? 2017-11-04 17:55:15 Shiz: have you tried to use opensmtpd-table-passwd lately? it's in testing and I even installed opensmtpd from there because it seems you changes the libexecdir for builds from /usr/lib/opensmtpd to /usr/lib. the driver crashes on me though. not sure why. 2017-11-04 17:59:47 I did pin both edge/main to @edge and edge/testing to @testing, hope that's the right thing to do on alpine. then installed opensmtpd@edge and opensmtpd-table-passwd@testing. 2017-11-04 18:14:50 bl1nk: you don't have to pin edge/main if you do, it sounds like you mix some release version like 3.6 with edge 2017-11-04 18:16:09 edge is like unstable on debian, a rolling repo with lastest versions. while versioned releases are stable branches, mixing them does not work very well 2017-11-04 18:17:25 wWrAR07vpPZ1: I actually just installed both @testing 2017-11-04 18:17:57 wWrAR07vpPZ1: how would I just installed opensmtpd and opensmtpd-table-passwd from testing but nothing else then if not by pinning? 2017-11-04 18:18:20 testing is not part of the versioned releases 2017-11-04 18:18:55 I know. opensmtpd-table-passwd is not available outside testing 2017-11-04 18:19:52 i was mostly responding to your statement "I did pin .. edge/main to @edge ...hope that's the right thing to do on alpine." 2017-11-04 18:20:24 I see 2017-11-04 18:20:39 if I'd just enable edge/main or edge/testing I'd update my whole system to that, right? 2017-11-04 18:20:58 Because that's what I wanted to avoid 2017-11-04 18:21:27 edge/main would update the system yes. adding edge/testing does really nothing, except for maybe fail because of missing dependencies not available in your release verion 2017-11-04 18:21:32 s/verion/version/ 2017-11-04 18:22:12 alright, thanks! 2017-11-04 18:22:29 mixing release versions with edge is a recipe for trouble though afaik 2017-11-04 18:22:38 edge/main i mean 2017-11-04 18:23:05 but it seems opensmtpd and opensmtpd-table-passwd do not depend on anything else 2017-11-04 18:23:08 besides musl 2017-11-04 18:23:35 then i think you dont need to "pin" edge/main at all, adding edge testing should be enough 2017-11-04 18:24:22 okay 2017-11-04 18:25:00 apk add package --update-cache --repository http://nl.alpinelinux.org/alpine/edge/testing 2017-11-04 18:25:25 for only a specific package from testing repo, maybe this is what you are looking for ? 2017-11-04 18:26:31 Is that any different to installing package@testing with a pinned repo or even installing package that's just available in testing? 2017-11-04 18:30:24 you won't get any updates etc 2017-11-04 18:31:31 also, nothing else will be installed 2017-11-05 15:01:16 hi 2017-11-05 15:03:17 Venker: hi 2017-11-05 15:03:54 Does anyone knows how do I get linux_spi programmer on alpine linux for flashrom? 2017-11-05 15:04:31 I had installed one on rpi2 but then I didn't know that alpine's flashrom does not have linux_spi programmer. 2017-11-05 15:04:38 T_T 2017-11-05 15:04:50 Laptop broke, now only pi's left to fix it. 2017-11-05 16:03:34 Argh, as well why does keyboard input for omxplayer in alpine does not work in framebuffer. 2017-11-05 16:22:05 Now it doesn't even have subtitles! 2017-11-05 16:22:09 ACTION is so frustrated 2017-11-05 21:05:45 Hi, are the update servers healthy? 2017-11-05 21:06:15 I'm getting these errors: https://pastebin.com/YLrF1RHi 2017-11-05 21:08:14 Oh, seems my DNS server is down, instead :) 2017-11-06 01:08:54 I noticed that Krita was not available in the repos. Is it easy to compile it on alpine? 2017-11-06 03:32:34 how would I change the time settings of "Alpine will boot automatically in 3 second" screen when I turn my computer on 2017-11-06 03:34:59 you fiddle with /etc/update-extlinux.conf 2017-11-06 03:41:12 thank you sir 2017-11-06 03:46:30 is that the only file that would affect this? I have changed it twice and its still the same. 2017-11-06 03:49:26 You'll want to run update-extlinux after changing it 2017-11-06 03:49:41 ah I see 2017-11-06 03:57:14 Is there a way to make udhcpc run at bootup? 2017-11-06 05:39:48 it is possible to config optional dependencies for a package? 2017-11-06 05:43:44 <_ikke_> aGaTHoS: It doesn't seem like it 2017-11-06 05:45:30 anyone know how i bootstrap go-bootstrap? 2017-11-06 05:45:31 but is possible to compile a package with support for something i dont have installed? 2017-11-06 05:45:32 lmao 2017-11-06 05:53:55 I dont really understand what I'm doing, ok im making a package for surf (light web browser) based on the PKGBUILD of archlinux , so this package dependencies are for instance webkit2gtk and as optional depends dmenu and mpv (for enter an address and play a video repectively) my questions are two 2017-11-06 05:55:07 <_ikke_> aGaTHoS: does the package work when you don't have them installed? 2017-11-06 05:55:20 1 - can i specify webkit2gtk in depends or webkit2gtk-dev in makedepends as i want, and that means in first case the package has to be installed in the system as is loaded dinamically and second case is linked statically? 2017-11-06 05:56:37 <_ikke_> aGaTHoS: *-dev packages in makedepends are not specifically for statically linked packes. 2017-11-06 05:56:54 <_ikke_> aGaTHoS: it's also needed for dynamically linked libs 2017-11-06 05:56:55 and 2 - for the optional depends, dmenu is something i considere essential but not mpv, that means that i need to add some makedepends to the program use this features or the y just need to be in the ystem to be used and if they are not just not works but the program still works? 2017-11-06 05:57:15 <_ikke_> the y? 2017-11-06 05:57:38 they* 2017-11-06 05:57:53 sorry for my english 2017-11-06 05:57:59 <_ikke_> no problem 2017-11-06 05:59:13 <_ikke_> aGaTHoS: Does surf require dmenu to be installed during building for it to support it? It does not have any libraries 2017-11-06 05:59:17 i want the package install minimal things to work, as i do with megatools i think i can put all dependencies i could as pakage-dev in makedepends so thies packages haven not to be installed at runtime right? 2017-11-06 06:00:13 dont know dont think so, is just a external progam it uses to enter an adress as is a very minimal web brower, without ir you just can open a website like in the command line you invoke surf 2017-11-06 06:00:24 <_ikke_> Yeah, if it's not a hard dependency, then it should not be in depends, correct. 2017-11-06 06:00:32 but i dont really think mpv has to be a needed dependency to play videos... 2017-11-06 06:00:34 <_ikke_> aGaTHoS: then I don't think it belongs in makedepends either 2017-11-06 06:00:40 is also a expternal progra 2017-11-06 06:00:42 m 2017-11-06 06:01:22 <_ikke_> I'm not sure how common it is, but you could use an install hook to notify about the optional dependencies (much like it happens in arch) 2017-11-06 06:03:01 i was refering to webkit2gtk and grc for makedepends... as in PKGBUILD of arch linux are set as depencencies 2017-11-06 06:03:06 hi 2017-11-06 06:03:10 alpine doesn’t support arch-style optional dependencies 2017-11-06 06:03:32 i mean probably they dont need to be as dependencies is i can put in makedepends and be compiled statically... 2017-11-06 06:03:43 abuild will figure out what the runtime dependencies are 2017-11-06 06:03:47 yes 2017-11-06 06:05:46 ok, and for optional i like this you said , warn the user in the installation hook that could also install mpv dmenu etc... to access this features, have to research where i can put the text how works the post_install trigger i suppose that exists 2017-11-06 06:07:25 such is a policy violation 2017-11-06 06:07:33 the package must have a static set of dependencies 2017-11-06 06:08:47 unless you mean program A invokes program B 2017-11-06 06:09:17 for example the way audacious is packaged in arch would be a policy violation in alpine 2017-11-06 06:11:59 <_ikke_> What policy in this case? 2017-11-06 06:13:22 <_ikke_> https://git.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/audacious 2017-11-06 06:13:52 in alpine a package must declare every shared library it depends on 2017-11-06 06:14:00 in arch this is not the case, they have optional dependencies 2017-11-06 06:14:37 no i mean this dependencies are really optional 2017-11-06 06:14:38 if you dont install dmenu or mpv you just can enter an address or play a video but can still surf the web 2017-11-06 06:14:38 are external programs 2017-11-06 06:14:38 surf just calls this programs 2017-11-06 06:14:38 you can install or not 2017-11-06 06:14:38 so I mean I put a message after installing warning the user that could install this packges also to get omre functionallity 2017-11-06 06:14:38 the program is gonna work even if you dont install them 2017-11-06 06:14:39 <_ikke_> But it does work when you don't have unzip installed, right? 2017-11-06 06:14:41 it’s -plug-ins 2017-11-06 06:14:43 that has the problem :p 2017-11-06 06:14:50 <_ikke_> kaniini: ah ok 2017-11-06 06:15:56 on arch if you start audacious 2017-11-06 06:16:12 you are greeted with error boxes 2017-11-06 06:16:14 for missing libraries 2017-11-06 06:16:58 <_ikke_> d'oh 2017-11-06 06:17:04 <_ikke_> Yeah, that should not be optdepends then 2017-11-06 06:17:07 <_ikke_> even in arch 2017-11-06 06:17:28 :D 2017-11-06 06:17:44 i think they solve it by patching out the error boxes now 2017-11-06 06:18:23 “solve” 2017-11-06 06:19:21 <_ikke_> heh 2017-11-06 08:14:10 How do we go about becoming an official mirror for alpine linux? 2017-11-06 08:16:35 <_ikke_> Perhaps by sending an email to alpine-infra@lists.alpinelinux.org 2017-11-06 08:18:35 _ikke_: thanks 2017-11-06 09:53:29 ncopa, with some lag, I've done it: https://bugs.alpinelinux.org/issues/8108 2017-11-06 09:54:19 markand: thanks! 2017-11-06 09:54:33 :) 2017-11-06 12:00:13 hello, guys I am trying to use python regex with nonascii characters, on ubuntu works fine on alpine never matches? any ideas? 2017-11-06 12:00:47 ondrysak: is it utf8? 2017-11-06 12:01:47 ncopa, should be i am doing something like this 2017-11-06 12:02:09 pattern = regex.compile("[čřžýáí]*") 2017-11-06 12:03:50 i am explicitly setting coding to utf-8 in my code 2017-11-06 15:57:03 Is there any way to get bash-4.3 on alpine edge? 2017-11-06 15:57:48 use pinning and pull it in from a repo which has it 2017-11-06 15:58:25 <_ikke_> The repos only have 4.4 2017-11-06 15:58:29 <_ikke_> or 4.3 2017-11-06 15:58:40 <_ikke_> on x86_64 2017-11-06 15:59:36 I'm running into the bash hang bug with <() and am looking for ways around it. downgrading seemed like a valid option but adding from the stable release shows conflicts. 2017-11-06 16:02:48 Unode: echo '@v3.6 http://dl-cdn.alpinelinux.org/alpine/v3.6/main'; apk add --update bash@v3.6 2017-11-06 16:09:27 <_ikke_> kaniini: you forgot a part 2017-11-06 16:09:37 <_ikke_> >>/etc/apk/repositories 2017-11-06 16:10:02 righto 2017-11-06 16:10:03 it's monday 2017-11-06 16:10:04 it's morning 2017-11-06 16:10:07 be glad i haven't blown my brains out 2017-11-06 16:10:27 that's usually where i'm at on a monday morning :D 2017-11-06 16:11:20 i'm not 100% serious there 2017-11-06 16:11:22 <_ikke_> heh 2017-11-06 16:11:37 but monday mornings = usually nowhere near A game 2017-11-06 16:25:55 kaniini: doesn't work. Reports conflicts: http://dpaste.com/1N1KREV 2017-11-06 16:28:17 humm 2017-11-06 16:28:35 well an alternative would be to build yourself a bash .apk from the 3.6 sources against 3.7 2017-11-06 16:41:56 i have decided to switch to alpine ... does anyone know if its possible to accomplish this https://pastebin.com/LghDyHcR in an /etc/default/grub? 2017-11-06 16:42:15 for an encrypted boot 2017-11-06 16:42:22 kaniini: does our grub have cryptodisk enabled? 2017-11-06 16:43:14 i am currently on void ... and i like it ... but i feel apline is more in-line with my needs 2017-11-06 16:43:54 metachr0n: cryptdevice is ignored in alpine grub 2017-11-06 16:44:03 and from what i've tried (several distros over the past month) ... alpine is the only binary distro that gives me about what i would have with configuring a gentoo or funtoo system 2017-11-06 16:44:17 ok .. so thats fine ... i will put my boot on a usb device 2017-11-06 16:45:06 however ... if someone has physical access its total access anyway 2017-11-06 16:45:20 so ... in the end ... i'm really not worried about the "evil maid" scenario 2017-11-06 16:45:22 it works but will need type passphrase twice (no hook which could deal with it) 2017-11-06 16:45:45 ok 2017-11-06 16:45:54 thanks everyone 2017-11-06 16:46:41 madknight: just add to kernel command line: cryptroot=/dev/sda1 cryptdm=lvmcrypt 2017-11-06 16:46:59 MY_R: cool 2017-11-06 16:47:24 i don't mind entering the password twice ... because its on a yubikey 2017-11-06 16:47:25 ofc adjust to your config and will ask you for passphrase, ofc regenerate grub config after 2017-11-06 16:47:26 :D 2017-11-06 16:48:44 i'm happy with this 2017-11-06 16:49:06 also ... another question has anyone got xen usb passthrough working? 2017-11-06 16:50:02 i assume this should work as expected 2017-11-06 16:50:04 https://wiki.xenproject.org/wiki/Xen_USB_Passthrough 2017-11-06 16:50:28 and just in case, these two lines in /etc/default/grub 2017-11-06 16:50:31 GRUB_CRYPTODISK_ENABLE=y 2017-11-06 16:50:33 GRUB_ENABLE_CRYPTODISK=y 2017-11-06 16:51:00 ok ... i will set it up like this instead of boot on a usb 2017-11-06 16:53:22 ofc cryptdm=lvmcrypt comming from 'pvcreate /dev/mapper/lvmcrypt" 2017-11-06 17:08:50 also one last question 2017-11-06 17:08:56 i want to setup a xen dom0 2017-11-06 17:09:10 why apk wants to install a lot of dependencies for an aport without depends and just two makedepends? 2017-11-06 17:09:12 do i want to install the xen variant listed on the download page 2017-11-06 17:10:03 alpine-xen-3.6.2-x86_64.iso ? 2017-11-06 17:10:58 Hi all, in https://github.com/alpinelinux/aports/commit/d60bf515aff21d0ef3eb0cfa45d92d5ee725e755#diff-10767f8ad203c7e171910be31b9e558bL42 a loop that applied the patches was replaced by 'default_prepare' and 'update_config_sub'. Do these 2 instructions also take care of applying all patches? 2017-11-06 17:11:40 default_prepare it seems does aply patches 2017-11-06 17:12:54 The reason I'm asking is: https://bugs.alpinelinux.org/issues/8098 , in particular the part about "bash --version" returning "version 4.4.0(1)-release" when if all patches were applied it should read 4.4.12 2017-11-06 17:14:57 but in the file unpatched version is 4.3.48 not 4.4.0... 2017-11-06 17:16:23 aGaTHoS: where do you see 4.3? 2017-11-06 17:16:39 https://github.com/alpinelinux/aports/blob/master/main/bash/APKBUILD 2017-11-06 17:16:50 in top of the link you have posted 2017-11-06 17:16:56 the first patch 2017-11-06 17:17:10 pkgname=bash 2017-11-06 17:17:12 -pkgver=4.3.48 2017-11-06 17:17:14 +pkgver=4.4.12 2017-11-06 17:17:21 yeah so, 4.4.12 2017-11-06 17:17:48 but you say bash version says "version 4.4.0(1)-release" 2017-11-06 17:17:58 not 4.4.12 2017-11-06 17:18:05 but also not 4.3.48 2017-11-06 17:18:08 so.... 2017-11-06 17:18:22 the patch is aplied why not? 2017-11-06 17:18:33 so, the base 4.4.0 is used and the patch isn't applied. That's my interpretation. 2017-11-06 17:18:42 that 4.4.0 comes from another location 2017-11-06 17:19:35 then I don't understand. If the APK builds 4.4.12 directly, what's the purpose of including all the incremental diffs from 4.4.0 ? 2017-11-06 17:19:40 is but is this really a patch? 2017-11-06 17:19:42 is not 2017-11-06 17:19:52 is the diff of previuos APKBUILD and new.. 2017-11-06 17:20:28 your link is the info of a commit 2017-11-06 17:20:28 aGaTHoS: the patches are from https://ftp.gnu.org/gnu/bash/bash-4.4-patches/ 2017-11-06 17:20:45 in the commit it shows the diff 2017-11-06 17:20:54 https://github.com/alpinelinux/aports/blob/master/main/bash/APKBUILD <- is the same 2017-11-06 17:20:57 of the previous APKBUILD and new one 2017-11-06 17:21:03 you cant patch the APKBUILD 2017-11-06 17:21:24 aGaTHoS: I don't understand where you are going 2017-11-06 17:22:00 I know that I was showing a diff. The point was, if the changes introduced with that commit could have bypassed the application of the patches. 2017-11-06 17:23:13 The latest master includes the same code. And given alpine@edge gives bash 4.4.0 I wonder if the entire bug-report is just apk actually building bash 4.4.0 instead of 4.4.12. 2017-11-06 17:23:26 i dont think so 2017-11-06 17:23:36 but I'm not familiar with the APKBUILD process so I can't tell for sure. 2017-11-06 17:23:52 are you sure you are installing bash from testing and not main? 2017-11-06 17:24:24 aGaTHoS: I do get 4.4 and not 4.3 . And yes. Double-checked on a docker container alpine:edge. 2017-11-06 17:24:30 prepare has prepare_default and this function aplies the patches... 2017-11-06 17:25:25 wel then the error if there is , should be another but i think patches are aplied 2017-11-06 17:25:27 are you sure about that? 2017-11-06 17:25:43 yes because ive recently write an aport and deal with this 2017-11-06 17:26:11 the default prepare function is call prepare_default || return 0 2017-11-06 17:26:22 the diffs set "#define PATCHLEVEL 11" in the code. At this point I'm guessing but I suppose that's what the compilation uses to increment bash's version. 2017-11-06 17:26:27 and prepare_default is the one function that aplies patches 2017-11-06 17:26:45 aGaTHoS: could you perhaps guide me through using this APKBUILD to compile this locally? 2017-11-06 17:26:57 i.e. which commands need to be executed 2017-11-06 17:27:01 but then this are not alpine patches 2017-11-06 17:27:16 but internal bash patches they aply in their installation process 2017-11-06 17:28:09 im still learning too to make this things.. 2017-11-06 17:29:06 what have you done? 2017-11-06 17:29:38 ok I'm going to try to compile bash locally and check if the patches increment the version. If they do, then I'm confident that the APKBUILD I linked doesn't apply the patches. 2017-11-06 17:30:22 but what you have posted is a commit in main branch... 2017-11-06 17:30:36 how have you get to this 2017-11-06 17:30:45 what have you done what is your concern? 2017-11-06 17:31:05 ok you put bash --version 2017-11-06 17:31:10 aGaTHoS: Can you read https://bugs.alpinelinux.org/issues/8098 ? 2017-11-06 17:31:49 My guess is that the <() issue was a bug present in bash 4.4.0 that got fixed (I cannot reproduce it locally on bash-4.4.12 - not running on alpine linux). 2017-11-06 17:32:16 If the patches aren't being applied then alpine:edge is actually shipping bash 4.4.0 despite calling it 4.4.12 2017-11-06 17:32:26 ok 2017-11-06 17:32:51 and the commit you posted is just because you see a chngue in version 2017-11-06 17:32:53 ok and at this stage I'm pretty confident that the patches are *not* being applied. 2017-11-06 17:33:24 or the rest of the patches do something to solve that problem of hanging the shell? 2017-11-06 17:33:35 aGaTHoS: the commit replaced what was an explicit loop calling "patch -p0 < patch" by some functions. If the functions don't apply the patches then that explains the result. 2017-11-06 17:33:53 ah... then yes 2017-11-06 17:33:54 aGaTHoS: the patches are what actually increments bash 4.4.0 to 4.4.12 2017-11-06 17:33:58 the patches is not aplied 2017-11-06 17:34:05 as prepare_default needs p1 2017-11-06 17:34:35 if the patches are prepared for -p0 you should aply manually 2017-11-06 17:35:02 the doc os alpine says the abuild system aplies automatically patches prepared as -p1 2017-11-06 17:35:16 https://ftp.gnu.org/gnu/bash/bash-4.4-patches/bash44-001 here's the first patch. can you tell if it's -p1 or -p0 ? 2017-11-06 17:35:27 for instance i have to modify a patch of a pkgbuild from archlinux 2017-11-06 17:35:33 prepared for -p0 2017-11-06 17:36:05 modify the lines where the files where named as blabla.orig blabla.h etc with dira/blabla.h dirb/blabla.h 2017-11-06 17:36:52 in the same file it tels to you: Patch (apply with `patch -p0'): 2017-11-06 17:37:42 you have to either aply the patch in prepare fuction manually or modify the patch 2017-11-06 17:37:48 first unpack 2017-11-06 17:37:53 see what there is in 2017-11-06 17:38:07 and figure out how to rewrite the patch 2017-11-06 17:38:42 for instance 2017-11-06 17:38:44 *** ../bash-4.4/lib/readline/history.c 2015-12-28 13:50:31.000000000 -0500 2017-11-06 17:38:46 --- lib/readline/history.c 2016-09-30 14:28:40.000000000 -0400 2017-11-06 17:38:51 if whe sources unpacked 2017-11-06 17:38:58 bash-4.4 ... 2017-11-06 17:39:07 is the dir of patches or something 2017-11-06 17:39:14 and lib is the new version files 2017-11-06 17:39:31 then you modify the patch to begin the second line as ./lib/readline... 2017-11-06 17:39:55 nop sorry 2017-11-06 17:40:20 as bash-4.4 is a dir out 2017-11-06 17:40:25 i mean in the parent 2017-11-06 17:42:33 well if you are using 2017-11-06 17:42:34 aGaTHoS: shouldn't the build fail then if a patch is included and the patch doesn't apply? 2017-11-06 17:42:37 bash from main 2017-11-06 17:42:45 you have to watch here : https://github.com/alpinelinux/aports/blob/master/main/bash/APKBUILD 2017-11-06 17:42:54 the actual APKBUILD 2017-11-06 17:43:36 source="http://ftp.gnu.org/gnu/bash/bash-${_myver}.tar.gz 2017-11-06 17:43:42 pkgver=4.4.12 2017-11-06 17:43:57 its not possible to bash --version return 4.0.0 2017-11-06 17:44:07 mean 4.4..0 2017-11-06 17:44:18 pkgver is just the name that is given to the package. Not what is downloaded 2017-11-06 17:44:28 see _myver 2017-11-06 17:44:31 of course is the source you download 2017-11-06 17:44:35 no it's not 2017-11-06 17:44:38 _myver=${pkgver%.*} 2017-11-06 17:44:47 do you know what %.* does ? 2017-11-06 17:44:52 what? 2017-11-06 17:44:59 makes 4.4.12 -> 4.4 2017-11-06 17:45:07 ah.. 2017-11-06 17:45:15 then this is the problem 2017-11-06 17:45:21 no, that's not the problem 2017-11-06 17:45:28 the problem is that the patches are not being applied 2017-11-06 17:45:42 which ones 2017-11-06 17:45:45 all of them 2017-11-06 17:45:50 i only see one 2017-11-06 17:45:55 all the 12 bash patches 2017-11-06 17:45:56 at least alpine one 2017-11-06 17:46:05 https://github.com/alpinelinux/aports/blob/master/main/bash/APKBUILD#L70 2017-11-06 17:46:09 that means that this patches are aplied for all distros right? 2017-11-06 17:46:12 ^ is the first of 12 2017-11-06 17:46:16 scroll to the right 2017-11-06 17:46:19 are bash patches from themselves 2017-11-06 17:46:34 aGaTHoS: depends, I don't know what other distros do. 2017-11-06 17:47:01 dont understand 2017-11-06 17:47:15 What other distros do is not relevant 2017-11-06 17:47:20 why so much checksums if in sources are only specified 2 files 2017-11-06 17:47:44 patches are applied incrementally 2017-11-06 17:47:47 what do you mean 2 files? there's 14 files in total. 2017-11-06 17:47:49 for bash 2017-11-06 17:47:53 it's how bash distributes the patch release 2017-11-06 17:48:00 i mean maybe this patches are bash patches not alpine patches 2017-11-06 17:48:02 so if you want 4.4.12 2017-11-06 17:48:11 you have bash 4.4 tarball 2017-11-06 17:48:17 plus the 12 patches after it 2017-11-06 17:48:22 thats why you download bash.4.4.0 sources and for having version 4.4.12 you have to aply patches 2017-11-06 17:48:29 but because of bash not alpine 2017-11-06 17:48:46 which give you 4.4.12 2017-11-06 17:48:48 yes, correct 2017-11-06 17:48:51 kaniini: can you confirm if indeed the patches are being skipped? And shouldn't this trigger an error on APKBUILD? 2017-11-06 17:49:23 patches shouldn't be skipped 2017-11-06 17:50:18 kaniini: I just tried to manually download the 4.4 tarball. Compiled it, ran bash --version and got 4.4.0. Then applied the first patch and got 4.4.1. alpine@edge has bash 4.4.12 but bash --version gives 4.4.0 2017-11-06 17:50:40 and the <() bug seems to have been solved in 4.4.3 2017-11-06 17:50:55 4.4.2 actually https://ftp.gnu.org/gnu/bash/bash-4.4-patches/bash44-002 2017-11-06 17:52:50 then patches are not aplied ... as prepare_default aplies only external patches 2017-11-06 17:53:29 i guess 2017-11-06 17:53:53 hmmmmmm 2017-11-06 17:54:06 seems the APKBUILD is busted then 2017-11-06 17:55:31 and i ask myself is alpine aplying some aplication patch (diferrent than external patches) there are many aplication that distribute sources as a version base plus patches... and they are never aplied? 2017-11-06 17:56:14 that was my question. If this is the case, then there's actually a bug in the patching code that can lead to silently skipping patches. 2017-11-06 17:56:24 is responsability of the packager to aply this patches or the configure make etc is what aplies patches? 2017-11-06 17:56:27 This could have a significant impact. 2017-11-06 18:03:02 ok i see now that there is code to download all patches 2017-11-06 18:04:02 it seems they are not aplied maybe because are in -p0 format? 2017-11-06 18:04:05 Unode: [master cec73eedb7] main/bash: fix applying vendor patches 2017-11-06 18:04:07 Unode: 4.4.12-r1 coming shortly 2017-11-06 18:04:33 what have you changued kaniini ? 2017-11-06 18:04:37 no 2017-11-06 18:04:59 they are not applied because they were kept separate on purpose 2017-11-06 18:05:29 default_prepare only applies patches that are -p1 and end in .patch 2017-11-06 18:05:45 ok 2017-11-06 18:07:01 hey help me with this ive write ana port for surf ive only put makedepends webkit2gtk-dev gcr-dev, it installs like +120 packages for compilation, then removes but when intall the package generated apk wants to install 16 more packages why?? 2017-11-06 18:07:19 where do this dependencies i have not specified come from? 2017-11-06 18:10:34 well i suppose they are at same time dependencies of webkit2gtk, but why is this being installed if i dont specify in depends just the dev package in makedepends 2017-11-06 18:10:37 ? 2017-11-06 18:29:06 use rootbld 2017-11-06 18:37:04 kaniini: thanks! 2017-11-06 18:37:13 i have to uninstall first abuild? 2017-11-06 18:37:33 as ive installed abuild-rootbld and apk info -L shows nothing 2017-11-06 18:37:43 kaniini: might want to close https://bugs.alpinelinux.org/issues/8098 too since it's the same issue. 2017-11-06 18:43:15 and also if i have already installed build-base i suspect lot of the packages installed in the build process come from there as it says intalling for build: build-base webkit2gtk-dev gcr-dev, and install packages like gettext etc... why ? if i already have it nstalled as apk version build-base says...? 2017-11-06 18:43:57 Unode: done 2017-11-06 18:44:53 well i supose the dependencies are automatically generated here: >>> surf*: Tracing dependencies... 2017-11-06 18:45:26 i have installed abuild-rootbld , rebuild and get same result 2017-11-06 18:45:50 should i speciy webkit2gtk as depend then? 2017-11-06 18:45:51 kaniini: GNU bash, version 4.4.12(1)-release (x86_64-alpine-linux-musl) all green. Thanks again 2017-11-06 18:48:13 in fact should i specify all dependencies encountered by abuild? that would be: libX11 libc.musl libgdk-3 libglib-2 libgobject-2 libgtk-3 libwebkit2gtk with the correct names, or what is the way i have to figure out which dependencies i must specify? 2017-11-06 18:49:57 or there is no need to specify in depends as it does automatically? 2017-11-06 19:17:53 too much dependencies for a lightweight web browser!! 2017-11-06 20:29:37 hi is there any python code to parse apkbuilds ? 2017-11-06 20:30:09 <_ikke_> apkbuilds are shell scripts 2017-11-06 20:51:22 ACTION scratches his head. kaniini oddly enough. Bash 4.4.12 still hangs on "cat <(ls)". I... dunno why. 2017-11-06 20:52:58 could blame it that I didn't have the package yet but that's no longer true... 2017-11-06 23:58:11 "apk fix " doesn't force a redownload of an apk before reinstallation? is that correct? 2017-11-07 06:21:41 We're interested in becoming one of alpine-linux's official mirror listed here: http://rsync.alpinelinux.org/alpine/MIRRORS.txt Who should we contact about it? 2017-11-07 06:23:35 kaniini: ^ 2017-11-07 06:34:09 i'm having deja vu 2017-11-07 06:34:11 something something 2017-11-07 06:34:13 00:14 < mikhael_k33hl> How do we go about becoming an official mirror for alpine linux? 2017-11-07 06:34:13 00:16 <@_ikke_> Perhaps by sending an email to alpine-infra@lists.alpinelinux.org 2017-11-07 06:48:29 lmao 2017-11-07 07:13:49 scv: so0rry about that, tried sending an email to that one, I'm just not sure if that mailing list is still active though 2017-11-07 07:40:42 <_ikke_> mikhael_k33hl: It's not a high volume list 2017-11-07 08:21:43 _ikke_: uhhm, i'm note entirely sure what you mean by that 2017-11-07 11:48:09 It there a way to do file compression in rar format on the arm architecture ? 2017-11-07 13:55:12 kaniini if you could reopen https://bugs.alpinelinux.org/issues/8098 . Seems like simply updating bash didn't solve the issue. It still hangs. :( 2017-11-07 13:56:56 looks like this one might be a bit tougher to fix. It only seems to affect alpine (musl c?). I can't reproduce on other systems. 2017-11-07 16:12:59 ncopa: does it make sense to ask for inclusion of sc1200 in i386 alpine builds? 2017-11-07 16:13:59 i mean filing a bug report 2017-11-07 16:14:47 already pushed it: https://github.com/alpinelinux/aports/commit/1760b393878c5a8b0eade96010cfe0c2f7829444#diff-b63892e5fd3d8c62e27169d5739b09fe 2017-11-07 16:17:26 oh 2017-11-07 16:17:38 that's cool. so next release it will be in there? 2017-11-08 00:41:40 what is alpine armhf built with? 2017-11-08 00:41:49 armv5-*-*-*eabi) _arch_configure="--with-arch=armv5te --with-tune=arm926ej-s --with-float=soft --with-abi=aapcs-linux";; 2017-11-08 00:41:51 armv6-*-*-*eabihf) _arch_configure="--with-arch=armv6zk --with-tune=arm1176jzf-s --with-fpu=vfp --with-float=hard --with-abi=aapcs-linux";; 2017-11-08 00:41:53 armv7-*-*-*eabihf) _arch_configure="--with-arch=armv7-a --with-tune=generic-armv7-a --with-fpu=vfpv3-d16 --with-float=hard --with-abi=aapcs-linux --with-mode=thumb";; 2017-11-08 00:41:55 one of those? 2017-11-08 09:45:27 Hi all. Why when I'm installing postgresql-client package, there is no /usr/bin/pg_dump file? 2017-11-08 09:46:17 <_ikke_> tarkus: https://pkgs.alpinelinux.org/contents?file=pg_dump&path=&name=&branch=v3.6&repo=&arch=x86_64 2017-11-08 09:47:00 @_ikke_ hm, I'm checking postgresql-client package's content, and /usr/bin/pg_dump is listed there.. 2017-11-08 09:47:45 <_ikke_> Only on edge 2017-11-08 09:47:48 <_ikke_> not on 3.6 2017-11-08 09:47:52 https://pkgs.alpinelinux.org/contents?branch=edge&name=postgresql-client&arch=x86_64&repo=main 2017-11-08 09:48:31 <_ikke_> So in edge, they moved those utils from the server package to the client package 2017-11-08 09:49:09 I'm on 3.5.2.. 2017-11-08 09:50:47 I see, I guess edge will be published as 3.7.. 2017-11-08 09:51:24 <_ikke_> yes 2017-11-08 10:19:30 having issues with the nginx package 2017-11-08 10:19:41 when trying to upload, getting [crit] 183#183: *1509 open() "/var/tmp/nginx/client_body/0000000051" failed (13: Permission denied), client: 172.17.0.1, server: localhost, request: "POST /wp-admin/async-upload.php HTTP/1.1"... 2017-11-08 10:20:10 nginx running as user www, folder owned by www 2017-11-08 10:24:45 ok, changing the ownership of /var/tmp/nginx from nginx:nginx to www:root fixed it 2017-11-08 17:16:21 I just found i had an alpine mirror at home 2017-11-08 17:16:30 forgot about it :) 2017-11-08 17:52:06 I seem to have a little problem again, this time trying to design a way to handle software configuration in various installed Alpine systems 2017-11-08 17:53:40 took some time to figure out how apk works with conf files in various situations (modified vs original configs) etc. and it makes sense, however, configuration files owned by various packages present some challenges 2017-11-08 17:55:11 obviously packages requiring config files usually provide them, and if changes are needed, they're made to those files; and after changes have been made, apk no longer overwrites them 2017-11-08 17:57:24 but, should they absolutely require to be overwritten or modified at package update time... quickly thinking I can come up with maybe three ways to do it, but I can't easily figure out what the correct strategy would be 2017-11-08 18:50:20 Hi, I've some issues when trying to install Alpine linux to Virtual box 2017-11-08 18:50:24 could someone help me ? 2017-11-08 18:53:27 just post your problem. someone who knows what's wrong will answer 2017-11-08 18:54:54 When trying to install version 3.6.2 from virtual iso. and when selecting install mode (sys), I get this error : ERROR: unsastifiable constraints: sfdisk (missing): syslinux (missing): required by: world[syslinux] 2017-11-08 19:13:03 kbeaugrand: either use an online repository or a more complete iso 2017-11-08 19:21:38 exit 2017-11-08 20:26:40 Does anyone want to help me installing alpine on an arm processor? 2017-11-08 20:29:58 like... the files for "Generic Arm" on the download page are just .tar.gz files with no ISO 2017-11-08 20:30:45 <_ikke_> What kind of ARM device? 2017-11-08 20:31:17 ARM Cortex A7 running on a Banana Pi 2017-11-08 20:31:35 it has to boot from an SD card 2017-11-08 20:32:04 <_ikke_> I think for those devices it only supports run-from-ram 2017-11-08 20:33:12 <_ikke_> https://wiki.alpinelinux.org/wiki/Alpine_on_ARM has some info 2017-11-08 20:36:52 Hmm, it has some info. I'm just confused as to what I actually put on the SD card 2017-11-08 20:37:24 <_ikke_> Not entirely sure, but I think you just need to add a single partition on it, and copy those files on it (just like you would boot from USB) 2017-11-08 20:37:35 <_ikke_> Though, not sure if you need to do something with a bootloader 2017-11-08 20:44:36 Arm devices use Das U-boot as the bootloader, which is confusing enough. But I'm just not sure what alpine boots from exactly... other OS's I just burn an iso to it 2017-11-08 20:45:12 <_ikke_> spacedoutstoner: You just need to untar the files from the tar file onto the sdcard 2017-11-08 20:45:22 <_ikke_> spacedoutstoner: it's a run-from-ram installation 2017-11-08 20:48:07 <_ikke_> https://wiki.alpinelinux.org/wiki/DIY_Fully_working_Alpine_Linux_for_Allwinner_and_Other_ARM_SOCs#How_to_burn_to_SD_Card 2017-11-08 20:48:40 Oh really? hmm I didn't think that would work 2017-11-08 20:49:09 would I have to add the files to a partition on the SD card or leave it blank> 2017-11-08 20:49:47 <_ikke_> That link I pasted has some details 2017-11-08 20:52:39 yeah... that's more of a tutorial on compiling everything from source 2017-11-08 20:52:57 <_ikke_> Correct, but it also has details on how to create the sdcard 2017-11-08 20:53:31 <_ikke_> I don't have experience with uboot, so I cannot say exactly what steps are necessary 2017-11-08 20:57:34 <_ikke_> spacedoutstoner: note that you will need to get the correct bootloader for your system and write it to the sdcard as well 2017-11-08 21:02:27 word on the street is openrc is on the way out 2017-11-08 21:02:29 is that true? 2017-11-08 21:04:30 is the kodi package supposed to work? 2017-11-08 21:04:45 it crashes for me if run under gdb, hangs if not 2017-11-08 21:09:21 *sigh* it has an idiotic race condition in use of pthread_create 2017-11-08 21:09:47 https://github.com/xbmc/xbmc/blob/2966375f853bb01c3be571a892190f4a128c224b/xbmc/threads/platform/pthreads/ThreadImpl.cpp#L49 2017-11-08 21:11:05 m_ThreadId (a class member) is set via pthread_create which sets it "upon successful completion", not before the new thread starts 2017-11-08 21:11:33 and the new thread accesses its own m_ThreadId member, resulting in a data race 2017-11-08 21:17:01 hey guys, quick question. running alpine in sys mode on virtualbox. i do have xf86-video-vmware installed but when I try to use it in Xorg.conf it seems to not detect the screen and close immediately 2017-11-08 21:18:01 why are you using bad virtualization software 2017-11-08 21:18:14 previously, the answer to that problem would've been three letters 2017-11-08 21:18:29 a Q, an X and an L 2017-11-08 21:18:47 virtualbox doesn’t emulate vmware graphics lol 2017-11-08 21:18:58 might still work; I don't know, I no longer do virtual installs 2017-11-08 21:19:19 kaniini: the wiki says to use the vmware driver for virtualbox installs 2017-11-08 21:19:39 SirCmpwn: i dont have a license for workstation and hyper-v isn't an option 2017-11-08 21:19:50 qemu is a thing 2017-11-08 21:19:58 you just named more bad virtualization software 2017-11-08 21:20:43 ... which you then continued doing :D 2017-11-08 21:20:58 qemu is better than everything he just said 2017-11-08 21:21:00 to each his own I guess; needs vary 2017-11-08 21:21:01 though it has flaws 2017-11-08 21:21:55 currently the best known virtualization platform in existence? 2017-11-08 21:22:07 judging by the news of the last two days, that would be Minix :D 2017-11-08 21:22:16 <_ikke_> lol 2017-11-08 21:22:19 since when is minix a virtualization platform 2017-11-08 21:22:41 ever since it's been underneath all your operating systems running on an Intel CPU I suppose 2017-11-08 21:22:52 it runs in a secondary CPU 2017-11-08 21:22:54 it's not a hypervisor 2017-11-08 21:22:55 don't really know the details that deeply 2017-11-08 21:23:00 it's a microkernel operating system 2017-11-08 21:23:10 has nothing to do with virtualization 2017-11-08 21:23:31 should find the time to read a deeper analysis on the Management Engine this weekend 2017-11-08 21:24:46 but you could very well be right about that 2017-11-08 21:25:02 I am right on that 2017-11-08 21:25:10 http://minix3.org/ 2017-11-08 21:25:21 minix is to virtualization as linux is to virtualization 2017-11-08 21:25:25 that is: unrelated 2017-11-08 21:25:31 anyway 2017-11-08 21:25:43 when is linode offering alpine kvm’s 2017-11-08 21:25:54 that’s the real question 2017-11-08 21:26:22 you can install anything you want on linodes, I have one running a hobby distro 2017-11-08 21:26:24 SirCmpwn: ah well, you never know what Intel might have used it for, but I guess we'll get more details very soon 2017-11-08 21:26:56 i want single click alpine vms man 2017-11-08 21:27:03 ACTION shrugs 2017-11-08 21:27:16 haven't been following minix enough to know what, if anything, intel has contributed back to the project 2017-11-08 21:27:29 if i have to do it myself 2017-11-08 21:27:35 I just moved out of linode and into a colo with 3 alpine 1Us (+2 others) 2017-11-08 21:27:39 i would just send a box to the colo 2017-11-08 21:28:03 see i want to harness the cloud 2017-11-08 21:28:07 for my prototyping needs 2017-11-08 21:28:15 said 3 machines are VM masters 2017-11-08 21:28:21 so it's kind of cloud 2017-11-08 21:41:00 ohmygosh I can't do this 2017-11-08 21:41:51 I have like 4 different wiki pages open all telling me stuff about compiling u-boot, and half the links are broken 2017-11-08 21:42:51 <_ikke_> can anyone give spacedoutstoner_ a nudge towards correct procedure for installing alpine on a banana pi? 2017-11-08 21:45:08 I just typed "sudo apk add ....", sudo is not available on alpine, so what am i actually looking for? 2017-11-08 21:45:18 doas? su -l? 2017-11-08 21:45:42 correction: sudo is available 2017-11-08 21:46:00 would be nice if doas was availble though 2017-11-08 21:46:10 so rephrase: whats the alpine way of doing it? 2017-11-08 21:46:48 I just use sudo 2017-11-08 21:47:16 nah 2017-11-08 21:47:19 i've read the manpage 2017-11-08 21:47:22 dont really wanna 2017-11-08 21:47:40 then port doas, you'd do me a favor too 2017-11-08 21:48:03 <_ikke_> Out of curiosity, what is better about doas? 2017-11-08 21:48:31 man sudoers 2017-11-08 21:48:35 grep despair 2017-11-08 21:48:36 ^ 2017-11-08 21:48:51 then man doas 2017-11-08 21:48:53 it fits in one screen 2017-11-08 21:49:07 btw, i just became an alpine user a few minutes ago 2017-11-08 21:49:34 and i really appreciate the i585 support 2017-11-08 21:50:43 actually, i upgraded from "casual user" to "regular user" 2017-11-08 21:51:16 also, i found that mpv on edge is sorta broken, but i figure thats why its named "edge" 2017-11-08 21:53:49 <_ikke_> Right, but they are going to release 3.7 soonish, so things should become quite stable soon 2017-11-08 21:54:15 3.6 works fine, too 2017-11-08 21:54:40 i didn't even bother to check for youtube-dl, since you sorta always need the git tip of it 2017-11-08 21:55:46 ACTION wondering how 'edge' will run on 'stable' lxc host, hmmm 2017-11-08 21:56:40 like, if there is an guest/border between it, whats the difference? 2017-11-08 21:56:48 *whats the matter? 2017-11-08 21:57:01 ye, I'm asking myself :D 2017-11-08 21:57:29 well... if can run debian, so why not edge pff 2017-11-08 22:00:25 MY_R: it shouldn't really matter 2017-11-08 22:00:57 hope so! 2017-11-09 02:59:17 Hi all. I just have a couple general questions: 2017-11-09 02:59:47 1. How many packages does apk support so far? 2017-11-09 03:03:09 2. Which unofficial port of grsec is Alpine using? 2017-11-09 06:48:43 are the setup scripts missing in "alpine-minirootfs-3.6.2" ARMHF tarball on purpose? 2017-11-09 07:07:54 ok, what is the purpose of that tarball? It has no init scripts at all, ,and the root login does not work....confused 2017-11-09 07:12:51 <_ikke_> jammers: I just replied in #alpine-devel 2017-11-09 09:31:21 Hi 2017-11-09 09:31:23 qemu 2017-11-09 09:31:38 mksh: qemu: not found 2017-11-09 09:31:39 why -sdl and -vga qxl do not work? 2017-11-09 09:31:56 huh 2017-11-09 09:32:09 ACTION sent a long message: cypa[m]_2017-11-09_09:32:09.txt  2017-11-09 09:32:17 cypa[m]: which version do you have there? 2017-11-09 09:32:38 minute 2017-11-09 09:33:15 $ qemu-system-arm -cpu cortex-a7 -machine cubieboard -m 2G -kernel openwrt-sunxi-uImage -spice ipv4,port=6 2017-11-09 09:33:15 001,password=gfhjkm -display sdl 2017-11-09 09:33:16 qemu-system-arm: -display sdl: SDL support is disabled 2017-11-09 09:33:25 3.6.2 2017-11-09 09:34:28 amd64 2017-11-09 09:35:50 hm, same arch & alpine version, cannot reproduce 2017-11-09 09:36:06 also, i just caught my lo interface not having any IP address 2017-11-09 09:36:12 is it qemu arm specific? 2017-11-09 09:37:33 cypa[m]: qemu-system-arm --version please 2017-11-09 09:38:15 $ qemu-system-arm --version 2017-11-09 09:38:16 QEMU emulator version 2.8.1 2017-11-09 09:38:58 my qemu has SDL support disabled at all 2017-11-09 09:39:12 indeed 2017-11-09 09:39:13 yours seems to have support for it, generally, even if it fails 2017-11-09 09:39:14 why? 2017-11-09 09:39:33 ? 2017-11-09 09:39:35 qemu-system-arm: -display sdl: SDL support is disabled 2017-11-09 09:40:23 $ qemu-system-arm -display sdl 2017-11-09 09:40:23 qemu-system-arm: -display sdl: SDL support is disabled 2017-11-09 09:41:09 you posted 'Could not initialize SDL(No available video device) - exiting' 2017-11-09 09:41:22 previously 2017-11-09 09:41:52 sorry, it was another server, debian actually ... my mistake 2017-11-09 16:36:05 Got a security vulnerability report from quay.io on an image running latest Alpine 3.6.2, the image had run `apk add --no-cache libgcrypt` which installs libgcrypt 1.7.9-r0, which has a known vulnerability CVE-2017-0379. I'm currently looking at how to install a newer version sans vulnerability, but thought this channel might appreciate the report. 2017-11-09 16:39:02 libgcrypt 1.8.1 resolves the vulnerability, but is unfortunately unavailable on http://dl-cdn.alpinelinux.org/alpine/v3.6/ 2017-11-09 16:41:46 <_ikke_> edge has 1.8.1 2017-11-09 16:43:38 Thank you @_ikke_. I also just found that the Alpine package repository requests CVE vulnerabilities to be reported as bugs at, so I am doing that now. I will see if I can resolve my issue locally by using the edge repo. 2017-11-09 16:43:53 s/at/at https://bugs.alpinelinux.org/projects/alpine/issues 2017-11-09 16:44:41 coryklein: http://bugs.alpinelinux.org/issues/7831 2017-11-09 16:45:15 coryklein: you should file a bug to quay.io. its a false positive 2017-11-09 16:46:10 the fix is included in 1.7.9 2017-11-09 16:46:22 https://git.gnupg.org/cgi-bin/gitweb.cgi?p=libgcrypt.git;a=shortlog;h=refs/tags/libgcrypt-1.7.9 2017-11-09 16:46:31 see the last commit before the 1.7.9 release 2017-11-09 16:46:44 https://git.gnupg.org/cgi-bin/gitweb.cgi?p=libgcrypt.git;a=commit;h=da780c8183cccc8f533c8ace8211ac2cb2bdee7b 2017-11-09 16:47:03 even the commit message says: CVE-id: CVE-2017-0379 2017-11-09 16:47:15 quay.io sources https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2017-0379, so I think the bug is in cve.mitre.org 2017-11-09 16:47:39 cve.mitre.org reads "Libgcrypt before 1.8.1 does not properly consider Curve25519 side-channel attacks" 2017-11-09 16:48:06 that was probably correct when the CVE was filed 2017-11-09 16:48:35 i dont think mitre follow up or investigate or anything 2017-11-09 16:49:00 its very common that the info there is incomplete 2017-11-09 16:49:37 in any case, we reported and fixed the issue 2 months ago 2017-11-09 16:49:59 so you're safe 2017-11-09 16:50:51 Let me repeat what I heard you say to make sure I understand correctly. 1.7.9 was patched after it had already been released, so there may exist libgcrypt versions in the wild that claim to be 1.7.9 that still have the vulnerability? 2017-11-09 16:51:15 Ah no wait 2017-11-09 16:51:45 When the CVE was filed, 1.8.1 may have been the latest with the fix, but 1.7.x was patched with the security fix, bringing it up to date. 2017-11-09 16:52:45 i have no clue what the people reported to mitre 2017-11-09 16:53:16 Ok 2017-11-09 16:53:33 my general feeling is that "security researchers" don't really care as long as they get their 15min fame 2017-11-09 16:53:53 But there was no 1.7.9 release that contained the vulnerability, do I understand that correctly? 2017-11-09 16:54:06 thats what i believe yes 2017-11-09 16:54:13 Okay thanks for your help. 2017-11-09 17:16:47 @ncopa Quay also reported the musl 1.1.16 vulnerability on the image, and I see it recorded on the bug tracker that you marked this resolved because 1.1.17, which doesn't have the vulnerability, is in the edge repository. Forgive me my ignorance, but why is the issue considered "fixed" when the vulnerability is still present in the latest release? Not trying to be confrontational, just curious to understand. 2017-11-09 18:50:58 Hello! I'm having a big problem and I'm kind of desperate 2017-11-09 18:51:31 I'm trying to boot alpine from an iso stored inside a VirtualBox VM 2017-11-09 18:51:51 I'm using alpine-virt 2017-11-09 18:52:12 but after days of try and re-try i still can't make it work 2017-11-09 18:52:45 is there somebody that can help me out? I'll be happy to explain everything in detail. Thank you! 2017-11-09 19:24:19 kmt-45_, please explain the problem, someone will help 2017-11-09 19:29:01 I wrote a server fault article 2017-11-09 19:29:02 https://serverfault.com/questions/882130/alpine-linux-boot-from-iso-stored-inside-an-hard-drive 2017-11-09 19:29:20 that can be more exhaustive 2017-11-09 19:30:22 the idea is to use GRUB to boot an iso stored inside the Hard drive of a VirtualBox VM 2017-11-09 19:30:39 I will have 2 partitions on the HD 2017-11-09 19:30:58 the first one is the `/boot` /dev/sda1 2017-11-09 19:31:15 and the second one is a ext4 "storage" partition 2017-11-09 19:32:55 `/boot` contains the isolinux files copied from the `alpine-virt-*.iso` maybe that's the problem 2017-11-09 19:33:46 anyways I tried to read everything on the internet and tried multiple solutions without any success, I wonder if I'm missing something 2017-11-09 19:33:57 (and I'm sure I am) 2017-11-09 19:38:57 hi there has alpine something like debians laptop mode 2017-11-09 19:39:07 laptop mode? i can't remember what that is 2017-11-09 19:39:30 m3tti: you mean laptop mode tools? 2017-11-09 19:39:48 yes 2017-11-09 19:40:11 well.. you could just search pkgs.alpinelinux.org, like i am, to answer your question 2017-11-09 19:40:17 anyway - there's no laptop mode tools in alpine 2017-11-09 19:40:31 yes that's why i was asknig 2017-11-09 19:40:40 you can create the package from https://github.com/rickysarraf/laptop-mode-tools or request it at bugs.alpinelinux.org (and wait!) 2017-11-09 19:40:42 is there something compareable 2017-11-09 19:40:52 no idea, nothing i can think of 2017-11-09 19:44:14 ok i'll take a look again into building an alpine package 2017-11-09 20:21:47 Hi there, I'm trying to install docker ce on alpine linux behind a proxy 2017-11-09 20:22:21 docker is installed but i cannot pull containers because system environments for http_proxy and https_proxy are not taken... 2017-11-09 20:22:34 could anyone say me where stands docker config files on aline ? 2017-11-09 20:41:35 Hi, any awnser ? 2017-11-09 23:44:04 Question, what's in your opinion 2017-11-09 23:44:29 the best way of booting alpine without installing it on the os 2017-11-09 23:44:48 I want to have a vm.. and be able to update the OS on the fly 2017-11-09 23:45:18 and being able to go back to the previous version if there's something wrong 2017-11-09 23:45:25 and an automated way 2017-11-09 23:56:32 kmt-45_: ramdisk, with apkovl files? 2017-11-10 00:02:01 @danieli 2017-11-10 00:02:04 danieli 2017-11-10 00:02:11 can you explain a little bit better 2017-11-10 00:02:48 I get the apkovl 2017-11-10 00:03:34 but I don't really know what to do to with the ramdisk 2017-11-10 00:04:55 I'm a tad preoccupied right now, I'm sorry 2017-11-10 00:05:15 use a ramdisk to run alpine without installing it 2017-11-10 00:05:19 apkovl for persistence 2017-11-10 00:06:34 I understand, ... so I create a ramdisk 2017-11-10 00:06:47 mount it and install alpine inside the ramdisk 2017-11-10 00:06:52 then boot from a ramdisk 2017-11-10 00:07:03 a ramdisk is a temporary disk, in memory 2017-11-10 00:07:09 https://www.jamescoyle.net/how-to/943-create-a-ram-disk-in-linux 2017-11-10 00:07:15 i believe there are some articles on this on the alpine wiki 2017-11-10 00:07:15 something similar to this 2017-11-10 00:07:29 you don't sound like you know your way around ramdisks or overlay filesystems 2017-11-10 00:07:33 i do warn you, alpine is meant for superusers 2017-11-10 00:07:53 if you don't (at least) know your way around google, it probably isn't a good fit 2017-11-10 00:09:54 danieli I'm just confused about what I should put inside the ramdisk 2017-11-10 00:10:18 a copy of the iso, or a full static installation of alpine 2017-11-10 00:10:59 can you explain me what a ramdisk is, and how you plan to use it? 2017-11-10 00:11:03 if not, this is probably a bad idea 2017-11-10 00:13:43 currently I'm doing something like this 2017-11-10 00:13:48 menuentry \"Test Menu Entry" --id menu_entry_1 { 2017-11-10 00:13:48 loopback loop (hd0,2)$isofile 2017-11-10 00:13:48 linux (hd0,1)/vmlinuz-virthardened iso-scan/filename=$isofile alpine_dev=sda2:ext4 modules=loop,squashfs,sd-mod,ext4 BOOT_IMAGE=/vmlinuz-virthardened 2017-11-10 00:13:50 initrd (hd0,1)/initramfs-virthardened 2017-11-10 00:13:52 } 2017-11-10 00:14:02 please use a pastebin for pasting code 2017-11-10 00:14:18 yes.. I've noticed the terrible mistake :S 2017-11-10 00:14:40 https://pastebin.com/d8W3FDTc 2017-11-10 00:15:08 where initrd initializes the ramdisk 2017-11-10 00:15:41 that is a mountpoint that stores everything in ram ... with tmpfs 2017-11-10 00:16:01 it's recreated on every boot 2017-11-10 00:16:07 yes sure 2017-11-10 00:16:17 for persistence, you can use an apkovl using alu 2017-11-10 00:16:32 yes.. In my first boot 2017-11-10 00:16:47 the iso starts and I do a first installation (I use packer) 2017-11-10 00:17:11 the VM boots .. i config stuff and I save an apkovl in one of the partitions 2017-11-10 00:17:15 of my vm 2017-11-10 00:17:35 the problem is the second boot... when grub is configured .. (with that menuentry) 2017-11-10 00:17:58 in that case it just breaks 2017-11-10 00:18:03 https://serverfault.com/questions/882130/alpine-linux-boot-from-iso-stored-inside-an-hard-drive 2017-11-10 00:18:38 that's the link of a serverfault I created with also some videos 2017-11-10 00:18:41 of the VM 2017-11-10 13:29:47 hm. 2017-11-10 13:30:13 i dont get ip addresses on the loopback interface at startup 2017-11-10 13:32:04 the /etc/network/interfaces definitely has the loopback interface tho 2017-11-10 13:34:37 ok after trying devuan i'm back to my good old alpine 2017-11-10 13:34:43 small fast awesowe !!! 2017-11-10 13:34:50 thanx for all the fish 2017-11-10 13:35:45 hmm. Why doesn't QubesOS base off of Alpine? 2017-11-10 13:36:07 I'd like to investigate what would be needed to get that done, and if viable, fund it 2017-11-10 13:36:45 not sure why so many security distros are using such massive attack surface bases 2017-11-10 13:37:38 sn0wmonster: Qubes is sorta an frontend for Xen virtualisation 2017-11-10 13:37:52 exactly 2017-11-10 13:38:16 but wouldn't it be lighter and faster to use Alpine for its base? 2017-11-10 13:38:27 what would that improve? 2017-11-10 13:38:29 (since every MB counts with virtualization) 2017-11-10 13:38:34 memory for one! 2017-11-10 13:38:54 but this is why I was curious what the benefits would be :) 2017-11-10 13:38:59 do you mean alpine for dom0 or the vms? 2017-11-10 13:39:03 dom0 2017-11-10 13:39:06 dom0 presumably lol 2017-11-10 13:39:31 other obvious benefit would be, no systemd 2017-11-10 13:41:32 hah 2017-11-10 13:42:07 didn't expect some distros to take security seriously to still use systemd 2017-11-10 13:42:46 sadly, both QubesOS and TailsOS do :| 2017-11-10 13:42:53 the excuse is always the same 2017-11-10 13:43:10 "too much work to remove, would require additional development when we can just ride on the main branch instead and focus" 2017-11-10 13:43:15 because nowadays, every shit depends on systemd 2017-11-10 13:43:26 it's a practical choice, but its like making CopperheadOS or LineageOS on a backdoored binary blob mobile device 2017-11-10 13:43:28 whats-the-point 2017-11-10 13:47:05 liwakura: every shit indeed, but you don't have to use it 2017-11-10 13:47:27 Gentoo provides OpenRC by default and patches stuff to work with it 2017-11-10 13:47:54 xterm -> xorg -> udev -> systemd 2017-11-10 13:47:54 sn0wmonster: and theres not too much work, can just take Gentoo's patches 2017-11-10 13:47:58 thank god there is eudev 2017-11-10 13:48:06 liwakura: I have all of those without systemd 2017-11-10 13:48:16 specing: i doubt you have udev 2017-11-10 13:48:49 gentoomen sacrificed their times to give us eudev 2017-11-10 13:49:03 to keep an non-systemd desktop still possible 2017-11-10 13:49:32 I have udev without systemd bastardisations 2017-11-10 13:49:46 the real udev 2017-11-10 13:49:55 just didn't upgrade? 2017-11-10 13:50:11 I did upgrade 2017-11-10 13:50:14 im calling on bullshit, proof pls. 2017-11-10 13:50:52 or tarball link pls. 2017-11-10 13:51:03 for gentoo without systemd? 2017-11-10 13:51:31 MY_R: udev without systemd is not possible, but thats why gentoo has the eudev fork 2017-11-10 13:52:03 and thats why every non-systemd desktop-able distro also uses gentoo's eudev 2017-11-10 13:52:04 he probably meant eudev 2017-11-10 13:52:54 i figure. 2017-11-10 13:53:54 package name is eudev but command name udevd so ye :P 2017-11-10 13:54:50 kernel binare is vmlinuz but its actually an bzImage 2017-11-10 13:55:30 :) 2017-11-10 13:55:57 also, i have yet to find static binaries in the /sbin folder 2017-11-10 13:56:03 and yet to find homedirs in /usr 2017-11-10 15:03:39 hi, does someone knows how to specify a DUID to udhcpc6? 2017-11-10 15:04:11 I cannot find in the RFC3315 which option is related to DUID 2017-11-10 15:06:37 I don't know if I can use the client-id (option 61) for that 2017-11-10 15:11:46 I understood some of those words. 2017-11-10 15:12:09 :p 2017-11-10 15:12:29 Sorry :( 2017-11-10 15:12:53 it's just an DHCPv6 (dhcp for ipv6) option dedicated to "authentication" with a token 2017-11-10 15:13:39 no worries 2017-11-10 15:13:50 ah, yea definitely no clue, but it's something I should learn 2017-11-10 15:15:31 Online, a french hosting provider, use DUID to request the routing of your ipv6 prefix to your host 2017-11-10 15:16:26 <_ikke_> anthony25: Isn't the dhcp client adding a DUID already? 2017-11-10 15:17:14 _ikke_: that's why I asked, I don't know if a DUID can be specified as the client-id (option 61) 2017-11-10 15:17:28 probably, so I'm going to test it :p 2017-11-10 15:18:10 <_ikke_> anthony25: as the DUID is part of the dhcpv6 spec, I would think the client should already provide it (it's the mac address equivalent of dhcpv4) 2017-11-10 15:18:13 and timeout to fallback here in ipv4, maybe :< 2017-11-10 15:19:24 <_ikke_> anthony25: RFC 3314 section 22.2 2017-11-10 15:19:33 <_ikke_> " 2017-11-10 15:19:35 <_ikke_> The Client Identifier option is used to carry a DUID " 2017-11-10 15:19:41 ooohhh 2017-11-10 15:19:43 neat 2017-11-10 15:19:47 _ikke_: thanks a lot! 2017-11-10 15:20:02 <_ikke_> RFC3315 ofcourse 2017-11-10 15:21:54 sorry, I looked for the info maybe too quickly (and badly) into this RFC… 2017-11-10 15:24:33 <_ikke_> I was first just looking for DUID, which gave many results. I saw there that it was part of the options, so as soon as I looked up the section about options, it was one of the first subsections 2017-11-10 15:27:07 l 2017-11-10 15:58:50 is there a way to just show what packages are available? 2017-11-10 15:59:02 like packages that need updates 2017-11-10 15:59:20 apk version 2017-11-10 16:00:43 thanks 2017-11-10 17:09:12 Howdy 2017-11-10 17:10:05 Any major updates made recently to edge? I've pulled a docker image alpine:edge and tried to apk upgrade in it, but get breakage 2017-11-10 17:10:50 Unsatisfiable constraints: /bin/sh (virtual) provided by ... baselayout stuff 2017-11-10 17:11:18 happens in alpine:latest, too 2017-11-10 17:17:32 nm, looks like I had a stale version of the edge image. Looks to be working after a pull. 2017-11-10 20:10:00 What compression format does .apk use? i'd like to extract it and investigate the files 2017-11-10 20:10:28 johnnyfive: gzip 2017-11-10 20:10:46 you can just unpack it with tar tho 2017-11-10 20:11:06 @liwakura, thanks! 2017-11-10 20:11:59 johnnyfive: there is an `file` command so you can easily check for yourself 2017-11-10 20:20:40 android-tools won't work 2017-11-10 20:20:50 ERROR: Could not find package 'so:libcrypto.so.41' in the aports folder and could not find it in any APKINDEX 2017-11-10 20:21:37 steamport: what repos do you have enabled? 2017-11-10 20:21:59 ((technically i'm using pmbootstrap, but this appears to be an alpine repo issue... i think it uses edge)) 2017-11-10 20:23:37 cant reproduce, i can install android-tools finely from edge 2017-11-10 20:24:36 weird 2017-11-10 20:25:13 do you have edge/testing in there? 2017-11-10 20:25:46 yeah 2017-11-10 20:25:56 edge/main, edge/community, and edge/testing 2017-11-10 20:26:20 libcrypto.so.41 is in libressl2.5-libcrypto 2017-11-10 20:26:49 oh well not in edge anymore 2017-11-10 20:29:00 steamport: which arch? 2017-11-10 20:29:05 armhf 2017-11-10 20:29:08 or x86_64 2017-11-10 20:29:10 not sure really 2017-11-10 20:29:18 steamport: uname -m 2017-11-10 20:29:27 im using a bootstrap program 2017-11-10 20:29:37 steamport: x86_64 2017-11-10 20:29:43 ye 2017-11-10 20:30:02 steamport: cat /etc/apk/arch 2017-11-10 20:30:24 x86 64 2017-11-10 20:33:09 liwakura: you probably have libressl2.5-libcrypto package in your syste, right? 2017-11-10 20:33:55 MY_R: yeah, on both 2017-11-10 20:34:21 i have one v3.6 and one edge 2017-11-10 20:34:29 but in edge this package is already gone, repalced by libressl2.6-libcrypto 2017-11-10 20:34:47 so you get this error if you /only/ have edge? 2017-11-10 20:35:12 because i dont have edge only, i have it optionally in addition to v3.6 2017-11-10 20:36:30 probably stuff is building now and just need to wait or bump 'android-tools' 2017-11-10 20:41:32 weird... 2.5 is in repository sooo 2017-11-10 20:41:47 steamport: install libressl2.5-libcrypto 2017-11-10 20:41:49 .. 2017-11-10 20:42:13 ok 2017-11-10 20:42:15 ACTION got problems with eyes :P 2017-11-10 20:43:03 so 2.5 will be removed when everything else will work with new 2.6 oook 2017-11-10 21:31:54 UHHH GUYS 2017-11-10 21:37:07 When do packages rebuild? 2017-11-10 21:38:39 <_ikke_> When they change? 2017-11-10 21:39:55 . . . 2017-11-10 21:40:13 alot of packages seem to be messed up on my side due to relying on older libressl 2017-11-10 21:40:34 <_ikke_> steamport: Might be that the rebuild is still running, ncopa said he was updating libressl 2017-11-10 21:40:40 k 2017-11-10 21:40:54 running edge will break shit 2017-11-10 21:41:57 ... once again, must remind, it's kinda frustrating to run into alpine-side problems while running pmbootstrap 2017-11-10 21:42:02 that's something I need to figure out actually; how to do something like edge without shit breaking 2017-11-10 21:42:28 well, you could have a "staging edge" or something like that 2017-11-10 21:42:28 steamport, well, the answer to that is obvious isn't it 2017-11-10 21:42:32 purely for development 2017-11-10 21:43:07 steamport: might be an idea for postmarket to target alpine stable (if it targets edge) 2017-11-10 21:44:15 from about one and a half years of experience I'd rather avoid tagged repositories completely, they're a constant source of headache and require full time jumping through flaming hoops 2017-11-10 21:44:35 ... at least the way I use them! 2017-11-10 21:44:55 (obvious, if indirect, disclaimer right there) 2017-11-10 21:45:16 <_ikke_> TBB: you mean pinned repo's? 2017-11-10 21:45:25 yup 2017-11-10 21:49:16 I've probably said this before, I have about 15 years of Gentoo under my belt with maybe 5 years professional experience in the RPM world... things don't get quite this complicated in either, although Gentoo certainly tries its best 2017-11-10 21:49:56 and by professional experience I mean packaging 2017-11-10 21:52:11 of course it's not a fair comparison, Alpine vs RedHat derivatives especially, the scale and the goals are very different 2017-11-10 21:52:35 the maturity as well, relatively seen 2017-11-10 21:59:17 yeah. while starting using a RH derivative can be quite an exercise in archaeology, (with the exception of point zero releases) it's also mature AF 2017-11-10 22:17:09 still alot of packages are BORKED 2017-11-10 22:17:10 on edge 2017-11-10 22:17:16 with libressl-2.6-libcrypto 2017-11-10 22:20:03 steamport: <@_ikke_> steamport: Might be that the rebuild is still running, ncopa said he was updating libressl 2017-11-10 22:20:08 oh ok 2017-11-10 22:20:30 read 2 more times if you want to ask the same thing - that's what I do 2017-11-10 22:21:43 does alpine track which packages were installed manually and which were installed as dependency? 2017-11-10 22:21:53 <_ikke_> "17:15 <@ncopa> i think i have pushed all libressl 2.6 rebuilds now 2017-11-10 22:22:00 hmm 2017-11-10 22:22:53 liwakura: I could be wrong about this, but /etc/apk/world might have just the toplevel packages, not necessarily deps 2017-11-10 22:23:22 TBB: seems so, right 2017-11-10 22:23:25 thanks 2017-11-10 22:23:27 I trust _ikke_ to correct me if I'm wrong :) 2017-11-10 22:24:09 (and just in case, in this crazy world of today where everything that can will be misunderstood... that was purely a compliment) 2017-11-10 22:26:25 <_ikke_> TBB: yeah, that sounds about right 2017-11-10 23:59:58 Hi, is there a particular reason why the "deny_new_usb" sysctl option from grsec isn't enabled in the alpine kernel? 2017-11-11 00:00:06 It's arguably useful on servers 2017-11-11 00:02:25 i can think of a lot of situations where it would be problematic for servers 2017-11-11 00:02:40 it should not be on by default 2017-11-11 00:02:51 just flip the sysctl if you want it enabled 2017-11-11 00:08:03 scv: It's not available at all 2017-11-11 00:08:13 ah 2017-11-11 00:09:04 Sorry, my question used the wrong terminology. I meant "available", not "enabled". 2017-11-11 00:23:38 If there is no particular reason, I'd like to suggest it is made available as a sysctl option in kernel packages. 2017-11-11 02:46:53 hello, I'd like to run Alpine from a USB thumbdrive, with a bunch of packages installed. I'm a bit confused about the best way to add those packages so they are included in the release on the USB, and don't need to be re-added on reboot. What is the correct way to add packages to a USB install? 2017-11-11 02:47:36 I have written the bootable ISO to a USB drive, do I need to boot that and then _install_ to the USB drive with setup-alpine? 2017-11-11 02:48:48 when I boot the installed on USB, I am unable to umount /media/usb, so i don't see how to do an install to that thumb drive 2017-11-11 03:13:16 dvetna: your installation media should be different to the target drive, of course 2017-11-11 03:13:58 yea, saying it out like that, it seems obvious 2017-11-11 03:19:33 when running setup-alpine, during the disk setup.. there is the question about how to use the disk.. you can choose either sys or data. For data, the OS runs from tmpfs. Here's my question: if you go that route, how would you customize the OS that runs from tmpfs? 2017-11-11 07:02:00 <_ikke_> dvetna: first, you need to make sure the usb disk is writable 2017-11-11 07:02:48 <_ikke_> dvetna: copying the iso does not work, because it's an iso9660 partition, which is read-only 2017-11-11 07:03:59 right, so what is the correct way to make a "live USB" that runs from tmpfs, with custom packages installed 2017-11-11 07:06:08 <_ikke_> dvetna: if you have alpine running already, setup-bootable 2017-11-11 07:07:53 <_ikke_> dvetna: otherwise it's a bit more involved 2017-11-11 07:08:36 ok, setup-bootable definitely looks helpful. If that uses an ISO for its source, how would you add additional packages? 2017-11-11 07:09:51 <_ikke_> dvetna: once you've booted from the usb and ran the setup, you can use lbu commit to store the changes you've made, like packags installed, config files changes 2017-11-11 07:11:26 OK, so a) boot into alpine somewhere; b) run setup-bootable with an ISO to setup bootable USB that runs from tmpfs; c) boot that new USB and customize with apk / etc; d) use lbu commit to save those changes to the USB? 2017-11-11 07:11:46 <_ikke_> yes 2017-11-11 07:12:04 great, I'll give that a shot, thanks for helping to set me straight here. 2017-11-11 15:39:22 _ikke_: testing out your suggestion to use `setup-bootable`, I am wondering if I'm doing it wrong. I have booted into a live/installer, then downloaded a release ISO, inserted another USB drive (no partitions on it), and used `setup-bootable release.iso /dev/sdd`, but I get: `mkdir: can't create directory '/media/sdd/.new': Read-only file system. 2017-11-11 15:43:20 <_ikke_> dvetna: did you follow something like https://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB#Creating_a_bootable_Alpine_Linux_USB_Stick_from_the_command_line / 2017-11-11 15:45:50 ah, ok.. I was confused about what setup-bootable was doing, that makes more sense now. thanks for the reference! 2017-11-11 15:46:27 <_ikke_> np 2017-11-11 15:47:45 <_ikke_> https://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB#Automated 2017-11-11 15:57:39 I see that setup-disk (when in sys mode), also writes a copy of the OS to the new partitions. How does that system differ from what `setup-bootable` sets up? I assume it's only the tmpfs detail 2017-11-11 21:33:54 oh well 2017-11-11 21:33:57 hi opal 2017-11-11 21:34:04 what 2017-11-11 21:34:33 ✿ what ❀ 2017-11-11 21:34:39 <_ikke_> .. 2017-11-11 21:34:44 how have you been doing? 2017-11-11 21:35:06 im actually surprised there are still channels you arent banned from 2017-11-11 21:35:06 ✿ fine, i have been using alpine for a while so decided to pop in ❀ 2017-11-11 21:35:11 ✿ im not ❀ 2017-11-11 21:35:11 oh! 2017-11-11 21:35:12 also switched 2017-11-11 21:35:22 ✿ im actually very tolerable ❀ 2017-11-11 21:35:32 lol 2017-11-11 21:35:33 seems legit 2017-11-11 21:35:56 ✿ might put alpine on my laptop ❀ 2017-11-11 21:36:06 ✿ cpu probably cant handle gentoo and all the compiling too well ❀ 2017-11-11 21:36:12 what DE do you intend to use? 2017-11-11 21:36:18 or WM? 2017-11-11 21:36:35 ✿ i keep it simple and stick with a wm lol... ive been using i3 but i may go back to awesomewm or something ❀ 2017-11-11 21:36:45 ✿ unless you know of any other good tiling wms ❀ 2017-11-11 21:36:51 im on i3 myself 2017-11-11 21:36:57 <_ikke_> What's with the unicode? 2017-11-11 21:37:07 ✿ it's 2017, unicode is standard ❀ 2017-11-11 21:37:11 _ikke_: flowers, for increased attention 2017-11-11 21:37:13 it's a flower, in case your client isn't set to use unicode 2017-11-11 21:37:14 ^ 2017-11-11 21:37:24 opal: it's very unnecessary though 2017-11-11 21:37:31 <_ikke_> I see the characters, just very small 2017-11-11 21:37:38 danieli: see? it works as attention catcher 2017-11-11 21:37:49 ✿ it only works as one because people want to point it out ❀ 2017-11-11 21:37:50 it really does 2017-11-11 21:38:03 opal: because it's a tad obnoxious, and very unnecessary 2017-11-11 21:38:10 ✿ so is this conversation about it ❀ 2017-11-11 21:38:10 opal: usually its rainbow colored, something up? 2017-11-11 21:38:16 ✿ +c ❀ 2017-11-11 21:38:25 ahh, nice 2017-11-11 21:38:25 ✿ oh youre that laintard arent you ❀ 2017-11-11 21:38:27 ✿ lol ❀ 2017-11-11 21:38:29 apropos that, let's keep this to #alpine-offtopic 2017-11-11 21:38:41 if you're going to insult people, you can leave the alpine channels. 2017-11-11 21:38:53 ✿ wasnt an insult? ❀ 2017-11-11 21:39:09 "that laintard" 2017-11-11 21:39:27 hah 2017-11-11 21:39:28 ✿ -tard and -fag are neutral modifiers the way i and the rest of imageboard users use them ❀ 2017-11-11 21:39:30 ✿ get over it ❀ 2017-11-11 21:39:36 this isn't an imageboard 2017-11-11 21:39:39 danieli: is this the first time you encounter opal? 2017-11-11 21:39:40 #alpine-offtopic, please 2017-11-11 21:39:46 ✿ theres 30 people in -ot ❀ 2017-11-11 21:39:47 liwakura: i've seem him before, but i'm kind of forgetful 2017-11-11 21:39:56 ✿ join #offtopic-offtopic ❀ 2017-11-11 21:40:08 > him 2017-11-11 21:40:11 them 2017-11-11 21:40:13 my bad 2017-11-11 21:40:17 bad habit 2017-11-11 21:40:26 ✿ lol ❀ 2017-11-11 21:40:29 <_ikke_> How can you be offtopic in an offtopic channel? 2017-11-11 21:40:36 :thinking: 2017-11-11 21:40:40 ✿ that's the question of the century ❀ 2017-11-11 21:40:45 by being on point, of course 2017-11-11 21:40:47 opal: weren't you like doing HRT? 2017-11-11 21:40:52 but that's beside the point 2017-11-11 21:40:52 ✿ i wish ❀ 2017-11-11 21:41:02 again - `/j #alpine-offtopic`, please 2017-11-11 21:41:06 ✿ i did ❀ 2017-11-11 21:41:15 move the convo there then 2017-11-11 21:41:21 ✿ waiting ❀ 2017-11-11 21:41:26 ✿ no= one else joined ❀ 2017-11-11 21:45:02 hey opal 2017-11-11 21:45:07 didnt you use to be on onionirc aswell? 2017-11-11 21:45:29 ✿ yeah but as danieli said you might want to take this to the offtoic chan ❀ 2017-11-11 21:45:34 hm sure 2017-11-11 21:48:54 opal: you just forgot to prolong lol 2017-11-11 21:49:04 ✿ to hwat ❀ 2017-11-11 21:49:09 oops 2017-11-11 21:49:14 ✿ :v ❀ 2017-11-11 21:49:14 wrong chan 2017-11-11 21:49:25 ✿ anyway yeah alpine ❀ 2017-11-11 21:49:41 ✿ seems like the sanest option compared to the systemds and the glibcs of the linux distro world ❀ 2017-11-11 21:49:47 ✿ its unintrusive ❀ 2017-11-11 21:49:49 ✿ lets me do what i want ❀ 2017-11-11 21:49:50 ✿ i like it ❀ 2017-11-11 21:50:02 glibc is quite bloated in my opinion, unfortunately a lot of patches are needed for some packages 2017-11-11 21:50:04 glibc-only extensions 2017-11-11 21:50:12 i was on musl void before, but craved for busybox 2017-11-11 21:50:32 its almost winter here 2017-11-11 21:50:35 ✿ i hate glibc because of its unicode handling ❀ 2017-11-11 21:50:37 get the flowers out 2017-11-11 21:50:45 ✿ it's also bloated but that's a secondary concern for me ❀ 2017-11-11 21:51:05 ✿ it's still autumn ❀ 2017-11-11 21:52:41 ✿ oh as far as window managers, ive gotten re-used to the windows way of alt-tabbing and using the taskbar ❀ 2017-11-11 21:52:52 ✿ taskbar groups and a system tray is nice ❀ 2017-11-11 21:53:04 ✿ i could probably get used to i3 or something again over time ❀ 2017-11-11 21:53:06 ✿ not sure ❀ 2017-11-12 05:40:46 hello? 2017-11-12 05:41:01 olleh 2017-11-12 05:41:18 ah goodie, someone is here 2017-11-12 05:41:51 I've become very interested in alpine after reading a bit about it, but there's 1 thing I can't find a clear answer for 2017-11-12 05:42:08 what's that 2017-11-12 05:42:12 is alpine init-agnostic? 2017-11-12 05:42:38 it says it uses "OpenRC by default", but that doesn't imply it can only use OpenRC 2017-11-12 05:43:06 it's openrc, but as it's not systemd it doesn't force 90% of the system to depend on it 2017-11-12 05:43:16 you could use another init 2017-11-12 05:43:30 ya, that's what I was hoping for. I noticed it has full on packages for runit and s6 2017-11-12 05:43:44 and it can be used perfectly fine for hosting servers? 2017-11-12 05:43:48 yep, it's ideal for that 2017-11-12 05:44:09 finally. every other distro I could find is hardlocked to systemd or something else 2017-11-12 05:44:58 it seems I would have to write the services myself for s6 or runit though, which is fine. if I have a base to work from I can do the rest of the work 2017-11-12 05:46:05 not a huge concern, the openrc files are quite simple 2017-11-12 05:46:35 I've used openrc before, but maintaining bash-like files is a bit of a hassle. s6 looks a lot less messy 2017-11-12 05:46:56 eh, if your service needs a lot of boilerplate stuff it is quite cumbersome 2017-11-12 05:47:08 you've used s6 before? 2017-11-12 05:47:14 nope 2017-11-12 05:47:22 openrc is sufficient for my needs 2017-11-12 05:47:41 it is for me as well, but I'm curious enough to try others. init freedom, you know? 2017-11-12 05:47:49 yep, good to have the option 2017-11-12 05:48:10 but that brings me to an issue I'm having getting it to work in a virtual machine. 2017-11-12 05:49:01 I've followed the directions on getting alpine working with qemu, as well as some videos, but I'm not getting an internet connection 2017-11-12 05:49:29 not sure what the deal is. I'm not quite used to the package manager yet 2017-11-12 05:50:33 what card did you emulate? 2017-11-12 05:50:35 or did you use virtio 2017-11-12 05:50:55 also, which alpine flavor 2017-11-12 05:50:58 oh, wait, I missed a section on networking. if you're curious, I'm using this wiki 2017-11-12 05:51:02 https://wiki.alpinelinux.org/wiki/Qemu 2017-11-12 05:51:16 ah user emulation 2017-11-12 05:51:30 i think it's an rtl by default 2017-11-12 05:51:33 so, which flavor of alpine? 2017-11-12 05:51:42 I'm using the iso optimized for virtual systems 2017-11-12 05:51:55 that would be why 2017-11-12 05:52:05 it only includes drivers for virtual interfaces 2017-11-12 05:52:28 so in qemu you'll need to adjust your nic settings 2017-11-12 05:52:28 1 sec 2017-11-12 05:52:32 oh. that would explain why >_>. should I just be using the standard iso? 2017-11-12 05:52:42 i typically use vanilla even in VMs 2017-11-12 05:52:59 but you can get this working if you just adjust your qemu settings 2017-11-12 05:53:01 which i recommend anyway 2017-11-12 05:53:06 using an emulated NIC is not great 2017-11-12 05:53:30 do you know how to setup a bridge on the host? 2017-11-12 05:54:30 I typically use a frontend to qemu called virt-manager 2017-11-12 05:54:58 I usually just run it through NAT, but from reading it appears a bridge is the best option for alpine? 2017-11-12 05:55:56 a bridge is just more convenient in general, it acts just as if you had it physically plugged in 2017-11-12 05:56:16 ah, I see 2017-11-12 05:56:30 so just use the standard iso then? 2017-11-12 05:57:02 I plan to use it for running servers at some point so I just want to give it a wirl 2017-11-12 05:58:28 you can do that 2017-11-12 05:58:35 but you still should use virtio 2017-11-12 05:58:40 rather than the emulated rtl or e1000 2017-11-12 05:59:05 not sure what you mean 2017-11-12 05:59:09 -netdev type=tap,id=net0,ifname=(unique interface name),vhost=on -device virtio-net-pci,mac=(interface virtual MAC),netdev=net0,id=net0 2017-11-12 05:59:21 by default qemu provides a virtual realtek nic for the VM 2017-11-12 05:59:27 emulating all the hardware mechanics is poitnless 2017-11-12 05:59:27 ah 2017-11-12 05:59:28 pointless* 2017-11-12 05:59:42 virtio is designed for virtual machines 2017-11-12 05:59:58 doesnt have any of that hardware emulation overhead 2017-11-12 06:01:45 right, so how would I use virtio then? 2017-11-12 06:02:55 the options i pasted above select the driver, would replace the -net nic -net user example from the wiki 2017-11-12 06:03:17 just run it through qemu? 2017-11-12 06:04:51 you'd need to specify a MAC for the VM as well as an interface name 2017-11-12 06:05:01 usually tap#, i.e. tap0 tap1 etc 2017-11-12 06:05:13 then once the vm boots you can bridge your physical interface with the tap interface 2017-11-12 06:07:38 so interface name would be tap0, but what about MAC? 2017-11-12 06:08:18 just gotta come up with something, personally i use a range that's extremely unlikely to ever collide 2017-11-12 06:08:29 00:90:2E:XX:YY:ZZ 2017-11-12 06:08:35 registered to atari, heh 2017-11-12 06:09:29 qemu-system-x86_64 -cdrom alpine-virt-3.6.2-x86_64.iso -netdev type=tap,id=net0,ifname=tap0,vhost=on -device virtio-net-pci,mac=00:90:2E:XX:YY:ZZ,netdev=net0,id=net0 2017-11-12 06:09:33 something like this? 2017-11-12 06:11:34 not sure if you can mix the old-style -cdrom option with explicit device layout, but might work 2017-11-12 06:11:44 also need to replace the last 3 groups of the mac with random hex digits 2017-11-12 06:13:07 like 01,02,03? 2017-11-12 06:18:37 ya, didn't really work. got this error: network script /etc/qemu-ifup failed with status 256 2017-11-12 06:22:17 oh, the qemu package doesn't include a file for that 2017-11-12 06:22:28 can just put an empty shell script there 2017-11-12 06:22:31 exit 0 2017-11-12 06:22:33 is all it needs 2017-11-12 06:23:06 exit 0 as an option, or? 2017-11-12 06:23:21 in /etc/qemu-ifup 2017-11-12 06:23:36 echo -e '#!/bin/sh\nexit 0' > /etc/qemu-ifup; chmod +x /etc/qemu-ifup 2017-11-12 06:23:43 ah 2017-11-12 06:24:59 yep, that seemed to do the trick 2017-11-12 06:25:26 you can put your bridging commands in there 2017-11-12 06:25:34 as well as /etc/qemu-ifdown to clean up the bridge 2017-11-12 06:26:19 you mean all the options after -cdrom? 2017-11-12 06:28:18 nah 2017-11-12 06:28:27 the commands to add the tap interface to the bridge 2017-11-12 06:29:32 well either way I can't get internet to work still I think. using the ping command returns a bad address 2017-11-12 06:30:07 did you setup the bridge? 2017-11-12 06:31:41 I used this command 2017-11-12 06:31:43 sudo qemu-system-x86_64 -cdrom alpine-virt-3.6.2-x86_64.iso -netdev type=tap,id=net0,ifname=tap0,vhost=on -device virtio-net-pci,mac=00:90:2E:01:02:03,netdev=net0,id=net0 2017-11-12 06:32:19 yep, that's just to launch qemu 2017-11-12 06:32:27 that creates a tap interface on the host 2017-11-12 06:32:28 tap0 2017-11-12 06:32:44 you now need to create a bridge, add your physical interface to the bridge as well as the tap0 2017-11-12 06:33:54 not exactly sure how to create a bridge. sorry if I'm a bit uninformed with qemu 2017-11-12 06:34:14 no worries 2017-11-12 06:34:25 it's not specific to qemu, bridges are just a standard networking feature 2017-11-12 06:34:29 basically a virtual switch 2017-11-12 06:34:37 you use the ip command to set them up 2017-11-12 06:35:01 ip link add name br0 type bridge; ip link set br0 up 2017-11-12 06:35:03 creates the bridge 2017-11-12 06:35:06 to add the interfaces you use 2017-11-12 06:35:18 ip link set eth0 master br0 2017-11-12 06:35:23 replace eth0 with your physical interface 2017-11-12 06:35:27 if I'm not mistaken, I think virt-manager handles all that 2017-11-12 06:35:28 and repeat with the tap interface too 2017-11-12 06:35:30 yeah 2017-11-12 06:35:31 not exactly sure though 2017-11-12 06:35:37 it does (optionally) 2017-11-12 06:35:54 well I'll just use that then 2017-11-12 06:36:38 so host device enp3s0: macvtap? 2017-11-12 06:36:54 enp3s0 would be it 2017-11-12 06:37:04 source mode is set to Bridge 2017-11-12 06:38:25 thing is, I've tried it this way before and the internet still doesn't work. I know this because packages fail to download 2017-11-12 06:50:47 I'm getting wget: bad address on all of the mirrors 2017-11-12 06:53:59 what address do you have on the guest? same subnet as your normal LAN? 2017-11-12 06:54:41 the ipv4? 2017-11-12 06:55:46 yep 2017-11-12 06:56:41 bad address usually indicates a dns problem 2017-11-12 06:58:29 I'm using ip addr show. do I need the localhost address, because ipv4 does not show in the vm 2017-11-12 06:58:49 well, that'd be why it isn't working 2017-11-12 06:58:53 you need to configure networking in the vm 2017-11-12 06:59:02 on my the actual host I get a lot more information 2017-11-12 06:59:25 remember when it's bridged it's just as if it was a real machine physically connected 2017-11-12 06:59:37 I set it to bridge mode 2017-11-12 07:02:30 I think I see what might be wrong. 2017-11-12 07:02:53 device model is set to rtl8139. it should be virtio as you said before 2017-11-12 07:12:58 ya, that didn't work. maybe I should deal with this tomorrow 2017-11-12 07:24:09 !coins 2017-11-12 07:28:00 DRAWING THE WORDS "HELP DEUCE IS RAPING ME" IN IT ON THE WALL 2017-11-12 07:28:17 <_ikke_> .. 2017-11-12 07:29:15 er 2017-11-12 07:29:32 sorry 2017-11-12 07:29:37 totally wrong channel 2017-11-12 12:41:50 hi all, i've got an issue understanding how a saved configuration and cached packages work. my understanding is that if i've done apk upgrade, have the packages cached, i've saved my config and then reboot, the upgraded packages will be installed on power on, is that correct? 2017-11-12 12:43:54 <_ikke_> marble_visions: when adding packages, apk adds the package name to a file called world. If you store that in the apkovl with lbu, then yes, at boot those packages will be installed again 2017-11-12 12:44:53 so the procedure of upgrading is repeated on boot, the new versions don't get "baked in" the core set of file of the distro 2017-11-12 12:45:57 how does that play with things like shared libs? if i've got a shared lib that's used by a service, if i upgrade the shared lib, i need to keep track of which services i need to restart, right? 2017-11-12 12:46:35 since a reboot won't solve the problem of the service using the old version of the lib, if it gets started before the upgrade process has completed 2017-11-12 12:46:38 <_ikke_> marble_visions: to your fist remark: the entire file system is built by installing packages and restoring the files form the overlay 2017-11-12 12:47:34 <_ikke_> marble_visions: so no, it's not that packages are upgraded on boot 2017-11-12 12:48:30 <_ikke_> it's just as if you completely installed the system from scratch with the newest versions 2017-11-12 12:48:51 _ikke_: that's great 2017-11-12 12:49:00 _ikke_: thanks for the info! :) 2017-11-12 12:49:13 seems like i need to reread the dev/apk wiki 2017-11-12 13:33:13 there seems to be a conflict between bash-completion and util-linux-bash-completion. both try to add rfkill: 2017-11-12 13:33:14 (1/2) Reinstalling util-linux-bash-completion (2.31-r0) 2017-11-12 13:33:14 ERROR: util-linux-bash-completion-2.31-r0: trying to overwrite usr/share/bash-completion/completions/rfkill owned by bash-completion-2.7-r2. 2017-11-12 13:56:24 <_ikke_> what repo/branch? 2017-11-12 14:04:04 edge, cdn-nl 2017-11-12 14:04:26 let me check main/community/testing (i do have testing enabled) 2017-11-12 14:06:17 eh. bash-completion is in main, and util-linux is also in main, but i can't font utils-linux-bash-completion ... 2017-11-12 14:06:30 can't find* 2017-11-12 14:08:50 _ikke_: main/utils-linux/APKBUILD has this, does it mean util-linux-bash-completion is part of util-linux? 2017-11-12 14:08:51 subpackages="$pkgname-doc $pkgname-dev $pkgname-bash-completion:bashcomp:noarch 2017-11-12 14:08:51 libuuid libblkid libmount libsmartcols libfdisk sfdisk cfdisk 2017-11-12 14:08:51 findmnt:_findmnt mcookie blkid setpriv" 2017-11-12 14:09:10 <_ikke_> avih: yes 2017-11-12 14:09:26 ok, so that's the conflict. both on edge/main 2017-11-12 14:10:08 that's what you asked, yes? 2017-11-12 14:16:28 <_ikke_> avih: https://pkgs.alpinelinux.org/contents?file=rfkill&path=&name=&branch=edge&repo=&arch=x86_64 2017-11-12 14:16:41 <_ikke_> That indeed looks like a conflict 2017-11-12 14:16:54 k. what's next? 2017-11-12 14:18:12 <_ikke_> I guess file a bug at bugs.a.o 2017-11-12 14:18:46 TBH historically, the two bugs I did file months ago never got even a single response. I don't trust anyone to notice it. 2017-11-12 14:19:50 i guess some bugs do get handled, but it did not inspire confidence to file more. 2017-11-12 14:22:31 <_ikke_> https://github.com/scop/bash-completion/commit/d962f3a3be2b5c3811e546890c99b19694c2a08e 2017-11-12 14:22:42 <_ikke_> So it has been fixed upstream already 2017-11-12 14:24:11 <_ikke_> hmm, it should be contained in 2.7 2017-11-12 14:25:45 <_ikke_> So I see that the bash-completion package has support to filter out conflicting files, but I wouldn't expect that version to conflict in the first place 2017-11-12 14:29:01 <_ikke_> ok, 2.7 does not contain it yet 2017-11-12 14:29:24 <_ikke_> so, simple fix is to add rfkill to the conflict list untill the next release of bash-completions 2017-11-12 14:38:45 <_ikke_> avih: https://github.com/alpinelinux/aports/pull/2708 2017-11-12 14:40:05 _ikke_: i'd be satisfied with waiting for upstream, but this is even nicer. thanks :) 2017-11-12 14:41:12 <_ikke_> The mechanism was already there, so was an easy fix 2017-11-12 14:41:37 cool. thanks. 2017-11-12 14:42:45 <_ikke_> Because 3.7 is being released, I think it's better to get it fixed in the mean time. 2017-11-12 16:47:50 sorry, been away. yeah, i wasn't aware of 3.7, but agreed. 2017-11-12 21:15:20 so i just followed https://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot but i can't seem to install some packages. specifically: ERROR: unsatisfiable constraints: 2017-11-12 21:15:22 abuild (missing): 2017-11-12 21:15:35 although i can install others, like gcc just fine 2017-11-12 21:17:44 oh abuild is not in 3.7 but in 3.6. the docs are a bit confusing there 2017-11-12 21:19:36 hello everyone 2017-11-12 21:23:10 aep: TIL there is an existing `abuild` already 2017-11-12 21:23:31 heyo liwakura 2017-11-12 21:24:36 I wanted to get someone's opinion. is alpine linux okay for a rolling desktop experience? I also checked out void linux to find some comparisons 2017-11-12 21:41:25 ServiceRobot: I'd say it is, but you'd better check the repos if there are all the programs you use regularly 2017-11-12 21:42:04 ya, that's an issue I noticed when searching certain packages. I think I'll just end up using it for hosting servers 2017-11-13 00:48:49 small bugfix on lxc's initd https://github.com/alpinelinux/aports/pull/2709 2017-11-13 14:57:44 I've asked this before, but... if there's a package in main that uses filecap to set capabilities for a binary in that package, shouldn't that package have the package containing the cap-setting tool in its depends= ? 2017-11-13 14:59:28 especially since this capability setting happens in one of the pre/post install scripts 2017-11-13 16:42:10 is it possible to run alpine inside Google Cloud Compute Engine VM 2017-11-13 18:15:44 <_ikke_> avih: ncopa merged the fix for bash-completion 2017-11-13 18:16:03 _ikke_: :) 2017-11-13 18:16:07 w00t! 2017-11-13 18:16:11 lol 2017-11-13 19:56:22 i get "ERROR: busybox-1.26.2-r7.post-install: script exited with error 127" if i attempt to bootstrap an different arch 2017-11-13 19:56:32 but i have binfmt enabled for the target arch 2017-11-13 19:57:04 /bin/busybox exists, but throws and ENOENT on execve 2017-11-13 19:57:10 which smells like broken libs 2017-11-13 19:57:30 command is `apk --arch armhf -X ${mirror}/v3.6/main -U --allow-untrusted --root ${ch` 2017-11-13 19:57:33 root_dir} --initdb add alpine-base 2017-11-13 19:58:16 i tried looking for cross chroot setups, but it doesn't seem like people do this in general 2017-11-13 20:15:14 ah yeah, i found it 2017-11-13 20:15:23 i just forgot to copy the qemu binary into the chroot 2017-11-13 20:15:35 and that caused the ENOENT 2017-11-13 20:15:54 <_ikke_> A real ENOENT 2017-11-13 20:19:24 ugh 2017-11-13 20:19:42 actually, i wanted to setup binfmt emulation for the whole system 2017-11-13 20:19:56 turns out there is no xf86-video-intel on armhf 2017-11-13 20:20:07 which is somehow to be expected 2017-11-13 20:21:06 *sigh* 2017-11-13 22:31:48 Hi folks, first time here ... AFAIK Apline's Reddit is not so much live place, so I'd like to ask here about the qemu ... 2017-11-13 22:31:52 I'm considering switch from Arch Linux to Alpine linux for my VM host. Is there any qemu-headless (as I know from Arch) option because I can't find this in the packages? 2017-11-13 22:32:30 lurid: alpine is designed for servers, try qemu directly? 2017-11-13 22:33:42 Xe: ok, thanks ... I hope in something like that. So if I understand well qemu package is very minimal, no balast for GUI, etc. 2017-11-13 22:33:56 lurid: to be honest, i have no idea. tias? 2017-11-13 22:33:57 <_ikke_> --disable-gtk is specified, so you can assume it's headless 2017-11-13 22:34:04 ah okay cool 2017-11-13 22:34:15 nice 2017-11-13 22:34:22 <_ikke_> https://git.alpinelinux.org/cgit/aports/tree/main/qemu/APKBUILD#n223 2017-11-13 22:35:34 <_ikke_> But apparently only for x86(_64) 2017-11-13 22:35:47 which is my case :) 2017-11-13 22:36:56 this is good ... thanks, I'm really very excited from Alpine (as former Gentoo and present ArchLinux user) and I'll give it a try! 2017-11-13 22:56:29 hi guys, Is there here an expert with GRUB and linux boot process 2017-11-13 22:56:54 I'm trying to boot from a disk image but I keep having problems 2017-11-13 22:57:17 if you want answers, just ask the question 2017-11-13 22:58:20 Ok.. I didn't want to bother the whole channel 2017-11-13 22:59:08 I'm trying to boot from an img stored on the hard drive 2017-11-13 22:59:36 I followed this guide step by step http://my-zhang.github.io/blog/2014/06/28/make-bootable-linux-disk-image-with-grub2/ 2017-11-13 23:01:16 but I can't figure out the root device when I do `linux /boot/vmlinuz-3.0.0-12-generic root=/dev/sda1` 2017-11-13 23:01:41 of couse I'm using the vmlinuz from the alpine cd 2017-11-13 23:03:07 I think that the `root=XXX` part is the wrong one because when I boot I can't mount to /sysroot 2017-11-13 23:05:03 does someone has any idea about how to accomplish this? 2017-11-13 23:06:57 I would also prefer to avoid an apk.static install but do exactly like the iso does.. reinstall everything everytime it boots 2017-11-14 03:38:23 before i ask this: yes, i know i'm doing a silly thing. i'm trying to get Xorg programs (atm, hexchat) to run in a Docker container with an alpine image. hexchat displays a warning the "iso-2022-jp character set is unknown". aside from that warning, it works fine. i'm... not really sure where to even start, tbqh. any ideas where to start poking for trying to debug this? 2017-11-14 03:38:54 (it shows it as a dialog i have to dismiss; if it weren't for that i'd just ignore it) 2017-11-14 03:53:38 are you japanese? 2017-11-14 03:53:44 if not, try figuring out where that charset is set 2017-11-14 03:55:00 i'm not. poking at the hexchat source, it looks like hexchat has a default list of servers and their character sets, and checks every server in that list even if you don't try to connect (lol...) to see if the charset is known by iconv. 2017-11-14 03:55:13 duckinator: set it explicitly then? 2017-11-14 03:55:55 not sure what you mean? hexchat explicitly checking if the iso-2022-jp character set is usable, to see if it can convert from it to UTF-8. 2017-11-14 03:56:16 strange, i'm not awfully familiar with hexchat's codebase 2017-11-14 03:59:18 mmm... it looks like the default server list is hard-coded into the codebase(?!), and thus it always checks if iconv supports iso-2022-jp no matter what. so i think i'll just find another IRC client to play with in Docker, then lol 2017-11-14 04:00:08 I have to say, a hard-coded server list was not something I expected to encounter: https://github.com/hexchat/hexchat/blob/6f8b8c3d3ca272f141f3fc36961802607d0c2289/src/common/servlist.c#L51-L363 2017-11-14 04:21:04 hexchat is broken on musl because dalias hasn't properly implemented locales yet 2017-11-14 04:21:26 it will always send utf-8 2017-11-14 04:21:30 but it is being worked on 2017-11-14 04:55:43 duckinator: that server list thing isn't too surprising, xchat is dead and hexchat's code gets minimal maintenance these days, i'm sure that's low on their list of issues to fix 2017-11-14 04:55:54 considering it works as expected despite being kinda bonkers 2017-11-14 04:56:18 scv: yeah, that's fair. 2017-11-14 04:56:33 pretty funny though 2017-11-14 04:56:36 didn't expect that personally 2017-11-14 04:56:37 lol 2017-11-14 05:11:04 kaniini, that narrative isn't really correct. locales have nothing to do with whether iconv can convert to iso2022jp 2017-11-14 05:12:30 yes and no. technically not, but users group it in with locales. 2017-11-14 05:12:56 either way, iconv does not support outputting to the requested charset 2017-11-14 05:13:05 and the user gets an error 2017-11-14 05:16:55 nod 2017-11-14 05:18:15 if youve been following on #musl we're investigating how to best implement conversion to legacy cjk encodings 2017-11-14 05:19:00 as well as ebcdic ncopa requested 2017-11-14 05:24:36 the latter is almost done 2017-11-14 08:04:38 ebcdic is relatively simple iirc 2017-11-14 08:05:17 cjk, shiftjis, and big5 though. 2017-11-14 08:13:45 ¯\_(ツ)_/¯ 2017-11-14 08:48:14 big5 is fun times 2017-11-14 08:48:41 cjk just means chinese/japanese/korean mepholic 2017-11-14 08:49:18 til 2017-11-14 08:54:52 shiftjis support means you can do tripcode calculation in musl directly 2017-11-14 08:55:07 among other things 2017-11-14 08:55:42 yes, i am sure the weebs will be thrilled 2017-11-14 09:31:49 lmao 2017-11-14 11:27:31 kaniini: i'd appreciate some help. consolekit2 seems to update much more that i _think_ it should. i don't know exactly what triggers its update, but it's always being "updated" to the same version it already was. no errors appear etc. is it expected? 2017-11-14 11:29:14 i don't think the things it depends on trigger it directly, as i _think_ many times none of its deps are updated but consolekit2 still gets updated (again, to the very same version it already was) 2017-11-14 11:30:48 first question would be: what does an update do if it keeps the same version? is it only for running some post-update scripts? 2017-11-14 11:31:31 well.. that would have been the second question. the first would be "what triggers it?" 2017-11-14 19:55:03 Hello, where can I find the vmlinuz kernel parameters for alpine 2017-11-14 19:55:42 ? I mean.. does alpine have specific parameters to pass to grub 2017-11-14 19:55:43 ? 2017-11-14 19:58:25 I know for a fact that one is `alpine_dev` but where can I find all the others? 2017-11-14 20:26:19 dj32: alpine uses extlinux, not grub, see /etc/update-extlinux.conf 2017-11-14 20:26:42 ok.. sorry.. I installed grub 2017-11-14 20:26:52 because I need fallbacks 2017-11-14 20:27:20 It seems udhcpd is not being added to recent busyboxes - there is a service "udhcpd", but when the service is started, the service complains that /usr/sbin/udhcpd does not exist, and it's not enabled as an available app in the latest busybox package in edge. 2017-11-14 20:29:51 <_ikke_> https://git.alpinelinux.org/cgit/aports/tree/main/busybox/busyboxconfig#n940 2017-11-14 20:30:17 I know, I read the config file. 2017-11-14 20:30:28 <_ikke_> It's moved to -extras 2017-11-14 20:30:44 <_ikke_> https://git.alpinelinux.org/cgit/aports/tree/main/busybox/busyboxconfig-extras#n920 2017-11-14 20:30:55 Ah, lovely. Thanks a bunch! 2017-11-14 20:34:20 <_ikke_> The conf file should probably also be moved 2017-11-14 20:53:16 Would it not make sense too to have the service be installed as part of the busybox-extras package, since it won't work without that package by default? 2017-11-14 20:55:49 <_ikke_> duncan^: I was looking into that 2017-11-14 20:56:01 <_ikke_> but it's busybox-initscripts that installs the services 2017-11-14 21:23:44 yeh, they should be included in the package that provides the actual service. just my opinion. 2017-11-14 21:24:21 <_ikke_> I agree 2017-11-14 22:44:23 can I put . in package names? 2017-11-14 22:44:41 and if so, how can I make split packages like that? The wiki is a bit lacking 2017-11-14 22:47:29 this is my (broken) APKBUILD: https://sr.ht/kpC3.txt 2017-11-14 22:47:54 eventually hits this: https://sr.ht/AlkH.txt 2017-11-14 22:53:21 does $pkgname have to be in subpackages? 2017-11-14 22:53:48 I mean 2017-11-14 22:53:50 I'd like it to be 2017-11-14 22:53:58 that causes a call to a function named $pkgname which doesn't exist (and isn't a valid shell function name) 2017-11-14 22:54:16 oh I see 2017-11-14 22:54:20 you don't specify the parent package in subpackages 2017-11-14 22:54:31 you can override the packaging function name though 2017-11-14 22:54:44 works without that 2017-11-14 22:54:46 many thanks 2017-11-14 22:54:56 glad to be of help :) 2017-11-14 22:55:40 well, almost 2017-11-14 22:55:51 I have run into another snafu 2017-11-14 22:56:09 removing that now leads me to: https://sr.ht/myKm.txt 2017-11-14 22:57:37 I can't tell you exactly the point where that happens but it has to do with the fact that for the main package you use $pkgdir, for subpackages $subpkgdir if memory servers me right 2017-11-14 22:57:48 something like that 2017-11-14 22:57:54 ah 2017-11-14 22:58:05 yep, that did it 2017-11-14 22:58:08 thanks very much! 2017-11-14 22:58:30 np 2017-11-15 07:34:43 all, can I simulate a stable release with alpine edge and a custom repo for updates? 2017-11-15 07:37:52 or, there might always be incompatibilities between versions so manual upgrade cannot be ruled out between future versions? 2017-11-15 09:24:33 when i sign up for the wiki i get "413 - Request Entity Too Large", should I file a bug? 2017-11-15 09:29:49 marble_visions: it's already been reported 2017-11-15 09:29:56 i personally have no idea who manages the wiki or who to ask 2017-11-15 09:29:59 i think clandmeter ? 2017-11-15 09:30:30 danieli: thanks 2017-11-15 09:38:47 marble_visions: does it happen when you submit the form? 2017-11-15 09:39:29 clandmeter: it happens when I click on the "Create your account" button 2017-11-15 09:39:35 and also if I try to login 2017-11-15 09:39:52 ah right i see it now 2017-11-15 09:40:02 i cant login anymore either :) 2017-11-15 09:41:05 i will ask the person responsible to look at it. 2017-11-15 09:43:17 clandmeter: thank you 2017-11-15 09:53:06 DOatJam: https://github.com/ariya/phantomjs/issues?utf8=%E2%9C%93&q=alpine 2017-11-15 09:56:11 clandmeter: OK, maybe I should the compile it with MUSL 2017-11-15 10:01:12 one more question, how to check the installed app with apk or other commands? as I'm newbone with alpine 2017-11-15 10:02:40 apk info is your friend 2017-11-15 10:02:59 it tells you more than what you want to know about packages with the right switches 2017-11-15 10:07:03 TBB: thanks anyway 2017-11-15 18:26:15 guys, i don't know what am i missing but I cannot start gunicorn in a docker container. installed it using apk add --no-cache --progress py-gunicorn, PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin 2017-11-15 18:26:29 error is /bin/sh: gunicorn: not found 2017-11-15 18:27:11 which gunicorn --> /usr/bin/gunicorn 2017-11-15 18:27:39 can you start a shell in the container and try to start it manually, shreyansh_k ? 2017-11-15 18:27:40 I can verify that gunicorn executable does exist in /usr/bin/ 2017-11-15 18:27:59 liwakura: Already tried that. 2017-11-15 18:28:06 what happens then? 2017-11-15 18:28:11 liwakura: I have a shell running right now. 2017-11-15 18:28:52 liwakura: /bin/sh: gunicorn: not found 2017-11-15 18:28:52 nice nickname liwakura ;) 2017-11-15 18:29:10 I am confused. which gunicorn return /usr/bin/gunicorn but cannot run this executable. 2017-11-15 18:29:16 shreyansh_k: Do you have /usr/bin/python2? 2017-11-15 18:29:25 It looks like py-gunicorn doesn't actually depend on python 2017-11-15 18:29:43 liwakura: /usr/bin/python3 2017-11-15 18:30:05 scv: i played Brood War alot 2017-11-15 18:30:06 i fixed that recently on edge 2017-11-15 18:30:09 try edge version 2017-11-15 18:30:26 shreyansh_k: But /usr/bin/gunicorn's shebang refers to python2, not python3 2017-11-15 18:30:30 You need python2 installed 2017-11-15 18:30:32 ha, my nickname isn't a sc reference but i get that a lot 2017-11-15 18:30:39 scv: lol. 2017-11-15 18:30:46 lets all love Lain then, instead. 2017-11-15 18:30:48 edge has separate py2-gunicorn and py3-gunicorn packages, but 3.6 does not 2017-11-15 18:30:49 indeed 2017-11-15 18:30:49 <3 2017-11-15 18:31:56 pardis: yeah, i cannot directly install py3-gunicorn. only py-gunicorn will install and during install, it will pull for whatever versoin of python that is installed. 2017-11-15 18:32:21 Are you sure? What does 'head -1 /usr/bin/gunicorn' say? 2017-11-15 18:32:41 Dockerfile: https://pastebin.com/MYFSZj3U 2017-11-15 18:33:05 pardis: #!/usr/bin/python2 2017-11-15 18:33:09 Okay, so you need python2 2017-11-15 18:34:14 pardis: Oh! so py-gunicorn must be installing for py2. what package do i need to install for py3? py3-gunicorn didn't work. 2017-11-15 18:34:29 3.6 does not have a py3-gunicorn package 2017-11-15 18:34:34 You're stuck with python2 if you want to use 3.6 2017-11-15 18:34:36 Or you can try edge 2017-11-15 18:34:43 But that might break 2017-11-15 18:34:59 I don't know if there's an edge docker image or not, I don't use docker 2017-11-15 18:37:53 pardis: in that case, when is the next release for alpine coming? 2017-11-15 18:38:20 No idea, I'm not a developer 2017-11-15 18:38:26 But it should be soon, 3.6 came out around 6 months ago I think 2017-11-15 18:39:06 pardis: Ahan!, okay thanks. 2017-11-15 18:44:44 shreyansh_k: end of the year +- 1 month 2017-11-15 18:46:57 liwakura: okay, thanks! 2017-11-15 20:56:44 marble_visions: wiki should work again 2017-11-15 22:00:52 clandmeter: great, i'll check it out tomorrow :) 2017-11-15 23:25:35 what do you all think of lighttpd? 2017-11-15 23:37:53 ash_workz: ask a more specific question 2017-11-15 23:38:40 is lighttpd popular in the alpine crowd? 2017-11-15 23:47:07 ash_workz: i don't see why that should matter to you 2017-11-15 23:47:11 use whatever fits you best 2017-11-15 23:47:13 i use nginx the most 2017-11-15 23:47:37 some alpine guys seem to like darkhttpd, some lighttpd 2017-11-15 23:53:32 danieli: yeah, I was just wondering why some people chose to take the routes they did 2017-11-15 23:57:59 then i suggest you just ask the question directly :) 2017-11-15 23:58:07 instead of asking around the question 2017-11-16 00:08:38 ah 2017-11-16 00:11:33 I asked in #lighttpd but if anyone wants to explain their use case, I'd appreciate it 2017-11-16 00:13:55 so 2017-11-16 00:14:12 you want to know why people prefer lighttpd over other httpds? 2017-11-16 00:30:21 the py-imaging package was built without zlib support which is not very useful (it cannot handle either jpg or png images) 2017-11-16 00:45:17 nsz: that sounds like a mistake.. who maintains it? 2017-11-16 07:07:20 Hey, I installed a bunch of -dev packages, so that I could compile and install lxml. But somehow, when I remove those -dev packages, lxml breaks. Why might this be? 2017-11-16 07:07:31 (Are the -dev packages more than just headers and compiler stuff?) 2017-11-16 07:08:14 Specifically, it breaks if I remove libxslt-dev 2017-11-16 07:11:25 Arahael: probably, some of them should be under depends="" rather than makedepends="" 2017-11-16 07:13:15 Ah, ok. 2017-11-16 07:31:59 nsz: re py-imaging, it looks dead upstream. can you use py-pillow instead? 2017-11-16 08:19:12 i use pillow instead of pil myself 2017-11-16 08:19:26 ncopa: thoughts on an #alpine-mips and alpine-mips@lists.a.o? 2017-11-16 09:04:08 alpine-mips sounds good to me, as long as it does not require too much of my time 2017-11-16 09:04:40 but why alpine-mips channel and mailing list when we have none for x86 and 86_64? 2017-11-16 09:04:44 or s390x? 2017-11-16 09:04:48 or aarch64? 2017-11-16 09:04:52 what makes mips special? 2017-11-16 09:15:57 well, what you're really saying is: why are the other architectures not properly organized 2017-11-16 09:16:25 so lets fix them all while at it 2017-11-16 09:16:35 +1 2017-11-16 09:17:04 a lot of people have shown up asking how to contribute to arm, to be met with "uhh..." 2017-11-16 09:17:30 so i suggested to danieli and foxkit that if they want to do mips, they should probably have it well-organized from start 2017-11-16 09:18:21 <_ikke_> But does that require them to seperate out to different 'channels'? 2017-11-16 09:25:52 i think that many people on the mailinglists won't want to see noise from things that aren't close to what their working on 2017-11-16 09:26:10 the problem that introduces is visibility into cross-architecture issues 2017-11-16 09:26:33 i wish there was a better way than mailing lists 2017-11-16 09:26:48 usenet!! 2017-11-16 09:36:24 _ikke_: i think it is useful to have architecture-specific channels for more specialized discussions. other distributions (basically all of them) have architecture-specific channels for discussing arch-specific work/issues/support 2017-11-16 09:37:20 it makes sense to me to have arch specific channels 2017-11-16 09:37:36 it's also useful because then you can join #alpine-arm for example and ask them for help with your packages on ARM 2017-11-16 09:38:18 right now a lot of people maintain packages which break on ARM and then they have no idea who is developing alpine on ARM and so they have no idea who to talk to, and just try to guess what the solution may be 2017-11-16 09:38:33 i'm not trying to pick on ARM here, but it really is kind of the land of the unknown 2017-11-16 09:38:58 the PPC team is easier to spot, if all else fails, look for somebody connecting to IRC from IBM :P 2017-11-16 09:39:13 and then there's the (erroneous) assumption that everyone is developing alpine from an x86/x86_64 machine 2017-11-16 09:39:34 so, arch-specific channels/mailing lists help to resolve these assumptions/lack of knowledge 2017-11-16 09:40:05 it also helps to serve as a rallying point for new people who care about a specific architecture 2017-11-16 09:40:21 which, for something like ARM is probably not a terrible thing 2017-11-16 09:47:34 agree 2017-11-16 10:25:16 ACTION sent a long message: cypa[m]_2017-11-16_10:25:15.txt  2017-11-16 10:25:29 something wrong? 2017-11-16 10:26:50 I could reproduce this with whois command 2017-11-16 10:27:05 I guess 2017-11-16 10:28:58 definitely ... something wrong with whois 2017-11-16 10:30:31 just out of curiosity, what record did you asked to whois? 2017-11-16 10:30:44 vps.eco 2017-11-16 10:30:50 and many others 2017-11-16 10:30:52 mpls.ru 2017-11-16 10:31:15 there are many fancy tlds nowadays 2017-11-16 10:32:02 are you using the built-in whois client or the packaged one? 2017-11-16 10:32:04 i'll look into it 2017-11-16 10:32:22 they are using busybox whois 2017-11-16 10:32:34 as noted in the dmesg output 2017-11-16 10:32:50 ACTION sent a long message: cypa[m]_2017-11-16_10:32:50.txt  2017-11-16 10:33:13 oh right 2017-11-16 10:33:25 i think busybox whois may be broken 2017-11-16 10:33:30 should I place a bug in the tracker? 2017-11-16 10:33:47 you should probably file a bug directly to busybox 2017-11-16 10:33:57 OK 2017-11-16 10:34:24 btw, who needs whois in busybox? 2017-11-16 10:36:20 i will need whois for security work 2017-11-16 10:36:42 on busybox? 2017-11-16 10:37:19 why? 2017-11-16 10:37:31 i will have to use alpine on systems that are airgapped for various reasons 2017-11-16 10:39:12 you could use packaged whois, maybe? 2017-11-16 10:39:55 implying i have the package available 2017-11-16 10:40:01 still a shame to throw away busybox features though 2017-11-16 10:40:06 the busybox whois client is really limited :/ 2017-11-16 10:40:35 it might fit if you are working with arin resources, but in ripe land it's pretty much annoying 2017-11-16 10:42:23 So I've setup a django project on alpine linux. It's *huge*, and almost all of it is in python's site-packages. Any obvious ways to cut down the size? 2017-11-16 10:42:43 Arahael: cut down on the size of what, exactly? 2017-11-16 10:43:05 danieli: My ~250MB of site-packages. 2017-11-16 10:43:45 Arahael: i'd run a `find site-packages` to have a look exactly whats there 2017-11-16 10:43:48 ^ 2017-11-16 10:43:57 that kinds sounds like more of a python issue though? 2017-11-16 10:43:58 ncopa: Loads of packages, the largest is django itself. 2017-11-16 10:44:05 danieli: Yes, it's definitely a python thing. 2017-11-16 10:44:17 so not an alpine-specific question 2017-11-16 10:44:17 depending on how you install your packages, you may have or may not have things like compiled sources there 2017-11-16 10:44:38 might be you have pip cache and similar 2017-11-16 10:44:49 also, i'd put install most things with pipX install --user 2017-11-16 10:44:54 and ^ 2017-11-16 10:44:54 I suppose I probably should delete the compiled sources. It'll slow down the initial launch, but that's only once. 2017-11-16 10:45:12 precompiled python, yes 2017-11-16 10:45:15 danieli: Nah, it's an epheremal setup. --user unneccessarily complicates it. 2017-11-16 10:45:21 aha 2017-11-16 10:45:31 in that case, nevermind 2017-11-16 10:45:47 danieli: (But I do feel that --user should be the default, but that's a whole different topic) 2017-11-16 10:50:44 you should switch from python to ada, maybe ;) 2017-11-16 10:51:04 cypa[m]: I'm switching to Go, whenever I have teh chance. :) 2017-11-16 10:51:38 does functional paradigm available there? 2017-11-16 10:52:18 ada, hahaha 2017-11-16 10:52:21 i've done a ton of ada and spark 2017-11-16 10:52:39 lately, mostly python, erlang and c++ 2017-11-16 10:52:45 wow ... it's surpise 2017-11-16 10:53:16 do you aware of ada's SOM? 2017-11-16 10:53:31 cypa[m]: Go isn't really functional. 2017-11-16 10:54:15 cypa[m]: But it's quite decent. The quality of it's libraries are generally excellent, and it's ecosystem and tooling are quite good, too. 2017-11-16 10:56:25 SOM is like windows COM-objects but better 2017-11-16 10:57:51 ACTION shudders 2017-11-16 10:58:42 I don't actually know anything about SOM. 2017-11-16 10:59:54 it's an old thing but cool 2017-11-16 11:00:18 It's interesting... As an industry we seem to be reverting back to crude messaging. 2017-11-16 11:00:36 i'm familiar with SOM but i haven't used it much 2017-11-16 11:01:58 and there is some magic which could lead to multiplatform binaries (unix/windows) elegantlly 2017-11-16 11:03:17 Meh, just cross compile it. 2017-11-16 11:03:32 no, it's a mess 2017-11-16 11:04:16 It's trivial in Go. :) 2017-11-16 11:04:25 could you imagine binary which works on linux/windows/macos flawless? 2017-11-16 11:05:36 GOOS=linux go build... GOOS=windows go build... GOOS=darwin go build... 2017-11-16 11:07:43 thats some sort of interlanguage ABI 2017-11-16 11:08:01 That's... Insane. 2017-11-16 11:08:13 http://octagram.name/pub/somobjects/other-impl/ilu/2.0b1/manual-html/manual_2.html#SEC42 2017-11-16 11:08:18 What about different architectures? 2017-11-16 11:08:47 many cool things looks inane first 2017-11-16 11:09:17 With an 's'. ;) 2017-11-16 11:09:43 But no, I meant it with a positive connotation, though I do wonder how they get the binary dependencies in without bloating it. 2017-11-16 11:10:22 Ah, it's an IDL-based RPC. 2017-11-16 11:10:43 it has now ABI by far 2017-11-16 11:10:51 *no 2017-11-16 11:11:52 Interesting. 2017-11-16 11:12:18 windows.. not really a target platform for me 2017-11-16 11:12:24 same with OS X 2017-11-16 11:12:29 occasionally, but rarely 2017-11-16 11:19:01 http://octagram.name/pub/somobjects/Release-to-Release%20Binary%20Compatibility%20in%20SOM.pdf 2017-11-16 11:19:16 http://octagram.name/pub/somobjects/ 2017-11-16 11:20:01 btw, firefox from time to time consumes 100% CPU on https://mail.yandex.ru/touch 2017-11-16 11:20:21 and on https://tweetdeck.twitter.com also 2017-11-16 11:21:56 cypa[m]: Probably for the ads, or the add blocker. 2017-11-16 11:22:22 It looks like your machine isn't beefy enough to run a modern web browser :') 2017-11-16 11:22:48 maybe 2017-11-16 11:25:09 holy crap 2017-11-16 11:25:15 facebook messenger has syntax highlighting 2017-11-16 12:37:12 Why would you care :P ? 2017-11-16 12:37:35 rhetorical question? 2017-11-16 12:38:04 hi, btw 🙂 2017-11-16 12:38:05 yes, mostly 2017-11-16 15:03:40 I want to install imapfilter but it isn't available via apk search imapfilter. It looks like an apk package exists but I don't know enough about alpine to understand what to google to resolve this. 2017-11-16 15:04:49 <_ikke_> Th3Mafia: apk add imapfilter 2017-11-16 15:05:08 <_ikke_> But it's only in testing, so no repo has it yet 2017-11-16 15:05:14 <_ikke_> You'd have to build it yourself 2017-11-16 15:05:42 _ikke_: can I just add the testing repo? 2017-11-16 15:05:55 <_ikke_> there is no testing repo 2017-11-16 15:05:58 in edge 2017-11-16 15:06:10 <_ikke_> testing just exists in the aports git repository 2017-11-16 15:06:19 ok 2017-11-16 15:06:21 <_ikke_> oh, my bad 2017-11-16 15:06:24 duncan^: is edge a repo? 2017-11-16 15:06:27 <_ikke_> duncan^: is right 2017-11-16 15:06:31 <_ikke_> edge *does* contain testing 2017-11-16 15:06:36 <_ikke_> I was confused with unmaintained 2017-11-16 15:06:49 <_ikke_> So you need to add the testing repo from edge 2017-11-16 15:07:12 Thanks, that should provide me with enough to read about and understand, thanks guys!! 2017-11-16 15:07:25 <_ikke_> Th3Mafia: protip: use pinned repos 2017-11-16 15:07:27 http://pkgs.alpinelinux.org/package/edge/testing/x86_64/imapfilter 2017-11-16 15:33:23 Is there a way to run a command when an Alpine Docker container first starts? I know the Alpine wiki says the init system is OpenRC, but afaict this doesn't apply to Docker images (OpenRC just isn't installed, and rc* commands aren't available). 2017-11-16 15:36:02 specifically, i'm trying to auto-launch Xvfb or a VNC server when the container starts running. I know the general approach for this is to have an entrypoint script for Docker, but the codebase is already cluttered as heck so i'd like to keep it in the Docker file if it's possible to do so. 2017-11-16 15:37:10 i mean i guess a horrifically hacky thing i could do is have the Dockerfile write something to a shell script in /tmp and use that as the entrypoint, but if there's a better approach i'd certainly like to know it haha 2017-11-16 15:37:57 <_ikke_> duckinator: A docker container is basically just a single process that starts 2017-11-16 15:38:05 duckinator: so you wnat run multiple service within the container? 2017-11-16 15:38:22 you could use some sort of service supervisor 2017-11-16 15:38:24 <_ikke_> duckinator: that single process can then start other proceses 2017-11-16 15:38:26 ncopa: i want to run a single service within the container, which the actual command being run relies on 2017-11-16 15:38:44 so 2 processes 2017-11-16 15:38:48 yeah 2017-11-16 15:38:54 <_ikke_> there is supervisord 2017-11-16 15:39:01 or s6 2017-11-16 15:39:03 <_ikke_> right 2017-11-16 15:39:23 <_ikke_> Are you running the container interactively? 2017-11-16 15:39:30 openrc is overkill 2017-11-16 15:41:29 _ikke_: nope. i want something that starts, initializes an Xvfb server, then runs the ENTRYPOINT which is a script that compiles some code and runs qemu 2017-11-16 15:42:12 is the command a service? 2017-11-16 15:42:20 or a oneshot thing 2017-11-16 15:42:36 <_ikke_> sounds like a oneshot thing 2017-11-16 15:42:47 you need to kill the Xvfb after command is done 2017-11-16 15:42:55 or the container will continue run 2017-11-16 15:43:03 the command is just a one-shot thing. 2017-11-16 15:43:32 Xvfb has options so that when the last X client exits it'll terminate the server 2017-11-16 15:43:44 ok 2017-11-16 15:44:12 i think a wrapper script should be enough as entrypoint 2017-11-16 15:44:18 the script woudl do: 2017-11-16 15:44:48 Xvfb -foo 2017-11-16 15:44:51 <_ikke_> duckinator: can't the Xvfb run in its own container? 2017-11-16 15:45:05 exec /path/to/the-real-command "$@" 2017-11-16 15:45:23 sorry 2017-11-16 15:45:25 Xvfb -foo & 2017-11-16 15:45:28 _ikke_: i'm not sure, tbqh. 2017-11-16 15:46:08 <_ikke_> duckinator: How does the compiling code communicate with the xvfb server? 2017-11-16 15:46:51 _ikke_: Xvfb is just a headless Xorg server, so the script compiles the code + runs qemu, and i just have to set the $DISPLAY environment variable appropriately for it to connect 2017-11-16 15:51:25 ... *facedesk* 2017-11-16 15:51:33 just realized qemu can be run with `-display none` 2017-11-16 15:51:43 lol 2017-11-16 15:52:21 <_ikke_> right, was thinking about that 2017-11-16 15:59:30 uhh, wat? the qemu package doesn't provide qemu-system-i386. O.o 2017-11-16 16:00:40 <_ikke_> separated out 2017-11-16 16:00:55 <_ikke_> qemu-system-i386 2017-11-16 16:02:08 huh. any idea why that doesn't show up on https://pkgs.alpinelinux.org/packages?name=qemu ? 2017-11-16 16:02:51 you need wildcard: https://pkgs.alpinelinux.org/packages?name=qemu* 2017-11-16 16:03:06 ahh, okay. so just me being silly lol 2017-11-16 16:03:22 nah 2017-11-16 16:03:36 its not very intuitive 2017-11-16 16:09:59 clandmeter: when i try to edit a page when i click save page i get "This action has been automatically identified as harmful, and therefore disallowed. If you believe your action was constructive, please inform an administrator of what you were trying to do. A brief description of the abuse rule which your action matched is: New users are not allowed to add ip addresses and phone numbers" 2017-11-16 16:10:14 clandmeter: page in question "Comparison with other distros" 2017-11-16 16:10:33 marble_visions: its related to uri's 2017-11-16 16:11:01 first 5 hours you cannot add/update pages with urls 2017-11-16 16:11:18 clandmeter: ooh. ok. will wait then. thanks 2017-11-16 16:11:45 you can break the uris and tell me what to replace it with if needed. 2017-11-16 16:12:09 we get a lot of spam on the wiki 2017-11-16 16:12:10 sorry about that. 2017-11-16 16:15:33 this is what i wound up doing, if anyone's curious: https://github.com/puppy-technology/awooos/blob/master/Dockerfile + https://github.com/puppy-technology/awooos/blob/master/build.sh 2017-11-16 16:22:21 when building a pkg for local usage, i added a key with abuild-keygen -a -i but get this err 'ERROR: APKINDEX.tar.gz: UNTRUSTED signature', gpg --lsign-key KEY says it finds no key 2017-11-16 16:23:03 paradoxspiral: you need to put your public key into /etc/apk/keys 2017-11-16 16:29:55 it is in there, xentec 2017-11-16 18:09:21 hello 2017-11-16 18:09:45 can someone confirm, that alpine rpi image works over uart? 2017-11-16 18:28:46 fugitive_: what do you mean? 2017-11-16 18:28:52 like console via /dev/ttyU0? 2017-11-16 18:28:55 yea 2017-11-16 18:28:58 well 2017-11-16 18:29:09 I'm guessing the answer is yes 2017-11-16 18:29:32 I bootstrapped and built optimized alpine packages for the beaglebone black (armv7a) 2017-11-16 18:29:39 and the UART console works fine 2017-11-16 18:29:44 i tried, as i usually do, screen /dev/ttyUSB0 115200, but not getting any output.. o,0 i see that framebuffer is built in module 2017-11-16 18:29:58 you might need to modify your initrd 2017-11-16 18:30:03 or kernel command line 2017-11-16 18:30:07 also check uh 2017-11-16 18:30:30 your /etc/inittab 2017-11-16 18:30:36 make sure there's a line similar to 2017-11-16 18:30:38 ttyS0::respawn:/sbin/getty -L ttyS0 115200 vt100 2017-11-16 18:30:43 ah.. i thought so.. maybe the module isn't loaded..so initrd needs a rebuld 2017-11-16 18:30:45 but for the proper serial device 2017-11-16 18:30:58 yeah forgot to check that 2017-11-16 18:31:01 it might already be in the initrd 2017-11-16 18:31:11 best bet would be to start with /boot/extlinux 2017-11-16 18:31:18 or whatever your bootloader is 2017-11-16 18:31:53 i am using native rpi image, 3.6 2017-11-16 18:32:05 but will see 2017-11-16 18:32:11 thank you for advices ;) 2017-11-16 18:32:41 APPEND BOOT_IMAGE=/boot/vmlinuz-grsec modules=sd-mod,usb-storage,et4 root=UUID=f5a1f2db-2754-46b9-80a5-54fb91546c28 rootfstype=ext4 console=${console} 2017-11-16 18:32:49 in extlinux.conf 2017-11-16 18:33:01 I'm not exactly sure what ${console} is, but it works 2017-11-16 18:33:19 mepholic: variable subsitution 2017-11-16 18:33:34 imagine it like the ruby #{} 2017-11-16 18:33:34 Xe: lol!? 2017-11-16 18:33:42 that's not at all what I meant 2017-11-16 18:33:51 i got it :) it is ok 2017-11-16 18:34:17 Xe: I meant I don't know what the contents of the console variable is 2017-11-16 18:34:18 also 2017-11-16 18:34:29 mepholic: oh, wanna know a cheap way to find out? 2017-11-16 18:34:32 can module be loaded from, config.txt like - modules=fb ? 2017-11-16 18:34:46 mepholic: $ cat /proc/cmdline :D 2017-11-16 18:34:54 APPEND sets the kernel arvg 2017-11-16 18:34:54 Xe: additionally, if I didn't know what bash command substitution was 2017-11-16 18:34:56 argv* 2017-11-16 18:35:02 what makes you expect that I'd know anything about ruby? 2017-11-16 18:35:11 just out of curiosity? 2017-11-16 18:35:15 :)) 2017-11-16 18:35:19 i made a staistical guess lol 2017-11-16 18:36:07 so fugitive_ 2017-11-16 18:36:24 in the case of this bbb, the console parameter on the kernel cli is console=ttyO0,115200n8 2017-11-16 18:37:03 okaay.. ? 2017-11-16 18:37:27 ttyO0 is the TI sitara uart driver devnode 2017-11-16 18:37:53 didn't know that :) 2017-11-16 18:40:07 hmm 2017-11-16 18:40:22 on the rpi2 it looks like it uses a broadcom uart which sits at /dev/ttyAMA0 2017-11-16 18:40:57 for the rpi3 2017-11-16 18:40:57 that is bluetooth uart if i am not mistaken? 2017-11-16 18:40:59 > NOTE FOR RASPBERRY PI 3: The Raspberry pi 3 has changed things around a bit: ttyAMA0 now refers to the serial port that is connected to the bluetooth. The old serial port is now called ttyS0. So if you have an RPI3, everywhere you see "ttyAMA0" below, you should read "ttyS0". 2017-11-16 18:41:07 yea 2017-11-16 18:41:12 so it's gonna look pretty normal 2017-11-16 18:41:25 console=ttyS0,115200n8 2017-11-16 18:41:28 let me place sdcard on my lap and investigate.. 2017-11-16 18:41:31 in your /boot/extlinux 2017-11-16 18:41:33 and then 2017-11-16 18:41:48 ttyS0::respawn:/sbin/getty -L ttyS0 115200 vt100 2017-11-16 18:41:51 in /etc/inittab 2017-11-16 18:42:18 yep yep, one sec 2017-11-16 18:44:18 hm.. but this would be solution for already installed alpine.. I will change configs to see if i can install alpine via uart 2017-11-16 18:45:06 hmmmmmm 2017-11-16 18:45:08 yea, look: 2017-11-16 18:45:09 cat cmdline.txt 2017-11-16 18:45:09 modules=loop,squashfs,sd-mod,usb-storage dwc_otg.lpm_enable=0 console=ttyAMA0,115200 console=tty1 2017-11-16 18:45:19 let me swap it.. 2017-11-16 18:45:40 ok so the rpi image is targeted at an rpi2 2017-11-16 18:45:50 fugitive_: are you using the aarch64 image? 2017-11-16 18:46:00 or armhf? 2017-11-16 18:46:01 nop, armhf 2017-11-16 18:46:13 use aarch64 if you're on an rpi3 2017-11-16 18:46:47 really? huh, why? 2017-11-16 18:46:50 that might have correct defaults for the uart 2017-11-16 18:46:59 fugitive_: because the rpi3 is 64 bit 2017-11-16 18:47:11 armhf is 32 bit arm executables 2017-11-16 18:47:28 i'm not even certain if they'll boot 2017-11-16 18:47:40 on the rpi3 2017-11-16 18:48:02 yes - "armhf is 32 bit arm executables" this is correct 2017-11-16 18:49:06 well it looks like armhf _should_ run on the rpi3 2017-11-16 18:49:25 idk that much about arm 2017-11-16 18:49:55 all I know is from the work i've done on my pine64 (haven't gotten alpine working on it yet) and the bbb 2017-11-16 18:52:09 black screen still.. ttyS0 didn't help :/ 2017-11-16 18:52:34 have you given minicom a shot? 2017-11-16 18:52:42 lets try that. 2017-11-16 18:52:51 also are you corrected to the proper uart pins? 2017-11-16 18:53:00 no reversed polarity crap going on? 2017-11-16 18:53:31 nah, same with minicom 2017-11-16 18:53:54 uart setup is ok.. i booted, bsd, debian,devuan xyz..distros.. 2017-11-16 18:54:07 must be a module 2017-11-16 18:55:27 i booted once alpine on H3 allwiner using this uart as well.. but it was homebuild kernel.. 2017-11-16 18:56:42 i am too lazy bastard to attach hdmi.. :( uart ftw :)) 2017-11-16 18:59:54 btw mepholic, there is no aarch64 for rpi out of box 2017-11-16 19:00:19 hmmm 2017-11-16 19:00:23 interesting 2017-11-16 19:00:51 fugitive_: there's been talks of adding architecture specific channels/mailing lists to alpine 2017-11-16 19:01:09 i see.. 2017-11-16 19:01:19 so there should soon be a good place you can go to for ARM help/development on alpine 2017-11-16 19:01:32 anyways, re: module 2017-11-16 19:01:34 possibly 2017-11-16 19:02:14 nevermind.. will rebuild a kernel, init when i get time.. just wanted to see is uart possible 2017-11-16 19:03:07 and it is - but the module needs to be loaded first heh, that is the catch 2017-11-16 19:06:35 good luck :) 2017-11-16 19:06:57 # grep -i 'FRAMEBUFFE' config-rpi 2017-11-16 19:06:57 CONFIG_FRAMEBUFFER_CONSOLE=m 2017-11-16 19:07:27 and - CONFIG_FB=y 2017-11-16 19:07:36 meh.. nvm :) 2017-11-16 19:07:41 thanks :)) 2017-11-16 19:12:43 no problem :) 2017-11-16 19:12:55 fugitive_: so since it's compiled as a module 2017-11-16 19:13:14 all you gotta do is figure out the relative kernel module path 2017-11-16 19:14:02 and you can create a .modules file in /etc/mkinitfs/features.d 2017-11-16 19:14:06 similar to 2017-11-16 19:14:11 $ cat /etc/mkinitfs/features.d/cryptsetup.modules 2017-11-16 19:14:13 kernel/crypto/* 2017-11-16 19:14:15 kernel/arch/*/crypto/* 2017-11-16 19:14:17 kernel/drivers/md/dm-crypt.ko 2017-11-16 19:14:46 so if you call it rpi-uart.modules 2017-11-16 19:15:07 you can add rpi-uart to the beginning of the list of features in /etc/mkinitfs/mkinitfs.conf 2017-11-16 19:15:15 if you don't wanna rebuild the whole kernel 2017-11-16 19:17:56 yep yep, but for no preinstalled alpine framebuffer needs to be loaded 2017-11-16 19:19:05 ah :) 2017-11-16 19:19:54 heh, yep, unless it can be loaded somehow o,0 (w/o initram rebuld) 2017-11-16 19:21:04 i love and hate arm equally ... 2017-11-16 19:28:29 just tried restarting firefox (not sure if this is the first time since upgrades) and it hangs without displaying anything 2017-11-16 19:28:46 strace shows it seems to be deadlocked (repeated timedwait on a futex) 2017-11-16 20:37:37 firefox hungs on https://tinkoff.ru 2017-11-16 20:38:35 process dissapears as it were never launched! 2017-11-16 20:38:55 ESR version 2017-11-16 21:33:00 try launching from terminal for more info. 2017-11-16 23:44:11 I'm trying to install/update my apk's. However, when I do apk-update, I ultimately get: ERROR: http://dl-cdn.alpinelinux.org/alpine/v3.6/main: temporary error (try again later) 2017-11-16 23:44:19 What may have gone wrong here? 2017-11-16 23:44:45 Arahael: sounds like fastly is having issues, or its backend is down 2017-11-16 23:44:46 lemme check 2017-11-16 23:45:41 danieli: It works on one computer I have, but not another. 2017-11-16 23:47:27 Arahael: is DNS configured and working? 2017-11-16 23:48:36 danieli: Should be, how do I confirm? 2017-11-16 23:49:25 Arahael: `ping google.com`? 2017-11-16 23:50:24 Wow, that's a bad address. I guess that answers it fo rme. 2017-11-16 23:50:37 now check if network is up 2017-11-16 23:50:44 if you can `ping 8.8.8.8` it should be 2017-11-16 23:51:10 if it is up, edit /etc/resolv.conf to use your preferred DNS servers (such as "nameserver 8.8.8.8\nnameserver 8.8.4.4") 2017-11-16 23:51:40 danieli: I'm actually inside of docker, I'll see what's gone wrong there. 2017-11-16 23:52:02 danieli: I'm a bit new to this, but I didn't think to check the internet and dns because it was just using it to grab the docker images and all that. :) 2017-11-16 23:52:05 Arahael: and you neglected telling me this is a docker container until now because? 2017-11-16 23:52:07 :P 2017-11-16 23:52:41 danieli: I see your point - appologies! :) 2017-11-16 23:53:05 if you use a domain in any internet related activity (such as downloading a package index), you use both the internet and DNS 2017-11-16 23:53:10 even containers do that 2017-11-16 23:53:44 Yep. 2017-11-16 23:55:24 Found the issue, an obscure firewall entry. Now chasing the guilty party... 2017-11-16 23:55:29 danieli: Sorry for bothering you! 2017-11-16 23:55:38 no worries 2017-11-17 00:38:38 guys, there is some errors in https://wiki.alpinelinux.org/wiki/Classic_install_or_sys_mode_on_Raspberry_Pi 2017-11-17 00:38:51 how can i contribute with edits? 2017-11-17 00:47:38 fugitive_: honestly i'd start by making an offline version of the page and editing things by hand until you can get a wiki account 2017-11-17 00:47:50 i remember you have to request something in order to get write access 2017-11-17 00:49:45 okay, will do tomorrow. mistakes are small.. but might confuse beginers.. thanks Xe 2017-11-17 00:50:26 no problem fugitive_, i didn't know that wiki page existed lol 2017-11-17 00:51:57 :))) 2017-11-17 00:52:12 i am good at googling hehe 2017-11-17 00:53:43 gnite everyone 2017-11-17 01:29:58 Hi everyone ! May someone help me with fstab to mount a cifs share ? The "mount -a" command works fine, but when I reboot my rPI 1, the shares are not mounted automatically 2017-11-17 01:33:12 here's an example of the fstab content "//192.168.1.15/Share /media/share cifs user=guest,password=,_netdev 0 0" 2017-11-17 01:41:18 Smeulf: does the system log (/var/log/messages or something similar) have any clues? 2017-11-17 01:42:46 Xe not it doesn't... Could it be just a service disabled, like netmount ? 2017-11-17 01:43:43 Smeulf: what does rc-status say? 2017-11-17 01:44:09 Runlevel: default ntpd [ started ] crond [ started ] acpid 2017-11-17 01:44:17 oops 2017-11-17 01:44:22 https://gist.github.com ^^ 2017-11-17 01:45:23 Xe: https://gist.github.com/anonymous/fe52e4063a5b92df96616ad0611b5f7f 2017-11-17 01:46:33 Smeulf: interesting. lemme compare to one of my x86 alpine boxes 2017-11-17 01:46:37 Xe: got it !! had to enable netmount, damn... that distro is "too" light for me lol 2017-11-17 01:46:49 ah yeah that'd do it! 2017-11-17 01:46:54 Xe: but that's why I choosed it 2017-11-17 01:47:10 Smeulf: rc-update add netmount and reboot 2017-11-17 01:47:21 the rebooting is so you know the FS mounts on boot 2017-11-17 01:47:34 Xe: Thanks !! yes it's working now ! 2017-11-17 01:47:42 yay 2017-11-17 01:47:50 Xe: btw, maybe you can e interessed by this : 2017-11-17 01:48:15 Xe: I installed my PI "diskbase" 2017-11-17 01:48:41 Xe: added a user, to gain ssh access with sudo, as described on the wiki, 2017-11-17 01:48:43 Smeulf: as in the normal sys install on the raspi? 2017-11-17 01:49:12 that one was really hard to do by hand but felt so worth it yeah 2017-11-17 01:50:50 Xe: sorry, disconnected... 2017-11-17 01:50:58 Xe: yes as described in the wiki. But my user was not allowed to run "/bin/ash", nor accessing to it's home folder until I did a "chmod 755 /" 2017-11-17 01:51:26 ahhhhhh, good old executable permission on / 2017-11-17 01:52:00 Xe: yep, maybe a script to fix for the next release ? ^^ 2017-11-17 01:52:42 i'm pretty sure if you're doing a sys install on raspi you're really flying by the seat of your pants tbh 2017-11-17 01:52:53 iirc it's not officially supported 2017-11-17 01:54:58 Xe: I'm a player lol, and it's to setup a "docker" box, i'm concerned about the performances if I have the containers on an img file on the SD card. But yes it's not officially supported, but still in the wiki :p 2017-11-17 01:55:53 Smeulf_: good luck to you, docker on arm was (for me at least a year ago) really interesting 2017-11-17 01:56:48 Xe: what do you think for that use ? better have a diskless install or the sys one ? docker on PI is so good, I use it under a libreELEC too, and it's very impressive 2017-11-17 01:57:25 honestly, i don't know. it really depends on a lot of factors 2017-11-17 01:58:51 Xe: you know what, maybe i'll run some benchmarks if I have some time, and report back in a few weeks, so we'll both know ;) (do I have 2 iddentical sd card :S ... lol) 2017-11-17 01:59:30 Xe: anyway, thanks a lot for your help and support ! Have a nice... day, night... 2017-11-17 01:59:35 you too 2017-11-17 01:59:51 ty 2017-11-17 04:23:04 ACTION sent a long message: cypa[m]_2017-11-17_04:23:04.txt  2017-11-17 04:23:16 how to collect core 2017-11-17 04:23:26 this segfault if reproducible 2017-11-17 04:23:30 & 2017-11-17 04:23:32 ? 2017-11-17 04:30:58 ulimit -c unlimited 2017-11-17 04:31:01 before executing 2017-11-17 04:51:20 oops ... 2017-11-17 04:51:26 ACTION sent a long message: cypa[m]_2017-11-17_04:51:24.txt  2017-11-17 04:51:39 core is useless? 2017-11-17 04:52:08 you need debug symbols 2017-11-17 06:00:49 why firefox conflicts with firefox-esr? 2017-11-17 06:01:29 apk packages 2017-11-17 06:02:09 <_ikke_> cypa[m]: because both provide /usr/bin/firefox 2017-11-17 06:02:14 <_ikke_> so they cannot co-exist 2017-11-17 06:14:10 Hi. I'm on a fresh Alpine 3.6.2. `apk info transmission` displays a wrong url ("tansmission", missing the "r") 2017-11-17 06:15:55 then why not /usr/bin/firefox-esr? 2017-11-17 06:16:08 KuroiKitsu: https://github.com/alpinelinux/aports/blob/master/main/transmission/APKBUILD#L6 <-- confirmed, filing a PR right now 2017-11-17 06:18:12 filed as https://github.com/alpinelinux/aports/pull/2744 2017-11-17 06:19:46 Xe: Perfect, thanks. 2017-11-17 11:14:01 ✿ alpine doesnt have tmpwatch? ❀ 2017-11-17 11:24:32 <_ikke_> opal: please stop with the 'flowers' 2017-11-17 11:42:09 <^7heo> oh that's what's been spamming all this UTF so far?! 2017-11-17 11:42:13 <^7heo> what a useless noise... 2017-11-17 12:33:26 ahoy all, just getting started w/ Alpine and using it for a recovery USB drive and SD card. while i see that Alpine local backup can be used for keeping changes persistent, is there any major reason for not simple installing Alpine directly to the USB/SD card? thanks. 2017-11-17 12:34:12 only reason would be so you can remove the usb after boot 2017-11-17 12:34:32 everything will run from tmpfs 2017-11-17 12:34:53 reason to install it directly to usb or sd might be to save ram memory 2017-11-17 12:35:26 but it may run slightly slower 2017-11-17 12:35:47 and if you unplug the usb bad things will happen 2017-11-17 12:38:53 ncopa: OK, thanks for confirming that. Alpine is very simple and fast, looking forward to trying it out. thanks for the help and Alpine 2017-11-17 16:08:15 Hi all. Could someone explain the --virtual flag for `apk add` to me? I'm not sure I properly understand the output of `apk --help`. I'm mostly interested in the --virtual flag in the context of Docker containers. 2017-11-17 16:08:55 jottr: a virtual package is an empty package with no contents at all 2017-11-17 16:09:02 but you can assign dependencies to it 2017-11-17 16:09:34 ncopa: Not sure I understand the use-case. 2017-11-17 16:09:42 I.e. https://github.com/gliderlabs/docker-alpine/blob/master/docs/usage.md#virtual-packages 2017-11-17 16:09:43 example: apk add --virtual .myvirt musl-dev 2017-11-17 16:10:12 if you try to do: apk del musl-dev, it will give you an error 2017-11-17 16:10:19 because .myvirt holds it 2017-11-17 16:10:45 if you do: apk del .myvirt it will uninstall the empty package + all deps 2017-11-17 16:11:12 for example, you want install some temporary build deps while building something 2017-11-17 16:11:25 then you can uninstall all the deps in one go 2017-11-17 16:11:40 ncopa: Ah ok. So this does not help with decreasing the size of the packages at all, but simply makes sure that certain packages are installed/deinstalled together. 2017-11-17 16:11:50 correct 2017-11-17 16:12:00 and you can 'lock' the packages 2017-11-17 16:12:12 so nothing else uninstalls it while you are building 2017-11-17 16:13:12 example: apk add --virtual .build-deps libpng-dev tiff-dev && pip install soemething && apk del .build-deps 2017-11-17 16:13:50 you could in theory apk add libpng-dev tiff-dev && ... && apk del libpng-dev tiff-dev 2017-11-17 16:14:05 but if you latre add another build dep, you need to also remember to remove it 2017-11-17 16:14:10 later* 2017-11-17 16:14:23 so it helps you make sure the add and del correspond 2017-11-17 16:14:41 also, if you on the ams machine at the same time build something else, that also depend on libpng-dev 2017-11-17 16:15:00 if the first build deletes libpng-dev before the sescond is done, it may cause the second to fail 2017-11-17 16:15:11 Ah ok. So this would help me reduce the image since it removes build-time dependenies. How can I identify which are only build time dependencies, and not packages I actually need to be available in the running container? 2017-11-17 16:15:33 Let's say I'm building a container for rails 2017-11-17 16:15:42 i understand 2017-11-17 16:15:53 thats slightly trickier 2017-11-17 16:16:12 ncopa: Example https://gist.github.com/jottr/8d9e27f6899150cde4e2c5d7903a6acb 2017-11-17 16:16:40 This container is currently 700MB in size, and I'd love to reduce its size 2017-11-17 16:17:09 https://github.com/docker-library/python/blob/cf179e4a7b442b29d85f521c2b172b89ef04beef/3.6/alpine3.6/Dockerfile#L77 2017-11-17 16:17:28 there you have an example how to find out all the runtime deps under /usr/local/ 2017-11-17 16:18:25 then it will && apk add --virtual .python-rundeps $runDeps 2017-11-17 16:18:31 and apk del the build deps 2017-11-17 16:18:53 if some of the build deps are also rundeps, then will it not remove the rundeps 2017-11-17 16:19:31 ncopa: Thx. Crazy cool. 2017-11-17 16:19:39 I'll try and adapt it to my needs. 2017-11-17 16:30:39 ncopa: Hm. https://git.io/vFS2Q 2017-11-17 16:31:49 ncopa: Can this be resolved? Or could aggresively optimizing for size this way bite me down the road? 2017-11-17 16:38:00 Hello, I have a question. on a fresh install in sys mode. I runned mkinitfs without any parameter. 2017-11-17 16:38:20 after that .. the system doesn't reboot 2017-11-17 16:38:46 syslinux's menu is shown but it gets stuck in a boot loop 2017-11-17 16:39:16 I'm testing in virtualbox with the `virt` image 2017-11-17 16:39:27 any ideas? 2017-11-17 17:26:24 weekend for me now sorry 2017-11-17 17:27:22 <_ikke_> ncopa: enjoy 2017-11-17 17:31:10 ncopa: before you go, is lua-aports lua 5.2 only? 2017-11-17 23:13:16 Hi 2017-11-18 00:23:12 hey, i run Xorg + udev on a qemu guest but only get the keyboard on not the mouse attached :/ any idea/ 2017-11-18 00:24:09 what flags did you pass for mouse input 2017-11-18 00:24:48 ✿ yeah it isnt like i -asked a question yesterday ❀ 2017-11-18 00:24:52 flags? i have defined InputDevice in /etc/X11/xorg.conf but not sure if its correct 2017-11-18 00:25:00 ✿ people are so quick to complain ❀ 2017-11-18 00:25:01 no, the flags to qemu 2017-11-18 00:25:48 its libvirt generic ps2 mouse 2017-11-18 00:25:55 ill check the qemu process 2017-11-18 00:26:51 typically i find you need to use an emulated usb tablet to get mouse input working reliably 2017-11-18 00:26:57 for absolute input rather than relative 2017-11-18 00:27:18 e.g. -device usb-tablet,(options) 2017-11-18 00:27:34 libvirt should expose similar functionality as an option 2017-11-18 00:27:50 cool thanks ill test that 2017-11-18 00:28:42 it works!!!! :P thanks 2017-11-18 00:30:29 i also cant use kms but would like XServer to automatically test different graphic drivers? is this somehow possible? 2017-11-18 00:30:43 19:23:12 hey, i run Xorg + udev on a qemu guest but only get the keyboard on not the mouse attached :/ any idea/ 2017-11-18 00:30:52 i have that problem on my host machine sometimes 2017-11-18 00:30:54 lol 2017-11-18 00:31:08 i just have to fuck with the mouse until libinput picks it up 2017-11-18 00:31:16 move it to differerent ports etc 2017-11-18 00:31:20 i have no idea why 2017-11-18 00:31:34 I thought about writing some try/error scripts.. 2017-11-18 00:32:16 i actually cant remove the mouse from libvirt but it works fine with touchscreen :) 2017-11-18 00:49:59 just use the tablet method mepholic 2017-11-18 01:35:20 anybody alive? 2017-11-18 01:37:13 I 2017-11-18 01:37:43 you should ask online instead of alive, I think 2017-11-18 01:38:51 Well I have never gotten any help from people who were on-line but they were dead 2017-11-18 01:39:04 oko: i'm not sure if i am alive or not, but i can try to help you if you can ask your question 2017-11-18 01:39:17 A quick question about the upgrade process 2017-11-18 01:39:39 you can just ask, you don't have to ask to ask or anything 2017-11-18 01:39:39 How does one upgrade from let say 3.5.1 to 3.6.2 2017-11-18 01:39:43 ah 2017-11-18 01:39:49 do you have backups? 2017-11-18 01:39:55 I am a n00b when it comes to Alpine (long term OpenBSD user) 2017-11-18 01:40:54 Yes and no. I am running Dom0. So I guess I could fire up another instance do hot migration, fresh installation and than put things back together 2017-11-18 01:41:24 because if so just change the alpine version major.minor in the url in /etc/apk/repositories (and *.list in /etc/apk/repositories.d iff it exists) and reboot 2017-11-18 01:41:40 er, apk -U upgrade and reboot 2017-11-18 01:41:55 That what I thought 2017-11-18 01:42:21 Wiki has a long explanation but I gathered it is for upgrading 2.xxx to 3.xxx 2017-11-18 01:44:18 Thank for the help! 2017-11-18 01:44:27 I meant to say thanks 2017-11-18 01:57:09 oko: no problem, safe upgrades 2017-11-18 05:30:04 oko: I'm not sure if I'm alive. (Coffee would help me know, but I'm trying to quit) 2017-11-18 10:33:50 Hi 2017-11-18 11:27:19 hi 2017-11-18 11:36:21 oops 2017-11-18 11:36:32 ACTION sent a long message: cypa[m]_2017-11-18_11:36:30.txt  2017-11-18 11:36:49 before the commands there was firefox-esr 2017-11-18 11:40:22 it wont install any firefox whatever I do with apk 2017-11-18 11:40:23 help! 2017-11-18 11:44:47 should I rebuild apk db? 2017-11-18 11:45:03 uh what 2017-11-18 11:46:32 ahoi 2017-11-18 11:46:58 i am trying to use alpine:edge in docker for a python app, following problem: 2017-11-18 11:47:04 apk add py3-httlib2 2017-11-18 11:47:09 python3 2017-11-18 11:47:15 import httplib2 2017-11-18 11:47:21 -> ModuleNotFoundError: No module named 'urlparse' 2017-11-18 11:47:38 does urlparse exists in alpine? 2017-11-18 11:49:28 afaik you need to install it separately 2017-11-18 11:49:46 which package would that be? didn 2017-11-18 11:49:50 t find anything 2017-11-18 11:52:27 ACTION sent a long message: cypa[m]_2017-11-18_11:52:26.txt  2017-11-18 11:52:28 and also 2017-11-18 11:52:48 ACTION sent a long message: cypa[m]_2017-11-18_11:52:47.txt  2017-11-18 11:53:25 and in fact no firefox binary on fs 2017-11-18 12:17:01 mik__: perl-uri probably 2017-11-18 12:17:35 cypa[m]: wtf did you do 2017-11-18 12:18:01 i think i found it 2017-11-18 12:25:32 first I did 'sudo apk add --force firefox' 2017-11-18 12:25:53 after that command firefox binary dissapeared 2017-11-18 12:26:49 before that command firefox-esr was installed and used for months 2017-11-18 12:50:26 is the maintainer from https://git.alpinelinux.org/cgit/aports/tree/main/py-httplib2/APKBUILD?h=3.3-stable here? 2017-11-18 12:51:04 https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Python#Python_2_vs._Python_3 -> Multiversion package for project using 2to3 2017-11-18 12:51:25 the python3-httlib2 package that gets built is the same as the python2 one 2017-11-18 15:10:17 $ sudo apk add libgcc 2017-11-18 15:10:18 1 error; 2525 MiB in 780 packages 2017-11-18 15:10:40 same "1 error" for libstdc++ 2017-11-18 15:10:50 I've tried 3 mirrors 2017-11-18 15:10:55 v3.6 2017-11-18 15:13:19 cypa[m]: could it just be that you somehow messed your apk db up? 2017-11-18 15:14:29 so how to fix it? 2017-11-18 15:14:52 btw, non-lib packages added ok 2017-11-18 15:15:02 how are we supposed to know what you did? 2017-11-18 15:16:10 problem with firefox was resolved commenting this line in repos file 2017-11-18 15:16:11 http://mirror.yandex.ru/mirrors/alpine/edge/testing 2017-11-18 15:16:34 so no reason for apk db corruption 2017-11-18 15:18:17 hm, upgrading to edge/testing might be able to block some main or community packages if they depend on an earlier version 2017-11-18 15:18:47 did you try "apk fix" already? 2017-11-18 15:20:13 ```$ sudo apk fix 2017-11-18 15:20:14 (1/1) [APK unavailable, skipped] Reinstalling glibc (2.25-r0) 2017-11-18 15:20:15 1 error; 2525 MiB in 780 packages``` 2017-11-18 15:22:12 huh 2017-11-18 15:26:03 im a bit surprised to see glibc on alpine 2017-11-18 15:26:43 https://github.com/sgerrand/alpine-pkg-glibc/releases 2017-11-18 15:27:23 I was triing run firefox quantum binary ... but it failed 2017-11-18 15:27:39 ACTION sent a long message: cypa[m]_2017-11-18_15:27:38.txt  2017-11-18 15:28:05 erm 2017-11-18 15:28:25 could it be that you disabled the repo where the glibc package came from? 2017-11-18 15:29:49 im on 3.6 stable and i dont even see it 2017-11-18 15:32:13 there is no glibc repo, it was downloaded from github and installed from local file 2017-11-18 15:33:39 honestly I've got 1Gb RAM old pc and hope firefox quantum would be better for me 2017-11-18 15:33:51 shall try to compile it then 2017-11-18 16:09:08 1 Gigabit? you poor thing 2017-11-18 16:27:59 it's on purpose ... since we all use minimalist OS 2017-11-18 16:28:21 I've got expirience for IoT 2017-11-18 16:28:46 which has low resources 2017-11-18 16:29:24 btw, is there .ashrc file or something to init /bin/ash 2017-11-18 16:29:34 set PS1=... 2017-11-18 16:29:59 /etc/profile ? 2017-11-18 16:34:23 cypa[m]: .profile is for login shells, you usually set ENV there 2017-11-18 16:34:43 and the ENV variable is used to find the config file for interactive shells 2017-11-18 16:35:05 this is like, the standard behavior, its just bash which is really dumb there 2017-11-18 16:45:29 ok but ~/.ashrc wont work 2017-11-18 16:46:34 what a surprise! 2017-11-18 16:47:27 what did you set your $ENV to? 2017-11-18 16:48:28 < liwakura> and the ENV variable is used to find the config file for interactive shells 2017-11-18 16:49:46 so, you set ENV in /etc/profile or ~/.profile 2017-11-18 16:49:53 $ echo $ENV 2017-11-18 16:49:57 export ENV="$HOME/.ashrc" 2017-11-18 16:50:06 this is invoked upon each login 2017-11-18 16:52:04 and then the config file from env is loaded for each interactive shell 2017-11-18 16:52:20 and you put your aliases and prompt there 2017-11-18 16:53:29 This is standardized via POSIX, but bash gives a shit about POSIX, people were using bash 2017-11-18 16:53:45 yes 2017-11-18 16:53:47 So people forgot the mechanism that works for like, every bourne-like shell except bash 2017-11-18 16:53:54 and relogin made it work 2017-11-18 16:54:00 thanks 2017-11-18 16:55:07 huh ... I'd like to build firefox quantum now 2017-11-18 19:11:55 I'm a wee bit confused on the Alpine package repositories 2017-11-18 19:12:38 According to pkgs.alpinelinux.org, there is a "firefox" package available 2017-11-18 19:13:33 According to apk in my environment (a Docker container), there is only "firefox-esr", and it doesn't install dependencies, and ultimately doesn't work 2017-11-18 19:13:54 the firefox package is only in edge 2017-11-18 19:13:56 are you running edge? 2017-11-18 19:14:32 My /etc/apk/repositories file includes http://dl-cdn.alpinelinux.org/alpine/edge/community 2017-11-18 19:14:52 <_ikke_> did you do a apk update? 2017-11-18 19:14:54 I believe so... the container is built from "alpine:edge" 2017-11-18 19:14:59 yes, I updated 2017-11-18 19:15:07 what does "flagged" on the pkgs.** site mean? 2017-11-18 19:15:18 because firefox on edge is flagged 2017-11-18 19:15:21 but im not sure for what 2017-11-18 19:15:40 <_ikke_> liwakura: It means someone thinks there is a newer version available 2017-11-18 19:15:44 (My total time knowing of Alpine's existence is now approaching 2 days, so I assume I'm doing something silly) 2017-11-18 19:15:49 <_ikke_> or an automatic process flagged it 2017-11-18 19:16:30 .....and I think I just realized what it is 2017-11-18 19:16:30 <_ikke_> firefox is in testing, not community 2017-11-18 19:16:36 yep, that 2017-11-18 19:16:44 I have main and community repos, not testing 2017-11-18 19:18:03 well, that installs, but still doesn't work 2017-11-18 19:18:24 Couldn't open libGL.so.1: ...: No such file or directory 2017-11-18 19:18:55 the only dependencies getting installed are libevent, hunspell, icu-libs, and libvpx 2017-11-18 19:23:50 <_ikke_> mesa-gl provides that 2017-11-18 19:24:39 Hey, it runs now 2017-11-18 19:25:07 thanks 2017-11-18 19:25:11 <_ikke_> Sarten-X: did you install mesa-gl? 2017-11-18 19:25:24 yep 2017-11-18 19:25:34 <_ikke_> ok, so that's missing as a dependency 2017-11-18 19:27:33 ...so not something altogether silly on my part? 2017-11-18 19:27:59 <_ikke_> Sarten-X: dependencies should be handled 2017-11-18 19:28:08 <_ikke_> so, yes, that part was not your fault 2017-11-18 19:28:20 heh 2017-11-18 19:28:27 a welcome surprise 2017-11-18 19:28:28 :D 2017-11-18 19:29:23 hi 2017-11-18 19:32:07 Hello there! Quick question, is it possible to trigger mkinitfs once I change the content of the files in /usr/share/mkinitfs ? 2017-11-18 19:32:55 I try but the initramfs that it generates it's pretty much unusable 2017-11-18 19:35:28 uhm. how will triggering it again solve this issue? 2017-11-18 19:35:58 mkinitfs -o ... usually does that 2017-11-18 20:12:02 sorry I just read liwakura 2017-11-18 20:12:33 liwakura: The thing I'm doing is pretty complex 2017-11-18 20:12:40 I'm installing alpine in chroot 2017-11-18 20:13:13 at the last step I add a linux-grsec kernel, this triggers automatically mkinitfs 2017-11-18 20:14:38 I copied the files (replacing the original mkinitfs files 2017-11-18 20:14:41 ) 2017-11-18 20:15:11 buit it doesn't seem to work (when I extract initramfs ... the /etc/fstab is still the original one) 2017-11-18 20:16:03 no idea about that 2017-11-18 20:16:33 all that manual initramfs pissed me off enough that i just self-compile my static kernels 2017-11-18 20:16:51 yes.. I totally understand 2017-11-18 20:17:00 can I ask you a nice guide 2017-11-18 20:17:02 to do that? 2017-11-18 20:17:35 I'm going this work a work project and I agree the initramfs part is something I would love to avoid 2017-11-18 20:17:59 hah 2017-11-18 20:18:04 i could write a guide 2017-11-18 20:18:12 but im not sure if it would be "nice" 2017-11-18 20:18:40 maybe you'd better off just figuring out how the initramfs stuff really works 2017-11-18 20:19:22 even some resource online would be fine 2017-11-18 20:22:28 liwakura, do you think it's feasible 2017-11-18 20:22:42 to change the initframfs init script 2017-11-18 20:22:49 to do exactly like the iso 2017-11-18 20:23:10 does? (install a ramfs system starting just from the apks)? 2017-11-18 20:23:20 dont really know 2017-11-18 20:23:42 it might be even officially supported 2017-11-18 20:24:45 because I tried and re-tried to launch the iso stored inside an harddrive 2017-11-18 20:25:25 and I ended up giving up 2017-11-18 20:30:54 like what? 2017-11-18 20:30:59 did you dd the iso to the drive? 2017-11-18 20:31:40 actually, i have such an setup, i unpacked one ISO and put the contents into my /boot so i can boot from it if i trash my system 2017-11-18 20:31:49 like an rescue disk... 2017-11-18 20:35:02 mm... 2017-11-18 20:35:24 liwakura that's interesting... I'm working with a vm for some customers 2017-11-18 20:35:46 and i'm using the http://dl-cdn.alpinelinux.org/alpine/v3.6/releases/x86_64/alpine-virt-3.6.2-x86_64.iso 2017-11-18 20:35:47 iso 2017-11-18 20:36:04 and? 2017-11-18 20:36:20 I installed grub in the main partition 2017-11-18 20:36:33 because the customer needs the fallbacks 2017-11-18 20:36:47 in my menuentry 2017-11-18 20:36:57 ah right. btw, syslinux can also do that 2017-11-18 20:37:14 I couldn't find that option 2017-11-18 20:37:34 its an module 2017-11-18 20:37:47 im still not sure how i could help you 2017-11-18 20:38:13 oh.. if you find a method to boot the iso from an harddrive 2017-11-18 20:38:22 I'll be the happiest man ever 2017-11-18 20:38:39 I really tried everything and it always goes in kernel panic 2017-11-18 20:39:16 I don't know if it's related to the fact that I'm using a VM 2017-11-18 20:39:35 usually the kernel panic says a reason 2017-11-18 20:40:38 that might help 2017-11-18 20:42:33 I wonder how would you boot and ISO 2017-11-18 20:43:17 I would copy /boot from the iso itself 2017-11-18 20:43:58 and then mount the loopback device (the iso) and boot from the loopback device 2017-11-18 20:46:07 im sure its possible 2017-11-18 20:46:15 but i'd need to try it myself first 2017-11-18 20:46:31 I'm trying right now... I'm going to paste the grub config soon 2017-11-18 20:46:34 here 2017-11-18 20:57:43 liwakura: this https://pastebin.com/Qhdc9ktv 2017-11-18 20:58:21 ends up in initramfs emergency recovery shell 2017-11-18 21:00:27 of course I'm using the official iso, and the location of the iso is correct 2017-11-18 21:02:56 when exiting from the shell ... the error is /sbin/init not found in new root 2017-11-18 21:12:40 oh, hm. 2017-11-18 21:12:50 i dont consider loading iso's from grub reliable 2017-11-18 21:13:52 mm.. 2017-11-18 21:14:01 hm, in this case, loading modloop would fail, but i dont know if the VM iso even has it 2017-11-18 21:14:01 liwakura: what would you do? 2017-11-18 21:14:17 I think it has it 2017-11-18 21:14:33 d4390f: install alpine regulary, copy the contents of the iso to /boot 2017-11-18 21:14:54 and edit extlinux.conf to have another entry for the ISO kernel / initramfs 2017-11-18 21:15:35 i remember that the default extlinux install by alpine already has an graphical menu 2017-11-18 21:15:51 oh.. so you would keep the whole contents of the "iso" inside the boot partition 2017-11-18 21:16:51 with my static kernel, i only need the ISO initramfs and the repodata 2017-11-18 21:17:48 with dynamic kernel, its probably a bit more difficult because the names from the ISO and the installed system might overlap 2017-11-18 21:18:56 mm.. that's interesting 2017-11-18 21:19:10 I will investigate what you just said 2017-11-18 21:19:21 that its possible and that i do it does not imply that its necessarily an good idea 2017-11-18 21:19:26 it just "works for me" (TM) 2017-11-18 21:19:31 I'd love to have your extlinux configuration 2017-11-18 21:19:37 just to have a good starting point 2017-11-18 21:19:53 https://w1r3.net/wrGyYP.txt 2017-11-18 21:19:56 here you go 2017-11-18 21:20:16 oh.. easy peasy 2017-11-18 21:20:23 its not graphical, i have to press caps lock in the right moment to even get a prompt for selecting 2017-11-18 21:20:36 yes.. sure 2017-11-18 21:20:51 I'm working at something without even a ui 2017-11-18 21:20:53 "initrd" is some self-compiled crap that exactly knows what to do, thats why no further bootparams 2017-11-18 21:21:22 do you generate that initrd? 2017-11-18 21:22:23 hah, "generate". I snatched things from sabotage linux together, glued it together via shell script 2017-11-18 21:22:28 it was done manually 2017-11-18 21:22:37 not good practice 2017-11-18 21:22:41 oh... i understand ;) 2017-11-18 21:23:18 on the other hand, because of the static kernel i never need to update it anyways 2017-11-18 21:23:42 its just invoking cryptsetup and doing other weird stuff im not going to recommend 2017-11-18 21:24:06 lol .. looks like you have a pretty un-standard setup 2017-11-18 21:24:38 I'm actually trying to do something repeatable .. have multiple images of the os in my vm 2017-11-18 21:24:44 and switch from one to the other 2017-11-18 21:25:26 cant you give the customer some way to boot a given iso? 2017-11-18 21:26:31 nope the customer doesn't touch the system.. 2017-11-18 21:26:42 it's the new os that gets downloaded 2017-11-18 21:26:48 as a new image 2017-11-18 21:26:58 and then we change the default boot 2017-11-18 21:27:13 that's why I wanted to use or images or iso 2017-11-18 21:27:26 I need to be able to repeat the process 2017-11-18 21:27:35 usually the boot order settings from the host are used for that.. 2017-11-18 21:27:36 hm. 2017-11-18 21:27:49 without bricking the machine 2017-11-18 21:51:40 liwakura: thanks for your help 2017-11-18 21:51:56 I really appreciated 2017-11-18 21:52:14 no issue 2017-11-18 21:55:17 d4390f: btw i just noticed my boot label is still set to "void", even if im not booting void anymore 2017-11-18 21:55:35 just didn't change after i switched 2017-11-18 22:08:58 hm. i need xf86-video-intel for arm. 2017-11-18 22:09:18 should i push it that its unmasked in alpine? 2017-11-18 23:33:08 hello 2017-11-19 10:04:12 > I'm installing alpine in chroot 2017-11-19 10:04:15 > I'm installing alpine in chroot 2017-11-19 10:05:24 @freenode_d4390f:matrix.org: A day later, so dunno if you'll be seeing this. Also - warning, I've never chrooted myself, so I really don't understand the issue here. 2017-11-19 10:07:43 However, since [PostmarketOS](https://github.com/postmarketOS/pmbootstrap) also install (what is basically) Alpine to a chroot (don't later boot from that - at least not directly though), I thought it might have some pointers / be a useful resource. 2017-11-19 10:08:13 (they use shell and python for the most part) 2017-11-19 10:20:28 seancjscherer[m]: perhaps consider #postmarketos? 2017-11-19 11:14:24 why do so many people here use matrix? 2017-11-19 11:17:36 it is convinient 2017-11-19 11:18:11 a good unified communication attempt attempt 2017-11-19 11:19:51 im a bit grumpy about matrix 2017-11-19 11:20:04 why? 2017-11-19 11:20:08 especially if it claims to be decentralized, but is actually just federated 2017-11-19 11:20:43 at least more decentralized than facebook, but same as email or open xmpp 2017-11-19 11:21:02 where half of the worlds population are @googlemail.com hah 2017-11-19 11:21:06 ok multi-centralized 2017-11-19 11:21:20 thats called "federated" 2017-11-19 11:21:54 I do not read you ... what would you offer instead? 2017-11-19 11:22:00 blockchain? 2017-11-19 11:24:51 guess how i am here 2017-11-19 11:25:01 i found that IRC stinks least 2017-11-19 11:25:21 mostly because it is an easy and dumb procotol you can even type manually in a netcat session 2017-11-19 11:27:31 but hey, at least JSON is better than XML 2017-11-19 11:59:56 danieli: Not sure what you mean ? I thought it might be helpful to user - d4390f . Did you mean I should refer to that channel, or that I should be posting there (the latter doesn't make sense to me, since then the addressee wouldn't see it... )? 2017-11-19 12:02:10 Ah, or you misunderstood and thought I want to chroot / learn understand about it - then your message makes sense to me :). Sorry I wasn't clearer! 2017-11-19 12:02:47 liwakura: JSON is rather fair tradeoff to universal communication, plain IRC is not universal 2017-11-19 12:04:05 liwakura - I think for "everyday" users (I guess I'd still semi-reckon myself in that group), its pretty easy to user ; no need to be logged into a shell somewhere with Irsii to be able to read backlogs, for example 2017-11-19 12:04:14 *to use 2017-11-19 12:12:40 liwakura: btw, you could use matrix.org via IRC interface I guess 2017-11-19 12:13:01 whxy 2017-11-19 12:13:07 im already on irc *shrugs* 2017-11-19 12:14:49 I guess you won't like this suggestion, but if you're serious / want to follow things regarding truely decentralized applications, there is #redecentralize:matrix.org on matrix :P . 2017-11-19 12:16:27 Though I don't see anything like that taking off for a looong time (whereas I would argue Matrix may have at least a chance against eg Whatsapp - ok, may be dreaming just the same...) 2017-11-19 12:24:05 liwakura: think and explore a little, you could use IRC client ---- bridge to matrix ---- martix server --- many apllication bridges to different IM networks etc. 2017-11-19 12:24:57 I think he was trying to say he's quite happy with IRC as it is :). So perhaps no need to keep poking... 2017-11-19 12:25:06 unify your every communication and land them to you precius IRC client 2017-11-19 12:25:29 ya 2017-11-19 12:27:33 sorry, he or she (guess I shouldn't assume, sorry) 2017-11-19 13:24:53 to hell with overly political correctness, you're using a language with a feminine and a masculine third person, and sometimes it goes wrong 2017-11-19 13:29:07 TBB: even better, i dont even care myself 2017-11-19 13:29:47 that's cool; some people get offended for simple things like that 2017-11-19 13:29:52 everybody knew what was meant, so it did its function 2017-11-19 13:55:30 I dunno - if I always wrote ~ "I think she was trying to say she's ..." 2017-11-19 13:55:37 it might seem strange as well 2017-11-19 13:56:00 anyway, didn't meant to enflame a discussion here. 2017-11-19 13:56:50 *mean 2017-11-19 14:26:16 no worries; if you guess wrong then you're corrected and then you'll know, it's no big deal :) 2017-11-19 16:25:50 Hi 2017-11-19 16:26:07 Having some problems with debootstrap 2017-11-19 16:26:28 I'm trying to run Source Dedicated Server on a musl system 2017-11-19 16:26:50 http://sprunge.us/YJQe this is what happens 2017-11-19 16:27:12 sy: "couldn't download" - hmm 2017-11-19 16:27:15 can you do "apk add udev"? 2017-11-19 16:27:24 oh right, debootstrap 2017-11-19 16:27:29 hi 2017-11-19 16:27:40 How can I add my own ipv6 dns nameserver to interfaces? 2017-11-19 16:28:06 I tried dns-nameserver and dns-nameserves, but the file resolv.conf is still the same 2017-11-19 16:28:17 danieli: trying 2017-11-19 16:28:23 dns-nameservers should work shafire 2017-11-19 16:28:44 also, on http://wiki.alpinelinux.org/wiki/Running_glibc_programs there are dead links on the srcds part 2017-11-19 16:30:33 danieli: do I need any package installed? 2017-11-19 16:41:46 danieli: same error 2017-11-19 18:50:52 Hi there! I am new to linux and wanted to use alpine on a rpi v3 to communicate with my sonos system - this works pretty fine. Now I wanted to install i2c-tools, to get the pi to read some 1wire sensors. Unfortunately I failed at installing i2c-tools, unless the package is in the repo. When I enter 'apk add i2c-tools' it fails with ERROR: unsatisfiable constraints (..); If i search for the package, i can't even find it... What 2017-11-19 18:51:24 <_ikke_> ads0x1: it's in edge/testing 2017-11-19 18:51:34 ^ 2017-11-19 18:52:07 <_ikke_> https://pkgs.alpinelinux.org/packages?name=i2c*&branch=&repo=&arch=&maintainer= 2017-11-19 18:52:43 ah, i see... so i have to change to the edge / testing repo? 2017-11-19 18:53:56 OK, this worked out... thanks for the answer! 2017-11-19 19:10:58 hey, just wanted to start a java application inside a docker container with the openjdk-8-jre alpine image on a alpine linux host 2017-11-19 19:12:23 i get a error message that i have insufficient menory for the JRE https://pastebin.com/6Ufxza6M 2017-11-19 19:12:52 the machine has 4gb of ram and 3.5 are free 2017-11-19 19:13:10 i guess it has something to do with that PaX stuff 2017-11-19 19:13:46 but i don't know how to turn off mprotect for the java binary inside the container 2017-11-19 19:14:04 do something of you have some idea? 2017-11-19 19:14:51 s/something/someone/ 2017-11-19 19:38:30 I am off again, thanks for your help! 2017-11-20 04:21:47 hi 2017-11-20 12:34:21 I've been having some problems with mkinitfs. If one specifies the -n option (to produce an initramfs without modules or firmware), it shows "sh: missing ]", says "initramfs: creating /boot/initramfs-generic" (for instance), but no initramfs is in fact produced 2017-11-20 12:34:56 the problem seems to be twofold: lack of actual confirmation of successfully building an initramfs as well as a broken configuration option 2017-11-20 12:35:09 Or, am I doing it wrong? 2017-11-20 12:41:08 duncan^: haven't used mkinitfs for some time, but could you add -x to mkinitfs's shebang and paste relevant details from the output (sprunge or wherever)? that way it should be slightly easier to track the problem. 2017-11-20 12:42:00 and just for the record, mention which AL version it is. 2017-11-20 12:42:23 Is there a tool for dealing with *.apk-new files after upgrades? 2017-11-20 12:44:31 j_v: I'm not aware of such thing. you should inspect them yourself using diff -u, vim, vimdiff or what you like the best. 2017-11-20 12:45:42 przemoc: sprunge.us/OQRh 2017-11-20 12:45:53 I'm using edge. 2017-11-20 12:46:01 OK, that's what I suspected, but just thought I'd check first before doing it 'by hand'. Thanks. 2017-11-20 12:47:08 j_v: maybe there is such tool, but rather nothing official, and at least it's not part of apk tooling AFAIK. 2017-11-20 12:48:13 duncan^: thanks, at least one typo (lack of space in mkinitfs.in) is clearly visible now 2017-11-20 12:48:52 btw, why no command aliases for apk, I'v misprinting frequently 2017-11-20 12:49:00 apk add ... 2017-11-20 12:49:06 apk install ... 2017-11-20 12:49:10 apk del ... 2017-11-20 12:49:18 apk remove ... 2017-11-20 12:50:50 cypa[m]: make some aliases 2017-11-20 12:51:12 alias 'apk install'='apk add'; apk install cargo 2017-11-20 12:51:19 this wont work for my 2017-11-20 12:51:24 *me 2017-11-20 12:51:46 you need global aliases for that, but that might even be an bashism 2017-11-20 12:51:50 and not posix-ly 2017-11-20 12:52:31 many shells do aliases, e.g. mksh 2017-11-20 12:53:09 its ash 2017-11-20 12:53:31 ``` 2017-11-20 12:53:32 $ alias 2017-11-20 12:53:33 apk install='apk add'``` 2017-11-20 12:54:07 so... verbose 2017-11-20 12:57:15 przemoc: actually, the error with sh appears not to be the root of the problem - after fixing the spacing issue (now there are no more errors), the initramfs still isn't produced, and I don't know why. 2017-11-20 13:11:34 duncan^: ok, but sh syntax error needed to be fixed anyway, so I sent the patch for it 2017-11-20 13:11:51 cool. 2017-11-20 13:14:23 sadly, I don't have time to investigate mkinitfs logic right now and why it doesn't work as you want or as it should 2017-11-20 13:16:33 The script is somewhat convoluted 2017-11-20 13:21:37 AFAICT it is to do with the initfs_kmods section 2017-11-20 13:21:49 duncan^: from quick glance, it seems that when you give -n, nokernel is set to 1, therefore initfs_kmods will return early with 1, therefore next stages won't happen 2017-11-20 13:23:34 duncan^: add 0 to return in first lines of initfs_kmods and initfs_firmware functions 2017-11-20 13:23:40 it should fix the problem 2017-11-20 13:25:00 Yes, that fixed it. Thank you! 2017-11-20 13:26:49 thanks for confirming, will send a patch to ML soon 2017-11-20 13:45:18 I guess not many people tried mkinitfs -n before you. :) 2017-11-20 13:45:33 those untested code paths, they're everywhere... 2017-11-20 13:50:11 <^7heo> yeah 2017-11-20 13:50:17 <^7heo> what does -n do again? 2017-11-20 13:50:34 <^7heo> nokernel? 2017-11-20 13:55:37 omits including modules and firmware 2017-11-20 13:58:59 oh, I see. it was Shiz who forgot that the exit status of the last command executed is returned if no value is provided to return. ;) 2017-11-20 13:59:13 well, happens to anybody 2017-11-20 14:00:00 we simply need more duncans 2017-11-20 14:01:07 (and more Shizes too, obviously) 2017-11-20 14:22:48 duncan^: thanks for reporting 2017-11-20 14:23:01 strange that it has always worked on my machine... 2017-11-20 14:23:13 and thanks for fixing przemoc 2017-11-20 14:24:52 mkinitfs has #!/bin/sh, so maybe you had/have some non-POSIX-conformant sh set? 2017-11-20 14:30:25 Shiz: sh -c 'iwant1() { false || return; }; iwant1; echo $?' # if it returns 1 for you, then I have no idea how it worked on your machine 2017-11-20 14:30:37 ACTION posted a file: dmesg (25KB)  2017-11-20 14:32:21 # if it returns 0, then it would explain why it worked and would mean that your shell is slightly broken 2017-11-20 14:32:38 segfault due to xfce4-power-manager? 2017-11-20 14:33:02 somefimes whole system freezes for a few seconds 2017-11-20 15:43:43 anyone using weechat with wee-slack? 2017-11-20 15:43:56 (on alpine linux) 2017-11-20 15:45:03 i just use their irc gateway 2017-11-20 15:50:50 same up till now - want to try wee-slack for better weechat/slack integration. 2017-11-20 15:51:17 websockets-client python module seems broken though on AL 2017-11-20 15:51:22 or is it just me 2017-11-20 16:33:43 hi 2017-11-20 16:35:33 how to install wee-slack? 2017-11-20 16:35:41 gimme link please 2017-11-20 16:39:40 ACTION suggest looking it up using your favorite search engine 2017-11-20 16:41:15 duncaen (IRC): thanks 2017-11-20 16:48:08 where can I find docs on alpine's addgroup/adduser command? 2017-11-20 16:48:23 man addgroup 2017-11-20 16:48:28 man adduser 2017-11-20 16:48:28 when I man them on linux it doesn't show the same flags I see in other people's dockerfiles 2017-11-20 16:49:01 dnull: are you suggesting I add the manpages to alpine? :\ 2017-11-20 16:49:32 does musl support any of glibc's ld.so wrapping mechanisms (-Wl,-wrap)? 2017-11-20 16:49:59 They would give you the correct flags 2017-11-20 16:50:10 Not sure if alpine supports info 2017-11-20 16:50:12 1 sec 2017-11-20 16:50:31 try: adduser -h 2017-11-20 16:50:33 (whoops, meant to ask on #musl) 2017-11-20 16:51:12 By default, apparently not 2017-11-20 16:51:36 j_v: that worked 2017-11-20 16:52:15 j_v: interesting... so addgroup -h on alpine and linux print different results.. so their not the same command right? 2017-11-20 16:52:35 j_v: is there an online manual for the basic alpine-linux stuff I can reference in the future? 2017-11-20 16:52:46 j_v: or just foo -h ? 2017-11-20 16:52:53 addgroup on AL is busybox applet, at least the default one is 2017-11-20 16:55:00 j_v: so the command descriptions are the most detail I can get on them? (ie: https://busybox.net/downloads/BusyBox.html) ? 2017-11-20 16:55:28 AFAIK, though someone else may know better 2017-11-20 16:58:28 j_v: thanks, that was a lot of help 2017-11-20 16:59:03 ash_workz: you're welcome 2017-11-20 17:07:23 ash_workz: to see a list of busybox applet links, i used (in terminal): for f in $(find /bin /sbin /usr/bin /usr/sbin -type l); do if readlink $f | grep -q busybox; then echo $f; fi; done 2017-11-20 17:07:52 thanks again j_v 2017-11-20 17:07:58 np 2017-11-20 18:11:05 hey! 2017-11-20 18:11:15 anyone here have experience using Alpine Linux on laptops? 2017-11-20 18:11:27 Like anyone have a starting point? 2017-11-20 18:12:34 vpndude, I'm using alpine edge on Dell Inspiron 7720 2017-11-20 18:13:22 got any tips? 2017-11-20 18:14:35 vpndude: like, just regular install? 2017-11-20 18:14:47 liwakura, laptops aren't like desktops 2017-11-20 18:15:40 it depends on what features you expect 2017-11-20 18:17:31 i have alpine myself on several laptops 2017-11-20 18:17:42 well power managment among other things 2017-11-20 18:17:44 GUI, etc 2017-11-20 18:17:58 gui is not special there 2017-11-20 18:18:02 you know.. Alpine as born for secure serving. 2017-11-20 18:18:24 alpine has wpa-supplicant, so wifi is supported 2017-11-20 18:18:37 that's pretty much it as far as laptop-specific stuff 2017-11-20 18:19:13 I had it on a T23 for a while, until an upgrade broke the graphics driver 2017-11-20 18:19:20 not a laptop problem, a crappy S3 graphics chip problem 2017-11-20 18:20:24 S3 and linux 2017-11-20 18:20:25 really? 2017-11-20 18:20:39 Do you hate yourself? 2017-11-20 18:20:42 the only other issue I had on it was a firefox bug with the context menus, which I'm sure is long fixed 2017-11-20 18:20:43 that was with 3.2 2017-11-20 18:21:26 qman__: one of my ThinkPads also has an S3 GPU 2017-11-20 18:21:28 Still working 2017-11-20 18:21:39 yeah, it was the specific one in the T23 2017-11-20 18:21:39 but afaik no xf86-video support for it 2017-11-20 18:21:56 oh, yeah, its beause i also have an T23 2017-11-20 18:21:57 I had to muddle with the X settings to get it to work in the first place, and after the upgrade to 3.3 it broke and I never got it working again 2017-11-20 18:22:51 but most distros have the same problem, ubuntu for example doesn't work out of the box either 2017-11-20 18:23:04 it's an issue with that specific chip and the drivers 2017-11-20 18:23:46 ah i remember what i did to that device 2017-11-20 18:24:01 i plugged an old HDD with Windows XP in and used it for games 2017-11-20 18:24:35 I did have windows 7 on it at one time, and even the windows driver is broken, it works better on software rendering with the generic driver 2017-11-20 18:24:50 if you install the S3 driver the performance drops dramatically 2017-11-20 18:26:15 i still have an Siemens Device with SiS GPU lol 2017-11-20 18:26:50 its i586, and bc of that it cant run most distros 2017-11-20 18:27:04 but alpine, because x86 is i586 and up 2017-11-20 18:27:26 oh, nice 2017-11-20 18:27:32 I have an old K6 in the same boat 2017-11-20 18:30:20 I was running an outdated version of ubuntu server, the last version that had the -386 kernel, until one day it didn't come back after a power outage 2017-11-20 18:30:34 I haven't touched it since as everyone had dropped pre-i686 support 2017-11-20 18:30:41 but I may again if alpine will work 2017-11-20 18:33:03 I was using it as my internet-facing SSH server, too slow to brute force it 2017-11-20 18:37:53 So 2017-11-20 18:38:12 I'm going to work on scripts to manage Alpine 2017-11-20 18:38:44 You know, things to make Alpine friendlier for Laptop / Desktop usage 2017-11-20 19:11:58 vpndude: you should look into contributing to adelie linux, http://adelielinux.org, they are basically forking alpine to make it better for end-user stuff 2017-11-20 19:12:53 (and contributing lots back upstream ofc) 2017-11-20 20:24:20 what is the path in alpine for auto hardware detection? Installed alpine on ARMHF device using chroot method (cant mod uboot or kernel on this board) ) It boots fine, but no hardware is detected or modules loaded. 2017-11-20 20:27:32 <_ikke_> usually eudev takes care of that (though no clue how it works) 2017-11-20 20:29:12 eudev is being called from open-rc, but is not loading any modules. modeuls are in correct path in /lib/modules/ 2017-11-20 20:48:44 Eudev doesn't deal with linux kernel modules, but block devices and the like. 2017-11-20 20:48:55 [e]udev 2017-11-20 20:52:40 @duncan - so what openrc convention is handling non block devices? (network etc...) 2017-11-20 20:54:04 huh 2017-11-20 20:54:28 afaik eudev is accepting module requests and inserting it into the kernel 2017-11-20 20:55:04 I was a bit surprised that it is not included (or at least active in the boot or sysinit runlevels) in the alpine-base dependency chain somewhere by default. 2017-11-20 20:55:38 hm, there is also mdev 2017-11-20 20:56:01 https://i.stack.imgur.com/GFA8G.png 2017-11-20 20:56:17 honestly i dont know how modules are load without eudev 2017-11-20 20:58:21 followed the process from : https://wiki.alpinelinux.org/wiki/Installing_Alpine_Linux_in_a_chroot 2017-11-20 20:59:45 so I did the - rc-update add devfs sysinit rc-update add dmesg sysinit rc-update add mdev sysinit 2017-11-20 21:02:17 those steps show during init as [ ok ] , but lsmod is empty, and not eth drivers loaded 2017-11-20 21:10:46 jammers: you mentioned that kernel couldn't be mod'ed, right? so perhaps the kernel does not have proper support for devtmpfs or uevent helper 2017-11-20 21:12:43 where do we check for those ? in /proc somewhere? 2017-11-20 21:14:47 if the kernel was configured with /proc/config support, that might be an option. 2017-11-20 21:16:21 look in /boot 2017-11-20 21:26:15 that chroot tutorial seems to be missing instruction to "rc-update add hwdrivers sysinit" 2017-11-20 21:26:25 have to test that when I get home... 2017-11-20 21:27:36 found that little gem somewhat referenced here: https://wiki.alpinelinux.org/wiki/Install_Alpine_on_Rackspace 2017-11-20 21:35:24 is it wrong to think that those links should not already be in the alpine-base ? If you use it for docker or something, they wont hurt anything, but it you try to use it for a setup you're left doing it manually (once you dig deep enough to figure all that out)) 2017-11-20 21:36:09 should be included... 2017-11-20 21:48:02 I'll quote "Alpine Linux is a security-oriented, lightweight Linux distribution based on musl libc and busybox." (thats from the front / splash page of the website. 2017-11-20 21:48:19 Why add stuff that some people would need to manually remove ? 2017-11-20 21:50:05 And to be honest, I'm no linux / unix crack, but I got a server up and running after installing from ISO. 2017-11-20 21:51:23 It only needs to be removed if it either adds alot to the size, or would need to be removed manually for the majority of usage. Those links violate neither of those key criteria. 2017-11-20 21:52:54 Perhaps I misunderstood what you're saying - as I said, I'm no "crack". 2017-11-20 21:54:28 setting up from CD on x86 architecture single machine is only one use for alpine. Alpines real value is for docker / server VM's / and other architectures where std distros are just too big. 2017-11-20 21:57:00 I just feel that the "apline-base" install need to be a little more versatile out of the box to better match its diverse intended usage model. 2017-11-20 21:59:16 > Alpines real value is for docker / server VM's / and other architectures where std distros are just too big. 2017-11-20 21:59:16 I would argue that some of that is true, and some not. Yes, I see that as great value that Alpine offers, but I wouldn't see that as its "real value" (sounds to me like other uses are only second rate / not so relevant... 2017-11-20 21:59:45 > Alpines real value is for docker / server VM's / and other architectures where std distros are just too big. 2017-11-20 21:59:49 I would argue that some of that is true, and some not. Yes, I see that as great value that Alpine offers, but I wouldn't see that as its "real value" (sounds to me like other uses are only second rate / not so relevant... 2017-11-20 22:00:42 I (and I think others) see Alpine as pretty darn great in just about any use :D . 2017-11-20 22:04:10 AGREED, or I would'nt be spending the time to learn "yet another distro" in the sea. All the pieces are there, just a small amount of tweaking in the base would yield big returns in deployment :) 2017-11-20 22:10:46 <^7heo> yeah, like adding systemd 2017-11-20 22:10:57 <^7heo> or the glibc 2017-11-20 22:11:05 <^7heo> so it's more compatible 2017-11-20 22:13:32 just take arch then 2017-11-20 22:14:00 7heo, i was resistant to musl for awhile, but it has gained alot of ground in the embedded community in the last few years. I like that they are using it, if you want to add std libc it can sit along side with no issues 2017-11-20 22:14:46 <^7heo> yeah but systemd is obviously more important 2017-11-20 22:15:52 <^7heo> and also the GNU coreutils 2017-11-20 22:16:03 <^7heo> it's much more advanced than busybox 2017-11-20 22:16:38 are you just here for trolling? 2017-11-20 22:16:56 i dont consider "causing breakage on every release" and advancement 2017-11-20 22:16:58 like liwakura said. if you want std stuff, pick a std distro like arch or deb 2017-11-20 22:17:03 s/and/an/ 2017-11-20 22:17:49 <^7heo> I dunno, I ignore liwakura 2017-11-20 22:17:51 is use std distros where they need to be used and embedded distros where they cant 2017-11-20 22:18:04 I use.. 2017-11-20 22:18:15 <^7heo> also std is a C++ namespace 2017-11-20 22:18:45 <^7heo> what you are searching for is GNU/linu 2017-11-20 22:18:51 <^7heo> x 2017-11-20 22:19:54 c++ simply means C + bloat + complexity ;) 2017-11-20 22:20:30 <^7heo> if it was that simple and shallow, itnwould be great 2017-11-20 22:20:42 <^7heo> s/itn/it / 2017-11-20 22:23:44 if I had my way all OS's would have to be written in assembly. you can write all the topware in somthing -on-rails and let the machine compile at every click if your feeling lazy or bored ;o 2017-11-20 22:25:08 <^7heo> sure, compiling is so fast 2017-11-20 22:25:21 <^7heo> I want to recompile firefox on every pageload 2017-11-20 22:25:31 <^7heo> it's gonna be much faster 2017-11-20 22:26:19 are you sure you dont work for Microsoft??/ ;) 2017-11-20 22:26:43 you two, stop that please 2017-11-20 22:26:47 Hello, 2017-11-20 22:26:54 u8432dsa: hi 2017-11-20 22:27:09 hi liwakura, I have a quick question, Is it possible 2017-11-20 22:27:30 to save the messages that get displayed during 2017-11-20 22:27:32 boot time 2017-11-20 22:27:38 in a file 2017-11-20 22:27:44 you looking for `dmesg`? 2017-11-20 22:27:55 yes but it doesn't display everything 2017-11-20 22:28:05 syslog 2017-11-20 22:29:00 mm.. can I see that from the recovery shell? 2017-11-20 22:29:11 I have boot errors while doing the first boot 2017-11-20 22:29:35 (caused by me touching initramfs) I'm trying to debug the errors 2017-11-20 22:29:38 u8432dsa: openrc errors go to /dev/console, so you see it there at bootup 2017-11-20 22:29:49 im not aware of an good method to grab those 2017-11-20 22:30:51 /var/log/rc.log is not turned on? 2017-11-20 22:31:23 never seen that on alpine before 2017-11-20 22:31:29 I don't have /var in the rescue console 2017-11-20 22:32:19 To enable boot logging, uncomment the rc_logger="YES" line in /etc/openrc/rc.conf. 2017-11-20 22:32:42 jammers: he is in the initramfs 2017-11-20 22:32:42 mm..cool 2017-11-20 22:32:54 can I do that in initramfs? 2017-11-20 22:33:14 ah, just saw that thanks 2017-11-20 22:33:17 dont think so 2017-11-20 22:33:28 mm.. let me check 2017-11-20 22:34:06 I think that I have just a bare metal system .. 2017-11-20 22:34:13 mm.. there's a bootchartd 2017-11-20 22:35:13 can I use that? 2017-11-20 22:35:55 what do you even want to do? 2017-11-20 22:36:08 or, whats your original problem? 2017-11-20 22:36:57 my original problem is that there's bug in my init script inside initramfs 2017-11-20 22:37:26 I'm trying to debug that but of course I can block the shell on erro 2017-11-20 22:37:36 error .. so I can't see what's going on 2017-11-20 22:39:08 does your initrd busybox have syslog? 2017-11-20 22:39:53 jammers: i dont think it would make sense for the initramfs /init script to write to syslog 2017-11-20 22:40:04 i'd expect the output to be on /dev/console 2017-11-20 22:40:19 because thats where /init's stderr points to per default 2017-11-20 22:40:55 I can edit /init 2017-11-20 22:41:02 just for debug, or redirect the output of your script stderr to a fille 2017-11-20 22:41:28 2>&1 2017-11-20 22:41:58 jammers: lol 2017-11-20 22:42:36 you CANT invoke it, because its the first PID 1 2017-11-20 22:42:47 so you'd need to edit it to do exec 2>&1 or sth 2017-11-20 22:42:59 but its an ramfs (rootfs actually), so its not persistent 2017-11-20 22:43:15 I can mount my hard drive 2017-11-20 22:43:18 in the first lines 2017-11-20 22:43:25 but yet... 2017-11-20 22:43:41 what error message does it show? 2017-11-20 22:44:00 eh.. that's my problem 2017-11-20 22:44:18 why dont you see it? 2017-11-20 22:45:20 because it scrolls up 2017-11-20 22:45:49 I also tried set +x 2017-11-20 22:46:28 u8432dsa: you can scroll a bit with Shift+PgUp 2017-11-20 22:46:45 also, if you reboot, the error message should be in sight 2017-11-20 22:46:58 ok.. now I want to die :S 2017-11-20 22:47:11 I didn't know I could scroll up 2017-11-20 22:47:12 <^7heo> u8432dsa: have you tried `find / -type f -exec chmod a+x {} \;` as root? 2017-11-20 22:47:25 u8432dsa: FYI ^7heo is a troll 2017-11-20 22:47:50 yes.. I mean .. I'm not an idiot XD 2017-11-20 22:47:59 just to make sure. 2017-11-20 22:48:03 hahah 2017-11-20 22:48:12 _ikke_: actually you could just kick or ban him 2017-11-20 22:48:18 I appreciate that ;) 2017-11-20 22:51:16 liwakura , yesy that s what I meant was call the script to debug from a dummy init script with redirection to a file 2017-11-20 22:52:00 mm.. jammers that's actually a great idea 2017-11-20 22:52:11 jammers: you cant manually re-invoke /init 2017-11-20 22:52:25 so you would need a reboot to get hold of PID 1 again 2017-11-20 22:52:36 = this would trash your changes 2017-11-20 22:53:43 but I can mount a disk 2017-11-20 22:53:48 and write the changes there 2017-11-20 22:54:19 and how do you instruct the initramfs to load it? 2017-11-20 22:54:44 eh.. that's a good question 2017-11-20 22:56:05 mount it from the dummy script before exec the suspect script 2017-11-20 22:56:43 so the dummy script will become init and my init will become the launched script 2017-11-20 22:57:07 :) 2017-11-20 22:59:21 you do realize that you need to regenerate the initramfs for that, right? 2017-11-20 22:59:41 yes 2017-11-20 22:59:41 sure 2017-11-20 23:00:11 I'm working on initramfs, I have a script that regenerates initramfs continuously 2017-11-20 23:00:31 uh but we already have mkinitfs 2017-11-20 23:00:45 or dracut 2017-11-20 23:01:01 yes.. but this is for a specific project 2017-11-20 23:02:11 mkinitfs is cool but sometimes 2017-11-20 23:02:32 it didn't generate a usable initramfs 2017-11-20 23:02:45 if launched without parameters 2017-11-20 23:03:07 i don't think mkinitfs is one of the new-style commands that will work without parameters 2017-11-20 23:03:41 dracut lets you put the kernel parameters in the initramfs 2017-11-20 23:03:44 eh.. it has the default files 2017-11-20 23:03:50 if that is what you mean 2017-11-20 23:04:53 duncan^, are you talking about https://github.com/alpinelinux/mkinitfs/blob/master/initramfs-init.in#L265 2017-11-20 23:04:58 those 2017-11-20 23:05:01 ? 2017-11-20 23:10:55 thank you for your help 2017-11-20 23:11:08 I'll try to see if I can fix my situation here 2017-11-21 04:39:54 is there a way to stress test my cpu to see if I have sufficient cooling? 2017-11-21 04:41:14 Will this be enough - md5sum /dev/urandom? 2017-11-21 05:37:08 trying to follow wiki for local apk cache from: https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management 2017-11-21 05:40:06 ran setup-apkcache to make link in /etc/apk/cache to local directory on drive. now what? apk wont load any of the packages there 2017-11-21 05:43:12 <_ikke_> "Note that apk will ignore any cache residing on a tmpfs volume." 2017-11-21 05:43:16 <_ikke_> Is that the case? 2017-11-21 05:43:27 no local drive 2017-11-21 05:43:53 directory on rootfs 2017-11-21 05:46:07 <_ikke_> How did you determine it doesn't get packages from there? 2017-11-21 05:47:16 it says any package I try to add is (missing) 2017-11-21 05:48:00 do we still need to use the "install loccal package" syntax? 2017-11-21 05:49:20 <_ikke_> how did you add packages there? 2017-11-21 05:49:28 <_ikke_> in the cache dir? 2017-11-21 05:49:57 copy from cd boot_repo 2017-11-21 05:50:23 <_ikke_> there is an `apk cache download` command 2017-11-21 05:54:09 that is to "replace missing packages" according to the help 2017-11-21 05:54:34 <_ikke_> it basically just fills the cache 2017-11-21 05:54:45 or clean out unused ones, which I dont want to do until I install them 2017-11-21 05:55:20 <_ikke_> jammers: is there an APKINDEX* file in your cache dir? 2017-11-21 05:55:27 typing that gave no error or output whatsoever 2017-11-21 05:56:10 there was, I removed it trying to get things working 2017-11-21 05:56:19 will put it back now 2017-11-21 05:56:34 and try again, but it was there before with no love 2017-11-21 05:57:21 <_ikke_> jammers: and you ran apk update? 2017-11-21 05:57:58 <_ikke_> (I'm not too familiar with the semantics of apk cache, so trying to find out with you :) ) 2017-11-21 05:58:35 update says 17 packages avai;able, bit thats how many are actually installed, so somethings amiss 2017-11-21 05:59:37 and TW for helping BW :) 2017-11-21 05:59:41 BTW 2017-11-21 06:08:39 only tried going the acahe route as manual apk load is broken if requested apk has deps, even if dep pkgs are in the same directory. how sad... 2017-11-21 06:08:45 cache 2017-11-21 06:11:13 <_ikke_> jammers: apk does rely on an index, it does not look in a dir to see what packages are available 2017-11-21 06:12:01 put the APKINDEX.tar.gz back on the cache dir. no change 2017-11-21 06:12:35 <_ikke_> and then apk update? 2017-11-21 06:13:59 still says only 17 packages avail 2017-11-21 06:14:28 there are over 50 in the folder 2017-11-21 06:16:19 <_ikke_> I don't know then, best wait until someone who does knows can answer here 2017-11-21 06:16:52 jammers: did you add the directory to /etc/apk/repositories? 2017-11-21 06:27:09 GEEZ, thats all it was...??!!! shouldn't the setup-apkcache script have done that? 2017-11-21 06:28:19 but TW j_v :) 2017-11-21 06:30:44 jammers: np 2017-11-21 06:33:14 could always submit a patch for it through the bug tracker: https://bugs.alpinelinux.org/projects/alpine/issues 2017-11-21 13:38:45 hey there can someone help me with the apk command ? I want to install nginx to a given path with apk add nginx -p /to/my/path but it always shows an the error: ERROR: Unable to lock database: No such file or directory any suggestions ? A normal installation works btw 2017-11-21 13:39:44 Guest4753, -p or --root assumes the directory is a chroot 2017-11-21 13:39:50 or requires a chroot that is 2017-11-21 13:40:27 <_ikke_> Couldn't apk also read repositories from a directory? 2017-11-21 13:42:02 ah yeah that could be the problem thanks ! 2017-11-21 16:13:52 I discovered that one of my 3 drives in a raid 5 array is dying last night :( 2017-11-21 16:15:09 what's weird (to me, at least) is that when it "hiccups", the whole system goes unresponsive for a while 2017-11-21 16:15:32 alpine is installed on a separate, non-raid, ssd 2017-11-21 16:16:11 it does this regardless of what I'm doing, or where I'm "at" in the filesystem 2017-11-21 16:25:16 only 3 drives in a raid5 is sketchy, no hotspare or extra parity? 2017-11-21 16:25:20 is it mdadm raid5 or hardware? 2017-11-21 16:25:35 mdadm 2017-11-21 16:26:10 interesting, how's the drive dying? i assume it's entirely going unresponsive rather than just some reallocated/failed sectors? 2017-11-21 16:26:28 it's my home server; I've run it on 2 drives for a while before xD 2017-11-21 16:27:36 smartd[23559]: Device: /dev/sdd [SAT], 2 Currently unreadable (pending) sectors 2017-11-21 16:31:02 I thought maybe I had a swap file (yes, file) on the raid, but I don't 2017-11-21 16:31:51 <_ikke_> shodan45: right, the latter was what I was thinking about 2017-11-21 16:32:27 yea that was my first concern, i have had cases where drives that are failing in a more severe manner have triggered similar symptoms 2017-11-21 16:32:41 but just reallocated sectors shouldn't trigger such problems in my experience 2017-11-21 16:33:15 I *was* doing some intensive-ish I/O on it - deleting a failed backup with rdiff-backup, but that was in another session 2017-11-21 16:34:31 its possible that's relevant, if the drive went unresponsive in the middle of an in-flight request it could've hung the process briefly in D state, which might affect other running tasks 2017-11-21 16:34:58 I was just ssh'd in, typing something at a bash prompt, and it just stops responding for 10+ seconds 2017-11-21 16:35:37 I just expect linux to handle this better :) 2017-11-21 16:35:50 or maybe it's a config thing? 2017-11-21 16:36:01 what type of drives? 2017-11-21 16:36:06 process and/or i/o scheduler? 2017-11-21 16:36:30 scv: cheap consumer-level 1.5tb 2017-11-21 16:36:33 certain models will attempt to recover for extended periods of time and can cause significantly worse symptoms than drives that are intended to run in RAID and have a short timeout 2017-11-21 16:36:42 yea, so that might be related 2017-11-21 16:37:56 Model Family: Western Digital Caviar Green 2017-11-21 16:38:09 yea definitely no TLER support :p 2017-11-21 16:40:29 like I said, it's just my home server... not a single point of failure in a multi-million dollar operation ;) 2017-11-21 16:40:53 yea, but the price difference between wd green and red is minimal 2017-11-21 16:41:00 yet reds are much more reliable and are compatible with raid 2017-11-21 16:41:07 and I've been meaning to upgrade the drives for a long time 2017-11-21 16:41:34 personally when i have spindle arrays i still use wd reds a lot of the time, even in production 2017-11-21 16:41:53 either wd red or hgst sas, both very affordable and highly reliable 2017-11-21 16:42:22 I've been (and still am, somewhat) financially "sucker punched" 2017-11-21 16:43:41 I was even thinking of getting one (or 2) of these: https://www.newegg.com/Product/Product.aspx?Item=1Z4-001J-00224 2017-11-21 16:43:44 :( been there, so i know the pain 2017-11-21 16:43:54 softraid is the best solution on a budget 2017-11-21 16:43:59 i highly recommend those 2017-11-21 16:44:12 i'm running an array of 8 of em in my personal storage machine 2017-11-21 16:44:14 those are refurbs 2017-11-21 16:44:14 that exact same model 2017-11-21 16:44:19 yep 2017-11-21 16:44:25 i'm usin refurbs too 2017-11-21 16:44:33 oh. well then. :) 2017-11-21 16:46:05 I might just get one then... but first, I need to get to work :/ 2017-11-21 16:47:00 bb 2017-11-21 16:47:01 :) 2017-11-21 16:54:56 Hi 2017-11-21 16:55:05 what is this mean? 2017-11-21 16:55:15 ```[ 5955.143772] grsec: denied resource overstep by requesting 135168 for RLIMIT_MEMLOCK against limit 65536 for /usr/bin/mpv[mpv:11637] uid/euid:1000/1000 gid/egid:1000/1000, parent /bin/busybox[sh:11635] uid/euid:1000/1000 gid/egid:1000/100``` 2017-11-21 16:55:21 ? 2017-11-21 16:55:44 process freezes for a few seconds 2017-11-21 16:55:46 there is a resource limit and mpv tried to exceed it 2017-11-21 16:55:47 also 2017-11-21 16:56:11 <_ikke_> "The maximum number of bytes of memory that may be locked into RAM. In effect this limit is rounded down to the nearest multiple of the system page size. " 2017-11-21 16:57:06 but 2017-11-21 16:57:25 ACTION sent a long message: cypa[m]_2017-11-21_16:57:24.txt  2017-11-21 16:57:27 ``` 2017-11-21 16:58:14 so how do I increase that limit? 2017-11-21 16:58:16 thats just your total memory, this is a per-process limit enforced by grsec 2017-11-21 16:58:18 <_ikke_> "locked into RAM" 2017-11-21 16:58:39 I see ... 2017-11-21 16:59:17 firefox and mpv affected by this 2017-11-21 17:08:37 no ulimit binary for Alpine? 2017-11-21 17:09:55 <_ikke_> type ulimit -> ulimit is a shell builtin 2017-11-21 17:16:32 $ sudo ulimit -l 1024 2017-11-21 17:16:33 sudo: ulimit: command not found 2017-11-21 17:16:54 $ ulimit -l 1024 2017-11-21 17:16:54 ash: error setting limit: Operation not permitted 2017-11-21 17:21:42 <_ikke_> ulimit is just like cd, you cannot sudo a builtin 2017-11-21 17:23:38 but with root privs it changes root's ulimit but not of mine user 2017-11-21 17:24:18 how do I change ulimit -l for a user, not root 2017-11-21 17:24:19 ? 2017-11-21 17:24:30 no /etc/limits.conf filex 2017-11-21 17:24:35 *file 2017-11-21 17:29:47 ```$ sudo paxctl -zxe /usr/bin/mpv 2017-11-21 17:29:48 file /usr/bin/mpv does not have a PT_PAX_FLAGS program header, try conversion 2017-11-21 17:29:51 ``` 2017-11-21 17:29:54 weird 2017-11-21 17:37:24 you can set limits in pam module as well.. 2017-11-21 17:41:49 <_ikke_> alpine doesn't come with pam 2017-11-21 17:42:21 cypa[m], paxctl is deprecated, use paxmark instead 2017-11-21 17:42:31 paxmark uses XATTRs where as paxctl modifies the file 2017-11-21 17:43:30 no pam installed indeed 2017-11-21 17:44:14 no man for paxmark 2017-11-21 17:44:16 ? 2017-11-21 17:44:17 ? 2017-11-21 17:45:17 <_ikke_> apparently just a binary (or script) 2017-11-21 17:45:25 cypa[m]: manpages are in usually -doc 2017-11-21 17:45:39 <_ikke_> there is no -doc for paxmark 2017-11-21 17:45:46 <_ikke_> https://git.alpinelinux.org/cgit/aports/tree/main/paxmark/paxmark 2017-11-21 17:49:05 btw, is it imposible to change ulimit -l at all? 2017-11-21 17:49:31 with paxmark one could disable MPROTECT which is another way 2017-11-21 17:53:27 ? 2017-11-21 18:01:37 $ cat /etc/limits.conf 2017-11-21 18:01:38 * soft memlock 128 2017-11-21 18:01:38 * hard memlock 1024 2017-11-21 18:01:49 and reboot give no luck 2017-11-21 18:02:01 memlock is 64 still 2017-11-21 18:03:35 btw 2017-11-21 18:03:42 ```[ 97.123720] xfsettingsd[2692]: segfault at 0 ip 00007d7b25b9d66a sp 00007fe1fe8f2370 error 4 in libupower-glib.so.3.0.1[7d7b25b8e000+221000] 2017-11-21 18:03:43 [ 97.123758] grsec: Segmentation fault occurred at (nil) in /usr/bin/xfsettingsd[xfsettingsd:2692] uid/euid:1000/1000 gid/egid:1000/1000, parent /bin/busybox[init:1] uid/euid:0/0 gid/egid:0/0 2017-11-21 18:03:43 [ 97.123818] grsec: denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/bin/xfsettingsd[xfsettingsd:2692] uid/euid:1000/1000 gid/egid:1000/1000, parent /bin/busybox[init:1] uid/euid:0/0 gid/egid:0/0 2017-11-21 18:03:45 ``` 2017-11-21 18:03:58 too many enforced limits or not? 2017-11-21 18:05:06 ah, grsec 2017-11-21 18:05:18 wrong channel 2017-11-21 18:05:30 is there a cure for this? 2017-11-21 18:05:47 or Alpine linux is not suitable for desktop? finally 2017-11-21 18:06:19 Xe: what do you mean wrong channels 2017-11-21 18:06:25 *channel 2017-11-21 18:06:26 ? 2017-11-21 18:06:41 i put my line in the wrong channel 2017-11-21 18:06:59 ok 2017-11-21 18:37:23 cypa[m] try to increase limits? would that help? 2017-11-21 18:40:32 heck ... how to do it? 2017-11-21 18:40:51 /etc/security/limits.conf do not work 2017-11-21 18:41:43 ACTION sent a long message: cypa[m]_2017-11-21_18:41:42.txt  2017-11-21 18:42:56 and no pam modules on Alpine 2017-11-21 18:48:55 btw, am I the only one who have this problem? 2017-11-21 18:58:46 check this: https://mtyurt.net/2017/04/06/docker-how-to-increase-number-of-open-files-limit/ 2017-11-21 19:29:56 thats not help 2017-11-21 21:18:18 cypa[m]: grsec is a series of hardening patches to the kernel, try installing linux-vanilla and see if you can replicate the problem? (installing linux-vanilla instead of linux-hardened is normally something that shouldn't be done unless you're running into problems like this you can't get around) 2017-11-21 22:29:26 hello! Is anyone here? 2017-11-21 22:30:55 maybe.. ? 2017-11-21 22:30:57 hello! dalias - do you think you could help me with an alpine question? 2017-11-21 22:31:04 whatsup? 2017-11-21 22:32:20 so, I know very little about Alpine, but I am using it to access some old emails I have on a server. I just want to forward all of the email in all of the inboxes to my regular gmail but I don't know how to do it it in bulk/mass. I am stuck forwarding them one by one. I tried using ";" followed by "A" to select all but that doesn't let me forward. 2017-11-21 22:32:23 Any suggestions? 2017-11-21 22:32:53 Alpine is a distro, not a mail client 2017-11-21 22:33:11 hah, bad naming, yeah 2017-11-21 22:33:20 RushNYC: this means this is the wrong alpine channel 2017-11-21 22:33:20 <_ikke_> alpinelinux to be exact 2017-11-21 22:33:22 ...Oh, ok. I need help with Alpine the mail client. Anyways, perhaps someone here knows? 2017-11-21 22:33:43 I just want to copy all my emails off in the most efficient way. 2017-11-21 22:34:08 I don't know, but find a better service provider than gmail 2017-11-21 22:36:59 heh, you can check Alpine linux How-to there are some docs related to mail setup, you might find what you need.. 2017-11-22 02:05:16 https://paste.ee/p/GGBJr#Rbq0d6Fbox4c8i1XvgWKFw5KGn3QRpkJ 2017-11-22 02:05:50 the source that the perl-mailtools package on main in 3.6 and edge is trying to fetch is 404 2017-11-22 02:07:40 correct url is: https://cpan.metacpan.org/authors/id/M/MA/MARKOV/MailTools-2.18.tar.gz 2017-11-22 02:08:19 i should probably submit my apkbuild-cpan patches 2017-11-22 02:08:28 adds support for metacpan since cpan is busted af for some packages 2017-11-22 02:08:35 and returns entirely different tarballs 2017-11-22 02:08:37 for unrelated modules 2017-11-22 02:11:20 huh 2017-11-22 02:11:22 ok 2017-11-22 02:11:27 that package even passes the checksum 2017-11-22 02:11:31 pr incoming 2017-11-22 02:37:08 https://github.com/alpinelinux/aports/pull/2787 2017-11-22 02:37:10 listen 2017-11-22 02:37:15 that travis shit aint my fault 2017-11-22 02:37:22 kaniini: ^ 2017-11-22 02:38:32 or whoever looks at that stuff 2017-11-22 03:04:03 mepholic: https://travis-ci.org/alpinelinux/aports/builds/305595721#L620 2017-11-22 03:04:23 oh 2017-11-22 03:04:28 i completely looked over that line 2017-11-22 03:04:28 you need to either add function check that runs the test suite or add `options="!check"` to the abuild 2017-11-22 03:04:32 i know 2017-11-22 03:04:39 still not my fault!!!!!!!!!!! 2017-11-22 03:04:43 hold on 2017-11-22 03:08:34 lets see if that does it 2017-11-22 03:10:21 boom 2017-11-22 03:10:23 i feex 2017-11-22 03:10:58 thanks Xe 2017-11-22 08:58:12 <_ikke_> We've added edge as a pinnen repository, but it causes dependency issues when installing non-pinned packaged. Shouldn't pinned repositories prevent at least this? 2017-11-22 08:58:21 <_ikke_> pinned* 2017-11-22 08:58:42 <_ikke_> apk add -U --no-cache 2017-11-22 08:59:02 <_ikke_> gives an unsatisfiable constraints error on musl 2017-11-22 09:03:52 <_ikke_> "Apk will now by default only use the untagged repositories.." 2017-11-22 09:10:01 morning 2017-11-22 09:11:27 <_ikke_> morning 2017-11-22 09:11:40 can you paste the error? 2017-11-22 09:12:30 need error message plus command line plus /etc/apk/{world,repositories} 2017-11-22 09:12:51 <_ikke_> It's on the alpine:latest docker image 2017-11-22 09:13:38 <_ikke_> https://pastebin.com/raw/ZgvZBSR8 2017-11-22 09:13:41 <_ikke_> That's the error message 2017-11-22 09:14:45 need to do apk upgrade first 2017-11-22 09:15:02 Hi, everybody, 3.7 should be released 2017.11.1, why delayed ? 2017-11-22 09:15:03 the image has stale packages 2017-11-22 09:15:32 <_ikke_> shouldn't the -U do that? 2017-11-22 09:15:41 no, -U runs apk update 2017-11-22 09:15:46 <_ikke_> oh right 2017-11-22 09:15:58 does that combo even make sense? 2017-11-22 09:16:08 it does, in limited cases 2017-11-22 09:16:12 -U --no-cache 2017-11-22 09:16:16 unfortunately docker people are sloppy 2017-11-22 09:16:25 in their dockerfiles 2017-11-22 09:16:54 the --no-cache seems pointless on docker 2017-11-22 09:17:08 but maybe i am missing something :) 2017-11-22 09:17:19 <_ikke_> The idea was to prevent files being left behind after the image was buitl 2017-11-22 09:17:21 <_ikke_> buitl 2017-11-22 09:17:23 <_ikke_> built 2017-11-22 09:17:35 <_ikke_> But I might have confused the what --no-cache does 2017-11-22 09:18:23 <_ikke_> But thanks, that solves it 2017-11-22 09:18:46 i guess it doesnt use the apk index from disk but from repo 2017-11-22 09:19:15 <_ikke_> We removed the --no-cache 2017-11-22 09:19:34 the problem is that `apk add` by default doesn't really want to upgrade any packages 2017-11-22 09:19:34 <_ikke_> and added a separate upgrade -U commadn 2017-11-22 09:19:40 when attaching to a virtual 2017-11-22 09:19:59 because that's weird 2017-11-22 09:20:07 you don't want a package upgrade as a side effect in those cases 2017-11-22 09:20:19 <_ikke_> Was more a copy-paste from someone 2017-11-22 09:20:23 (maybe you do when building docker images, but that again, doesn't seem like a very good idea to me) 2017-11-22 09:21:12 arguably, apk is doing the right thing and telling you that you're doing something weird 2017-11-22 09:21:13 (: 2017-11-22 09:21:34 <_ikke_> I agree, I didn't notice that we only did add -U 2017-11-22 15:28:32 hi, i would like to update wiki, how? I do not have permissions.. thnx 2017-11-22 16:07:58 zixor: dont submit pages with urls, or wait 5 hours. 2017-11-22 16:21:23 clandmeter: thnx... 2017-11-22 16:55:48 Xe: I've reported bug http://bugs.alpinelinux.org/issues/8178 2017-11-22 16:56:29 since no one knows how to change ulimit -l for a user 2017-11-22 16:58:22 I shall learn grsec and see if there is a way 2017-11-22 16:58:26 also 2017-11-22 17:00:22 anyway for desktop such a freezes of firefox are undearable and we should provide some tool or info for desktop apline users 2017-11-22 17:00:54 anyway for desktop such a freezes of firefox are unbearable and we should provide some tool or info for desktop apline users 2017-11-22 17:04:16 cypa[m]: you might want to get involved with adelie linux, they are a fork of alpine designed for desktop users 2017-11-22 17:05:03 yeah, the thing with firefox. 2017-11-22 17:05:57 anyone know why 3.7 release has been delayed ? it's should be release at 2017/11/1 2017-11-22 17:06:11 huh? source= 2017-11-22 17:07:05 https://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2017-11-22 17:08:43 wener1: in 2016, it was released in january 2017-11-22 17:09:07 v3.3 2017-11-22 17:09:13 and 3.5 went in late december 2017-11-22 17:09:27 so,any release schdule for 3.7 ? 2017-11-22 17:09:30 why do you think 3.7 should come out on 1st november? 2017-11-22 17:09:50 i think the problem here is some different 2017-11-22 17:10:05 why do you need 3.7 ? 2017-11-22 17:10:24 wener1: honestly if you need packages that are in newer versions of alpine, edge is fairly stable actually, just make sure you patch weekly 2017-11-22 17:10:27 if you want it more like rolling release like, just go on testing 2017-11-22 17:11:46 I setup a lot images based on 3.6, but if 3.7 come out soon, I need to do this again. 2017-11-22 17:11:50 i personally base my containers on edge, but about half of the advantage there is that they're containers and the base filesystem can be rolled back lol 2017-11-22 17:12:16 wener1: you should look into using something like `box-builder` instead of just straight docker 2017-11-22 17:12:23 wener1: if 3.7 comes out, doesn't mean that 3.6 is outdated 2017-11-22 17:12:23 not just for containers, but also server os 2017-11-22 17:13:07 honestly 2017-11-22 17:13:19 i don't think you really need to upgrade to 3.7 the second it comes out 2017-11-22 17:13:24 ^ 2017-11-22 17:13:30 wener1: 3.6 has 2 years support 2017-11-22 17:13:41 if stuff is stable and will be supported until 2019, who cares for now? 2017-11-22 17:13:48 wait until 3.8 comes out and upgrade to 3.7 then 2017-11-22 17:14:10 i'd like to point to the core problem 2017-11-22 17:14:28 that crappy software makes users think they need to have the most recent version of fucking everything 2017-11-22 17:14:34 thx, i'll stick to 3.6 for now. 2017-11-22 17:15:25 liwakura: usually because of ubuntu/debian/centos packaging old versions of everything useful lol 2017-11-22 17:16:09 it took me years to realize that "no updates" sometimes means that software is "finished", not abandoned 2017-11-22 17:17:09 and i encounter alot of people who are totally flabbergasted about my concept of "finished software" 2017-11-22 17:17:35 whatever. im on v3.6 now, lets hope this gets proper security support 2017-11-22 17:17:50 fuck firefox (because its a steaming pile of crap to compile) 2017-11-22 17:25:52 I wanted to setup PAN on my RPi3. Is the 'pand' daemon available? I couldn't find it in any package. 2017-11-22 17:28:08 sw: i dont see it 2017-11-22 18:38:15 <_ikke_> sw: What is it? 2017-11-22 18:43:22 hey, is there a way to tell setup-bootable which kernel flavor to use? Would be cool if I could get zfs-ready usb stick with alpine 2017-11-22 19:39:09 hi people 2017-11-22 19:53:29 hey Venker 2017-11-22 19:53:46 hi kahiru 2017-11-22 19:53:54 whats up? 2017-11-22 19:54:39 deploying alpine, having a Bud... 2017-11-22 19:56:06 sounds like a fine evening 2017-11-22 19:56:10 or whatever it is now in your tz 2017-11-22 19:56:20 it is evening, yes 2017-11-22 19:56:45 what will be the purpose of the machine you're deploying? 2017-11-22 19:57:05 for the moment, I'm only testing 2017-11-22 19:57:35 but I'll like to use alpine in some old machines, running the OS from an SD card 2017-11-22 19:58:28 also in openstack, but I couldn't build a proper kvm image 2017-11-22 19:58:48 what was the issue there? 2017-11-22 20:00:00 the openstack template has to be able to autoresize, and the iso in which I tried to install it, only allowed me to deploy in a fixed partition schema 2017-11-22 20:00:13 as far as I could remember when I did it 2017-11-22 20:01:05 there are templates for centos, debian, ubuntu and even gentoo... but not for alpine 2017-11-22 20:01:25 maybe a build script or package such as Oz would help 2017-11-22 20:02:23 I never had to deal with openstack 2017-11-22 20:02:31 so I didn't 2017-11-22 20:03:01 I have a small openstack account where I try new things 2017-11-22 20:03:15 and alpine has a lot of potential 2017-11-22 20:04:00 there are apks for almost everything and it's lightweigh 2017-11-22 20:04:25 (I miss the burp backup package, BTW) 2017-11-22 20:14:55 <_ikke_> Venker: If you ask someone nicely, they want to package it for you :) 2017-11-22 20:15:30 thanks ;-) 2017-11-22 20:15:52 I'd like also to learn how to package apks by myself 2017-11-22 20:16:11 <_ikke_> https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2017-11-22 20:16:30 <_ikke_> https://wiki.alpinelinux.org/wiki/APKBUILD_Reference 2017-11-22 20:19:25 it seems easier than in Debian... ^_^U 2017-11-22 20:19:45 <_ikke_> It does a lot more than debian 2017-11-22 20:30:43 I think it's more flexible in what reminds to dependencies 2017-11-22 20:32:19 (sorry for my bad english) 2017-11-22 21:35:32 _ikke_: I'm curious about packaging (I've never packaged anything myself though...). I don't understand what your saying - the packaging tools do more work for you (ie you need to do more manually in debian ? or something similar ? 2017-11-22 21:35:40 *you're 2017-11-22 21:35:46 <_ikke_> seancjscherer[m]: exactly 2017-11-22 21:36:05 ah, ok thanks! 2017-11-22 21:36:15 I'll look into it some time :) 2017-11-22 21:36:31 <_ikke_> you mostly need to provide metadata, and some build steps, and abuild does all the rest 2017-11-22 21:37:44 I was following (and using) Solus (the OS) for a while, and they wrote about their packaging tools some - and it sounded in some ways similiar (ie "let the tools do the heavy lifting, easier than for debian") 2017-11-22 21:38:06 (oh, I neither follow nor use Solus anymore, just btw) 2017-11-22 21:43:00 Sorry, just saying that because I thought back then "cool", and so now I'm definitely thinking "Cool !" :D . 2017-11-22 23:01:33 hi guys 2017-11-22 23:01:59 im having a bit of difficulty as a newb, i was wondeing whether anybody could help 2017-11-22 23:05:38 see you, people :) 2017-11-22 23:08:16 anyone got experience using cron with alpine? 2017-11-22 23:09:13 no different than running cron elsewhere 2017-11-22 23:13:34 for some reason i have a script that won't run on start-up 2017-11-22 23:13:38 using cron 2017-11-22 23:13:48 i did crontab -e 2017-11-22 23:13:56 inserted: @reboot "file path" 2017-11-22 23:14:01 but nothing 2017-11-22 23:14:33 does script works? w/o cron? 2017-11-22 23:14:47 permissions maybe? 2017-11-22 23:14:54 is it executable 2017-11-22 23:14:56 ? 2017-11-22 23:15:11 @reboot is not a standard cron feature 2017-11-22 23:15:16 and the stock alpine crond doesnt support it 2017-11-22 23:16:40 @fugitive yeah sorry, it does work without it 2017-11-22 23:16:52 although i was just testing it then and if i was in a differnet folder i wasn't able to execute it 2017-11-22 23:17:34 with the full path? 2017-11-22 23:17:46 yeah 2017-11-22 23:17:53 might depend on being in the current directory then 2017-11-22 23:18:07 which means you should likely be wrapping this into a startup script instead of using cron in the first place 2017-11-22 23:18:27 are there any guides for that? 2017-11-22 23:18:38 it starts a daemon btw if that means anything 2017-11-22 23:19:14 https://wiki.alpinelinux.org/wiki/Writing_Init_Scripts 2017-11-22 23:19:31 yes.. and set it to load on boot by init 2017-11-22 23:20:24 okay thanks 2017-11-22 23:21:37 starting/stopping daemons should be managed by init.. cron is more for... one shoot scripts at time 2017-11-22 23:27:16 ah okay 2017-11-22 23:27:50 so im copying the example given, am I just meant to be replacing the command= value with my own script? 2017-11-22 23:28:00 or are there other things i need to be replacing in the example 2017-11-22 23:28:46 depends, how your script looks like, and what it does :) 2017-11-22 23:36:40 my script has a sh command in 2017-11-22 23:36:52 it basically starts a dhcp server 2017-11-22 23:37:06 so it'll do: sh startGateway 2017-11-22 23:37:16 and then another dhcpd command 2017-11-22 23:49:09 okay i managed to do it 2017-11-22 23:49:18 i didnt use an init.d script, i got too confused lol 2017-11-22 23:49:24 made a script in profile.d 2017-11-22 23:54:18 simpleton: that is wrong. Sorry i don't have time to check your script. The easiest way for you will be to add it in /etc/rc.local . So it will be ran on the boot. 2017-11-22 23:54:55 er, what dhcp server are you using? 2017-11-22 23:55:02 if it's in a package, it already has an init script 2017-11-23 00:25:35 hi 2017-11-23 00:25:51 thanks for keeping grsecurity alive! 2017-11-23 00:26:52 what is the recommended browser for alpine? 2017-11-23 00:28:03 it seems the firefox package is a bit old, maybe i can build other sources with -fPIE etc. enabled? 2017-11-23 00:29:25 edge has a new firefox package 2017-11-23 00:30:55 old is also ok if it is secure, but it seems like the hell of a task for a small team 2017-11-23 00:32:28 ah, edge has 55.0, nice! 2017-11-23 00:34:19 could the source from the debian source package theoretically work? 2017-11-23 00:35:54 fssdd: for firefox quantum or something? 2017-11-23 00:36:42 b/c if so that is the second request for quantum today lol 2017-11-23 00:39:45 no, i just want normal secure browsing. i love the concept of alpine, but debian naturally has a bigger security team, which may be important for a giant mass of sourcecode like firefox. i have to admit i have no idea how the alpine packages are created though. 2017-11-23 00:40:15 lemme grab the APKBUILD for firefox 2017-11-23 00:40:54 https://github.com/alpinelinux/aports/tree/master/testing/firefox <-- APKBUILD is the package definition 2017-11-23 00:44:09 thanks 2017-11-23 06:19:18 so 2017-11-23 06:19:21 $ sudo apk add virt-viewer 2017-11-23 06:19:22 ERROR: unsatisfiable constraints: 2017-11-23 06:19:24 so:libspice-client-gtk-3.0.so.4 (missing): 2017-11-23 06:19:26 required by: virt-viewer-2.0-r0[so:libspice-client-gtk-3.0.so.4] 2017-11-23 06:19:28 in edge 2017-11-23 06:20:36 ah 2017-11-23 06:20:48 it just needs a rebuild 2017-11-23 14:03:12 hey there, is there any chance of backporting nodejs-npm for v3.5? 2017-11-23 14:13:54 3.5 is pretty old by now, I doubt it 2017-11-23 14:14:03 If you need a newer npm and can't upgrade, you might want to try nvm 2017-11-23 14:18:00 oh, I didn't realise npm wasn't split out from nodejs-npm, never mind then :) 2017-11-23 14:39:43 Hi! Where can I read about how to add a self-signed certificate to alpine linux? 2017-11-23 14:40:26 <_ikke_> Do you know how to build packages that get signed? 2017-11-23 14:41:18 No, I don't, my base use case is to make a program which honors system certificates to be able to query my api end point 2017-11-23 14:41:32 which runs inside a alpine linux docker container 2017-11-23 14:41:40 <_ikke_> https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account 2017-11-23 14:57:10 hello guys ! I desire to generate a working linux system that is loaded from u-boot and resides entirely on RAM. I've seen that there is a tool https://git.alpinelinux.org/cgit/mkinitfs that can be used to generate ramdisk images for alpine linux. I have downloaded the tool on my ubuntu host machine but I have not figured out how to proceed. The target HW is an arm 32 bit device 2017-11-23 15:10:15 ilkappe: apline is designed to run entire from RAM 2017-11-23 15:10:20 and be modular 2017-11-23 15:20:57 <_ikke_> And unless you have special needs, there is no need to do anything iwth initramfs 2017-11-23 15:25:02 ow 2017-11-23 15:25:21 that's quite amazing 2017-11-23 15:26:24 but how can I generate the ram image to be used with uboot in order to load a minimal alpine distro ? 2017-11-23 15:26:39 I know it's a big question but I'm pretty new to this world 2017-11-23 15:26:43 <_ikke_> just use setup-bootable for example to create a bootable usb key, boot, run alpine-setup with diskless mode, and you're ready 2017-11-23 15:26:46 so please, be patient 2017-11-23 15:27:35 <_ikke_> ilkappe: all you would actually need is a bootable usb disk for example 2017-11-23 15:28:37 <_ikke_> if you boot alpine, by default it's already from ram 2017-11-23 15:28:55 ilkappe: whats your boot media? usb? sdcard? 2017-11-23 15:30:19 basically, take tha alpine generic armhf tar.gz from https://alpinelinux.org/downloads/ 2017-11-23 15:30:26 put the files on your boot media 2017-11-23 15:30:32 and configure uboot 2017-11-23 15:30:48 how you configure uboot etc depends on what board it is 2017-11-23 15:31:10 we have a tarball for rpi too, with the rpi kernel 2017-11-23 15:31:54 once you get it booted, you'll have it running from RAM 2017-11-23 15:32:12 minimal distro from ram 2017-11-23 15:32:42 with the exception of kernel modules, which we mount a loopback image for 2017-11-23 15:32:53 but you can umount the modloop (module loopback) 2017-11-23 15:33:04 and then you can eject the bootmedia if you want 2017-11-23 15:33:09 works the same on x86 2017-11-23 15:34:18 when you run setup-alpine to "install" alpine it wil will give you an option to run diskless 2017-11-23 15:34:27 just select "none" when it asks for disk 2017-11-23 15:34:54 there is a tool 'lbu', to make a local backup of the config 2017-11-23 15:35:00 which it will look for at boot 2017-11-23 15:35:15 so you can have your package reinstalled and config restored on reboot 2017-11-23 15:35:17 Are there Virtualbox-guest-addtions that have X11 integration? 2017-11-23 15:35:30 azarus: not sure 2017-11-23 15:45:36 ty ncopa 2017-11-23 15:46:11 as now I have one SD card with 3 partition 2017-11-23 15:46:28 1 FAT partition with my custom u-boot.bin 2017-11-23 15:46:45 1 EXT4 partition with a custom raspbian rootfs 2017-11-23 15:46:55 1 EXT4 partition for user data 2017-11-23 15:47:37 what bot is used in alpine-commit (algitbot)? 2017-11-23 15:47:39 i would like to have (from what I have understand) to have an initrams image to be booted from my u-boot 2017-11-23 15:47:49 hubot? 2017-11-23 15:48:24 in order to upgrade the existing raspbian rootfs if needed 2017-11-23 15:48:36 <_ikke_> ScrumpyJack: a custom bot iirc 2017-11-23 15:48:43 <_ikke_> ScrumpyJack: ask clandmeter 2017-11-23 15:48:51 ACTION nods 2017-11-23 15:49:53 i'm trying to get hubot to relay an mqtt sub into an irc channel and wondering how it's done here 2017-11-23 15:50:17 I do not want to have a fs that hosts the file of the minimal distro 2017-11-23 15:51:35 <_ikke_> ilkappe: when alpine boots, it creates a tmpfs where the filesystem is created 2017-11-23 16:09:33 ty ikke 2017-11-23 16:15:10 <_mitch> Hello everyone. 2017-11-23 16:15:44 <_mitch> I'm coming from Sabotage linux, looking to learn more about Alpine 2017-11-23 16:15:52 <_mitch> and just its status of development/community in general 2017-11-23 16:16:56 <_ikke_> _mitch: this lists they key points of alpine: https://alpinelinux.org/about/, in case you did not read that yet 2017-11-23 16:17:06 _mitch: oh hai 2017-11-23 16:17:10 <_mitch> Of course. 2017-11-23 16:17:16 <_mitch> Oh hello liwakura. 2017-11-23 16:17:33 <_mitch> @liwakura: Sabotage has so many problems, its painful to get much accomplished 2017-11-23 16:17:39 <_mitch> I've heard alpine is much father in development 2017-11-23 16:17:48 <_ikke_> _mitch: 3.7 is about to come out 2017-11-23 16:17:52 <_mitch> I'm spending hours every day debugging problems and getting things to compile 2017-11-23 16:18:07 <_mitch> which Im fine with, I mean experimental distros is a hobby 2017-11-23 16:18:11 <_mitch> but man 2017-11-23 16:18:12 <_mitch> gets old 2017-11-23 16:19:30 <_mitch> @_ikke_: Here are some of the things that are not working in Sabotage. Are they easy to get running Alpine? 2017-11-23 16:19:37 <_mitch> Java 8, Haskell, Webkit2gtk, Chromium 2017-11-23 16:19:50 <_mitch> wicd or network manager 2017-11-23 16:20:15 <_ikke_> alpine ships openjdk8 in edge 2017-11-23 16:20:30 <_mitch> i love it. alright 2017-11-23 16:20:51 <_ikke_> ghc is provided for armhf and x86_64 2017-11-23 16:20:56 <_ikke_> in 3.6 2017-11-23 16:21:02 <_mitch> great! 2017-11-23 16:21:21 <_mitch> right now I have stack working in sabotage, but fails ghc compile 2017-11-23 16:21:37 <_ikke_> I hear ghc is difficult to get to work 2017-11-23 16:21:48 <_mitch> unfortunately 2017-11-23 16:21:52 <_mitch> but alpine has it, yeah? 2017-11-23 16:21:54 <_ikke_> yeah 2017-11-23 16:21:57 <_mitch> Seems good 2017-11-23 16:22:04 <_ikke_> http://pkgs.alpinelinux.org lists the packages btw 2017-11-23 16:22:08 <_mitch> Fantastic. 2017-11-23 16:22:18 <_ikke_> webkit2gtk and chromium are both available 2017-11-23 16:22:39 <_ikke_> and the last 2 also 2017-11-23 16:22:54 <_mitch> aside from package availability/ease of use 2017-11-23 16:23:03 <_mitch> what might you argue for alpine vs sabotage? 2017-11-23 16:23:11 <_mitch> I see alpine uses OpenRC where sabotage uses runit 2017-11-23 16:23:12 <_ikke_> I don't know sabotage 2017-11-23 16:23:15 <_mitch> oh really? 2017-11-23 16:23:18 <_ikke_> nope 2017-11-23 16:23:22 <_mitch> http://sabotage.tech 2017-11-23 16:23:30 <_mitch> it's another musl c distro 2017-11-23 16:23:33 <_ikke_> ok 2017-11-23 16:23:46 <_ikke_> alpine has grsec kernels (still) 2017-11-23 16:23:54 <_ikke_> run-from-ram 2017-11-23 16:24:00 <_mitch> Yep, that sounds great. 2017-11-23 16:24:05 <_mitch> I'm not super familiar with it, but I like the concept 2017-11-23 16:24:14 <_ikke_> imho, easy to create packages 2017-11-23 16:24:32 <_mitch> yep, I'm familiar with the syntax 2017-11-23 16:24:45 <_mitch> having looked at them to create packages for sabotage 2017-11-23 16:25:11 <_ikke_> packages install fast 2017-11-23 16:25:18 <_ikke_> apk is really fast 2017-11-23 16:25:31 ^^ 2017-11-23 16:25:34 <_mitch> how can it be "fast" when its a compile based distro 2017-11-23 16:25:40 <_mitch> or do you just mean the package manager searches fast 2017-11-23 16:25:49 what do you mean "compile based" 2017-11-23 16:25:51 <_ikke_> _mitch: it's not compile based 2017-11-23 16:25:55 alpine installs binaries 2017-11-23 16:25:57 <_mitch> O.O wow. 2017-11-23 16:25:59 <_ikke_> packages are precompiled 2017-11-23 16:26:05 <_mitch> I just assumed musl c = compilation 2017-11-23 16:26:13 never make assumptions 2017-11-23 16:26:13 :d 2017-11-23 16:26:18 <_mitch> sabo is compilation based, which is what creates all the headaches 2017-11-23 16:26:28 <_mitch> packages may install on one machine, and not another, without much reason 2017-11-23 16:26:38 <_mitch> same archetecture, etc 2017-11-23 16:26:56 <_ikke_> packages are also small (many subpackages for optional components) 2017-11-23 16:28:11 <_mitch> well i'm convinced. 2017-11-23 16:29:24 <_mitch> @_ikke_: "standard", "min root", or "xen"? 2017-11-23 16:29:30 <_mitch> One of the three looks to be what I want 2017-11-23 16:29:51 <_mitch> does the "standard" not include the grsec? 2017-11-23 16:30:32 standard has grsec 2017-11-23 16:30:39 <_ikke_> vanilla does not have grsec 2017-11-23 16:31:07 <_ikke_> _mitch: standard is probably what you want 2017-11-23 16:31:18 <_ikke_> unless you have a different arch 2017-11-23 21:02:48 _mitch: welcome 2017-11-23 21:05:24 he is gone already 2017-11-23 21:05:31 btw, i was also using sabotage before 2017-11-23 21:05:39 but self-compiling is a pita quite often 2017-11-23 21:37:43 I don't know how Gentoo managed to stay usable for so long being source based 2017-11-23 21:38:06 ... well, as long as you remembered to update at least once per week :D 2017-11-23 21:40:20 related question - what does Void linux look like (I mean, how does it compare - it also being musl based, with binaries and non-systemd ; runit if iirc) 2017-11-23 21:41:06 Ok, package management / compilation is via essentially bash scripts. I guess that partly answers my own question :P 2017-11-23 21:53:56 <_mitch> hey, just installed alpine. 2017-11-23 21:55:12 <_mitch> gcc can't find "stdio.h"? 2017-11-23 22:15:39 <_mitch> what is x11-libs named under alpine? 2017-11-23 22:15:48 <_mitch> i need Xlib.h 2017-11-23 22:16:07 <_ikke_> http://pkgs.alpinelinux.org 2017-11-23 22:16:15 <_ikke_> on the content tab, you can search for files 2017-11-23 22:16:25 <_mitch> right... or use apk search.. 2017-11-23 22:16:58 <_ikke_> apk search only searches for package names 2017-11-23 22:17:30 <_mitch> i cant search for something i dont know the name of 2017-11-23 22:18:34 <_mitch> should be libx11 but gcc doesnt seem to find it 2017-11-23 22:19:04 <_mitch> ahhhhh it was libx11-dev 2017-11-23 22:19:09 <_mitch> nevermind. 2017-11-23 22:19:12 <_ikke_> :-) 2017-11-23 22:20:47 _mitch: btw, i also use alpine linux on my host system 2017-11-23 22:20:58 self-compiling was a bit too time-consuming forme 2017-11-23 22:21:27 and since i was an POSIX-Shell nazi before, my stuff works well with busybox 2017-11-23 22:22:01 seancjscherer[m]: void has both an musl and an glibc variant 2017-11-23 22:22:06 runit instead of openrc 2017-11-23 22:22:16 and gnu coreutils instead of busybox 2017-11-23 22:22:34 i choose alpine over void because of coreutils vs busybox 2017-11-23 23:30:01 ah, k thanks :) 2017-11-23 23:30:20 minimalism all the way then ? 2017-11-23 23:31:13 (sorry, just blurted that out - I'm not really too worried. But always curious :P ). 2017-11-23 23:33:28 I was wondering as well about "ecosystem" and the community. Information, eg on decision processes and who basically is the team running the show is rather sparse both here and there : D . 2017-11-23 23:35:15 But thats a more general question I have :) . And I'm going to be gone in minute, so won't be able to follow up right now... Like I said, thanks for sharing! 2017-11-23 23:56:40 what is the priority default for /etc/apk/repositories? 2017-11-23 23:56:57 like, when i have a package with the same name & version in two different repos 2017-11-23 23:57:10 <_mitch> no keyboard input after startx, i have evdev and libinput installed 2017-11-23 23:57:39 <_mitch> grepping "input" from xorg.0.log gives no results 2017-11-23 23:57:43 _mitch: did you use setup-xorg-base ? 2017-11-23 23:57:48 or manually? 2017-11-23 23:58:01 <_mitch> lmfao wow 2017-11-23 23:58:04 <_mitch> that worked 2017-11-23 23:58:06 setup-xorg-base reminded me that i needed eudev 2017-11-23 23:58:07 <_mitch> i had done it manually 2017-11-23 23:58:19 <_mitch> but it installed pciutils and something else 2017-11-23 23:58:24 <_mitch> i had udev already 2017-11-23 23:58:38 i uninstalled the other utils later 2017-11-23 23:58:44 <_mitch> pciutils and "hwdata-pci" 2017-11-23 23:58:47 <_mitch> now it works 2017-11-23 23:59:03 <_mitch> ive never needed those on other distros 2017-11-23 23:59:11 <_mitch> in fact i normally uninstall usb/pci utils 2017-11-24 00:02:07 <_mitch> "make: tic: command not found" 2017-11-24 00:02:11 <_mitch> what is tic...? 2017-11-24 00:03:13 installing terminfo 2017-11-24 00:03:24 might be part of ncurses package 2017-11-24 00:04:01 <_mitch> Yep. youre right 2017-11-24 00:05:47 is alpine considered GNU/Linux? 2017-11-24 00:06:09 hah, thats what uname -o is for 2017-11-24 00:06:11 answer is "no" 2017-11-24 00:10:26 <_mitch> okay, next problem, trying to open st: "XOpen failed. Could not open input device" 2017-11-24 00:10:38 <_mitch> never seen this error in my life 2017-11-24 00:10:44 <_mitch> what is xopebim 2017-11-24 00:10:55 <_mitch> XOpenIM* 2017-11-24 00:11:23 no idea 2017-11-24 00:11:25 need to sleep 2017-11-24 00:11:26 gn8! 2017-11-24 00:21:38 <_mitch> what package is c99? 2017-11-24 00:22:04 gcc 2017-11-24 00:23:04 <_mitch> i have gcc... 2017-11-24 00:23:46 you mean an actual 'c99' command? 2017-11-24 00:23:54 <_mitch> trying to compile st 2017-11-24 00:23:56 <_mitch> simpleterm 2017-11-24 00:24:11 <_mitch> "/bin/sh: c99 command not found" 2017-11-24 00:24:25 does it not have some sort of configuration detection? 2017-11-24 00:24:46 <_mitch> suckless tools dont come with configure 2017-11-24 00:24:55 i don't mean autoconf in particular 2017-11-24 00:25:04 but literally anything to try and probe tool names 2017-11-24 00:25:24 gcc -std=c99 would be essentially the same so you'll probably want to edit the makefile 2017-11-24 00:26:54 <_mitch> what am i looking for? 2017-11-24 00:27:26 <_mitch> i see "CC $(STCFLAGS) -c $<" 2017-11-24 00:50:43 <_mitch> so i do not have /usr/bin/c99 2017-11-24 00:50:51 <_mitch> what package includes that? 2017-11-24 00:50:54 <_mitch> its not gcc 2017-11-24 00:55:53 there is no /usr/bin/c99 2017-11-24 00:56:03 16:25 < scv> gcc -std=c99 would be essentially the same so you'll probably want to edit the makefile 2017-11-24 00:56:13 usually a variable named CC or similar 2017-11-24 00:58:51 <_mitch> why would this compile fine elsewhere but not on alpine? 2017-11-24 00:59:07 <_mitch> compiled fine on another muslc distro 2017-11-24 01:00:03 <_mitch> found it, fixed and working 2017-11-24 01:00:10 <_mitch> weird 2017-11-24 01:00:39 <_mitch> still wont open for me however 2017-11-24 01:01:04 <_mitch> "XOpenIM failed: Could not open input device" 2017-11-24 01:09:17 doesn't say what's missing? 2017-11-24 01:09:26 or what the path is rather 2017-11-24 01:09:39 <_mitch> nope thats the message 2017-11-24 01:09:45 <_mitch> its st/simpleterm 2017-11-24 01:09:52 <_mitch> never seen this error before 2017-11-24 01:10:23 it should list what device it failed to open though 2017-11-24 01:10:27 a path or something 2017-11-24 01:10:31 an errno 2017-11-24 01:15:38 <_mitch> nope :( 2017-11-24 01:15:47 <_mitch> maybe try lsof or something 2017-11-24 01:15:50 <_mitch> tomorrow 2017-11-24 01:16:17 strace probably 2017-11-24 01:16:24 since it failed to open wont be in lsof : 2017-11-24 09:17:10 Question: I recently wanted to setup Nextcloud ; in [the documentation](https://docs.nextcloud.com/server/12/admin_manual/configuration_server/occ_command.html#run-occ-as-your-http-user) it says to run the configuration manager as html-user . 2017-11-24 09:17:16 Is "html-user" simply a (daemon/system) user with permissions to write at eg var/www/html ?? . Or is there anything else involved ? I realise its a newbie question, sorry ! 2017-11-24 09:22:25 *HTTP-user 2017-11-24 09:38:44 hello guys ! I have downloaded alpine linux for raspberry: http://dl-cdn.alpinelinux.org/alpine/v3.6/releases/armhf/alpine-rpi-3.6.2-armhf.tar.gz. 2017-11-24 09:39:15 I have copied the files contained in the archive in on ext4 partition on my 2017-11-24 09:39:19 sd card 2017-11-24 09:39:52 the very first partition of my sdcard is a FAT partition with my custom u-boot.bin 2017-11-24 09:39:58 and boot.cr 2017-11-24 09:40:26 By the way I can't boot alpine linux correctly. 2017-11-24 09:40:40 In my boot.scr I have used this lines: 2017-11-24 09:40:55 ext4load mmc 0:2 ${kernel_addr_r} /boot/vmlinuz-rpi2 2017-11-24 09:41:12 ext4load mmc 0:2 ${ramdisk_addr_r} /boot/initramfs-rpi2 2017-11-24 09:41:40 bootm ${kernel_addr} ${ramdisk_addr} ${fdt_addr} 2017-11-24 09:42:03 it seems that the kernel has a wrong image type 2017-11-24 09:42:17 Does anyone has any advice ? 2017-11-24 09:42:39 <_ikke_> does the raspberry pi even support uboot? 2017-11-24 09:49:06 yep 2017-11-24 09:49:19 I can boot raspbian correctly with u-boot 2017-11-24 09:49:28 on the third partition 2017-11-24 09:49:30 <_ikke_> ok 2017-11-24 09:50:08 I don't know how to configure u-boot to load alp-linux correctly 2017-11-24 09:50:45 <_ikke_> me neither, sorry 2017-11-24 09:51:00 yep ! 2017-11-24 11:00:46 Hello. I’m running Alpine on a RPi 3, and I’d like to get IPv6 to work. However, I haven’t had any luck yet. It gets a SLAAC address but the default route is missing for some reason: https://pastebin.com/BEsP3wZ6 2017-11-24 11:01:55 Do you have any suggestions for what could be going wrong? At the moment, the RPi is the only device I have trouble with IPv6 2017-11-24 11:19:05 Actually, I just found a solution. echo 2 > /proc/sys/net/ipv6/conf/eth0/accept_ra seems to do the trick 2017-11-24 11:20:05 Now, the question remains that if I put this in the pre-up section for eth0, it doesn’t really persist 2017-11-24 11:20:35 So, I need to do this manually. What could be the issue there? 2017-11-24 11:21:54 <_ikke_> It's odd that you get a slaac address (unless it's on link-local) without ra enabld 2017-11-24 11:31:07 hello guys ! I'm running alpine linux on a raspberry pi3. 2017-11-24 11:31:33 I have made some changes to the /etc folder with respect to the original ones 2017-11-24 11:32:25 how can I make these changes presistent ? I0ve tried lbu commit -v /media/mmcblk0p2 but still nothing. the syntax seems not to be coorect 2017-11-24 11:32:41 <_ikke_> ilkappe: not just lbu commit? 2017-11-24 11:33:02 <_ikke_> Should not be necessary to specify the volume if you set it up 2017-11-24 11:33:50 if i run only lbu commit it exits showing its usage 2017-11-24 11:33:54 :/ 2017-11-24 11:42:04 _ikke_: actually, RA gets enabled by default for some reason 2017-11-24 11:43:18 but, I need it to get «2»-enabled instead of «1»-enabled, and ideally with temporary addresses as well 2017-11-24 11:45:18 <_ikke_> corrideat: RA's are enabled by default in the kernel 2017-11-24 11:47:22 <_ikke_> corrideat: so you have enabled forwarding and still want to accept RA's? 2017-11-24 11:50:04 ilkappe: you might need lbu add /path/of/new/file. If you make some changes, and do lbu st, what do you see? If the file you changed is not on the list, try lbu add’ing it. 2017-11-24 11:51:42 <_ikke_> corrideat_: did you see my last messages? 2017-11-24 11:52:38 the last message I saw was «so you have enabled forwarding and still want to accept RA’s?» 2017-11-24 11:53:03 <_ikke_> right 2017-11-24 11:53:37 did you see my response to that? 2017-11-24 11:53:40 <_ikke_> no 2017-11-24 11:54:15 I said yes, because I’m running some Docker containers there 2017-11-24 11:54:50 <_ikke_> right, so ip_forward disables RA, unless what you see you set it to 2 2017-11-24 11:55:00 <_ikke_> You can do that on specific interfacecs 2017-11-24 11:55:36 right. The issue is that the pre-up section in the network config file gets ignored on boot 2017-11-24 11:56:03 I mean, it probably runs, but then something overwritten by something else 2017-11-24 11:56:28 <_ikke_> if ip_forward is enabled afterwards, that would reset it 2017-11-24 11:56:36 <_ikke_> see: https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt 2017-11-24 12:00:19 Thanks. I’ll now try setting ip_forward first. I’m not sure when ip_forward gets set, but it must be after the interface is up 2017-11-24 12:02:33 Setting forwarding explicity and before accept_ra didn’t seem to help 2017-11-24 12:04:45 <_ikke_> corrideat_: dockerd might set it 2017-11-24 12:07:05 It might, but I disabled most of docker’s most auto configuration stuff. More importantly, if forwarding is 1 already it shouldn’t matter 2017-11-24 12:07:19 «This variable is special, its *change* resets all configuration» 2017-11-24 12:15:20 explicitly echo 1 > /proc/sys/net/ipv6/conf/all/forwarding also doesn’t help. Actually, manually changing /proc/sys/net/ipv6/conf/all/forwarding doesn’t reset the others. 2017-11-24 12:18:06 I think it’s only ipv4 ip_forward that causes resetting 2017-11-24 12:27:17 I suppose I can fix this with a startup script, since the network interface status doesn’t change often 2017-11-24 12:27:26 Still ugly though 2017-11-24 12:53:52 corrideat: I've added the file using lbu add. running lbu status -a I can see the previous file marked as A along with some other files I don't care about. 2017-11-24 12:54:07 by the way running lbu commit always exits printing its usage 2017-11-24 12:56:52 can you try lbu ci? 2017-11-24 12:57:04 or lbu_commit? 2017-11-24 13:01:57 i've tried lbu commit 2017-11-24 13:02:13 with no other arguments. I only get the usage printed outù 2017-11-24 13:06:09 ilkappe: lbu ci or lbu_commit are alternative ways of calling lbu commit 2017-11-24 13:07:35 trying to configure the printer with cups but does not found my printer, kernel messages are ok, what im missing? 2017-11-24 13:08:32 corrideat_: ok, by the way still can't commit. Maybe there is r/w permission issue somewhere.. 2017-11-24 13:10:13 ilkappe: possibly. You can try mounting the SD RW and commiting 2017-11-24 13:10:37 mount -oremount,rw /dev/mmcblk0p1 2017-11-24 13:12:00 I’ve had that issue (lbu commit showing usage) before, but I don’t remember what caused it or how I fixed it 2017-11-24 13:12:09 doh ! 2017-11-24 13:12:51 you shouldn’t need to remount RW though. Alpine does it automagically 2017-11-24 13:13:13 ok is permission related ive changued to 0666 how do i make this permanent in /etc/mdev.conf? 2017-11-24 13:13:41 lp[0-9]... i suppose? 2017-11-24 13:13:59 ilkappe: lbu add /etc/mdev.conf? 2017-11-24 13:14:15 corrideat_: no success also with rw 2017-11-24 13:15:01 ilkappe: Sorry, I got confused with SoHTaGa messages 2017-11-24 13:15:29 lbu add works, by the way since I did not modify /etc/mdev.conf if I run lbu status -a there is not a corrisponding entry 2017-11-24 13:17:00 SoHTaGa: maybe you can just add your user to the lp group 2017-11-24 13:17:13 corrideat_: probably my fault I've seen it only now.. lbu also says: Create a backup of config to writable media 2017-11-24 13:17:43 ilkappe: is this a fresh install? 2017-11-24 13:17:44 still have to modify mdev.conf as owner of /dev/usb/lp0 is root:root 2017-11-24 13:17:48 yup 2017-11-24 13:18:06 fresher than ever 2017-11-24 13:19:03 SoHTaGa: usb/lp[0-9]* root:lp 660 should work 2017-11-24 13:20:33 should i ok 2017-11-24 13:20:50 * 2017-11-24 13:21:06 SoHTaGa: I think so, but I haven’t tried it 2017-11-24 13:22:13 ilkappe: what’s LBU_MEDIA in /etc/lbu/lbu.conf. More importantly, does it have the correct value? 2017-11-24 13:22:47 it's all commented out 2017-11-24 13:23:10 ilkappe: try LBU_MEDIA=mmcblk0p1 2017-11-24 13:23:19 ok ty 2017-11-24 13:27:05 corrideat_: now the warning is no longer shown, but still can't commit 2017-11-24 13:27:10 only usage printed out 2017-11-24 13:27:12 aaaaaaaaaaaa 2017-11-24 13:27:31 ? 2017-11-24 13:30:14 how do i generate a ppd? 2017-11-24 13:32:07 ilkappe: did you solve your issue? 2017-11-24 13:32:49 SoHTaGa: I’m not sure. I think generally you get them from your printer manufacturer? 2017-11-24 13:33:16 idk 2017-11-24 13:33:38 SoHTaGa: or try http://www.openprinting.org/drivers 2017-11-24 13:34:24 alpine can only print with cups? 2017-11-24 13:34:32 SoHTaGa: You can probably make your own too, but you need to know how your printer works and find the PPD specification 2017-11-24 13:35:00 SoHTaGa: if you don’t find an appropriate PPD, maybe ‘generic PostScript printer’ works 2017-11-24 13:35:32 i dont have choices 2017-11-24 13:35:51 the configuration page does not allow me to select manufacturer and model 2017-11-24 13:36:02 thats why i searching directly for the ppd 2017-11-24 13:36:14 also tried :cups-genppd.5.2 -p ~/ "EPSON Stylus CX3200" 2017-11-24 13:36:17 Driver not found: EPSON Stylus CX3200 2017-11-24 13:38:45 corrideat_: nope, still struggling 2017-11-24 13:39:19 SoHTaGa: gutenprint (http://gimp-print.sourceforge.net/p_Download.php) should work for your printer 2017-11-24 13:40:10 i have installed gutenprint 2017-11-24 13:40:19 but does not have any ppd file 2017-11-24 13:40:39 and the generator does not work or i dont know the model exact name to it to generate 2017-11-24 13:42:01 acccccccccccccccccccc 2017-11-24 13:42:04 got it ! 2017-11-24 13:42:28 corrideat_: setup-lbu seems to be a nice choice... 2017-11-24 13:52:16 SoHTaGa: I see. I haven’t printed from Alpine before. The name of the printer may be Stylus-CX3200 or Stylus CX3200 2017-11-24 13:53:09 The easiest thing (without figuring out how to generate the PPD from Alpine) I think would be to install, for example, Debian, and copying the ppd file to Alpine 2017-11-24 13:53:18 ilkappe: glad to hear 2017-11-24 13:53:53 well im extracting the ppd from debian gutenprint package 2017-11-24 13:55:14 I looked into that. It seems like the gutenprint package is a wrapper that generates them locally. Maybe I looked at the wrong package 2017-11-24 13:57:05 OTOH, generating them locally could fix your problem as well 2017-11-24 14:02:09 ok already configured but now i want to use a program to print with options, firefox printing pdf files does not alow me to changue any option like back and white etc.. 2017-11-24 14:07:34 SoHTaGa: I believe that’s got to do with your desktop environment and the options packages are compiled with 2017-11-24 14:08:19 my dektop environment is openbox + fbpanel 2017-11-24 14:08:28 i3wm :3 2017-11-24 14:08:30 One more thing and I think I'm done.. I have to enable thw watchdog service that is /etc/init.d/watchdog. It needs to variables, so if I simply run service watchdog starts it fails. I've added them in /etc/default/watchdog but this does not seem the place 2017-11-24 14:08:32 I’d expect the options for this would be limited in Alpine, but, as I said, I don’t use Alpine for printing 2017-11-24 14:08:43 Any suggestions ? 2017-11-24 14:08:45 and independent programs as xpp dont have a package in alpine 2017-11-24 14:08:59 or gpr 2017-11-24 14:09:17 anyway ive sended a job and the printer is idle 2017-11-24 14:09:21 does not work 2017-11-24 14:09:59 SoHTaGa: do you get an error? 2017-11-24 14:10:03 the job i printed says its completed but the printer has not printed anythig 2017-11-24 14:10:07 no 2017-11-24 14:10:11 ilkappe: /etc/conf.d/watchdog is probably it 2017-11-24 14:10:18 epson-1 Desconocido Retenido 11030k 1 completado el 2017-11-24 14:10:21 Fri Nov 24 15:04:07 2017 2017-11-24 14:12:11 SoHTaGa: try restarting the printer or the job. Normally you’ll get some kind of error, even if CUPS errors tend not to be very helpful 2017-11-24 14:13:21 You can also add LogLevel debug to your cupsd.conf 2017-11-24 15:19:22 this is just sad, https://www.phoronix.com/scan.php?page=news_item&px=Munich-Back-To-Windows 2017-11-24 15:19:31 yeah 2017-11-24 16:00:44 whats the difference between depends and makedepends, i mean i know but for instance im writting a package for xpp has as dependencies cups and fltk and fltk-fluid (dont know if more as my last error was not finding fluid) the point is that ive put cups in depends and is a runtime dependency but have moved fltk from depends to makedepends as fltk-dev expecting is gonna be linked statically (thats what 2017-11-24 16:00:47 really happens? so i dont need at run time this librarythat is only going to be used by this program in my system??) 2017-11-24 16:09:42 some libraries are needed during build but not during operation, e.g. -dev packages for headers 2017-11-24 16:09:51 makedepends is fpr those. 2017-11-24 16:09:54 for* 2017-11-24 16:30:02 hi 2017-11-24 16:50:55 Hello, stupid question: How do I rename the source file in an APKBUILD file? (currently is just the $pkgver.tar.gz, from github) 2017-11-24 16:51:38 Is there something I can do with the source variable or do I rename it from a function that happens before that check? 2017-11-24 16:52:58 Never-mind, found the reference on the wiki, thanks 2017-11-24 17:19:45 hi 2017-11-24 17:20:14 what is the difference betweens packages with `hardened` and those without it? 2017-11-24 17:20:19 -s 2017-11-24 17:23:24 <_ikke_> dffdd: are there packes other then kernels with hardened? 2017-11-24 17:24:21 not sure 2017-11-24 17:25:13 <_ikke_> apparently. yes 2017-11-24 17:25:32 <_ikke_> https://pkgs.alpinelinux.org/packages?name=*-hardened&branch=v3.6&repo=&arch=x86_64&maintainer= 2017-11-24 17:26:16 <_ikke_> So either kernels and modules 2017-11-24 17:26:53 <_ikke_> hardened contains grsec 2017-11-24 17:27:26 are ok, thanks 2017-11-24 17:27:33 *ah ok 2017-11-24 17:28:37 so the repo supports https? that is very cool and secure. 2017-11-24 17:32:50 <_ikke_> dffdd: It doesn't add that much security though 2017-11-24 17:32:56 <_ikke_> all packages are signed 2017-11-24 17:33:11 <_ikke_> and the index also 2017-11-24 17:34:23 well i remember some guy on youtube talking about getting arbitrary code execution by crafting a package - it is not bad to have multiple layers of security 2017-11-24 17:35:31 to be fair he a to switch off a memory randomization feature though 2017-11-24 17:41:50 the problem was that tar did some work before the signature was checked iirc 2017-11-24 17:42:28 its hard to avoid slight mistakes like this all the time, it is easier to use https :-) 2017-11-24 18:20:25 so I finally got around to ditching rhel and installing alpine on my home server 2017-11-24 18:25:11 I'm really glad I did that 2017-11-24 18:25:23 edge or a stable release? 2017-11-24 18:26:50 I went with stable 2017-11-24 18:32:10 mm, you're gonna get rock-solid stability 2017-11-24 18:35:25 that's exactly what I'm hoping for 2017-11-24 21:08:53 hello. i look for a cd iso burner package like cdrecord, for terminal, but do not see one in the packages 2017-11-24 21:14:42 oh i need to use asterisk in package search - found cdrkit 2017-11-24 21:53:27 it is great so far. but when i run a command with --help |less or |more the help scrolls past screen 2017-11-24 21:53:33 in console 2017-11-24 21:59:11 try cdrkit --help 2>&1 | less 2017-11-24 21:59:29 i believe they are weird and print usage info on stderr instead of stdout 2017-11-24 22:04:06 thank you i keep forgetting how to redirect stderr 2017-11-24 22:11:21 do you know an alpine program that can play video in console/framebuffer? 2017-11-24 22:16:45 i am on 3.6 i386 and apk search does not find mplayer 2017-11-24 22:17:04 only avail for x86_64 it seems 2017-11-24 22:20:04 i suppose i dont need video 2017-11-24 22:20:08 ttyl, thanks 2017-11-25 00:28:27 Is it possible to cryptomine on the Alpine OS? 2017-11-25 03:23:48 You'll use loads of power and waste loads of money, which makes mother earth very sad. 2017-11-25 06:30:21 hello 2017-11-25 07:45:56 my printer does not print anything 2017-11-25 07:46:06 cups is working 2017-11-25 07:46:12 jobs are sent 2017-11-25 07:46:28 it says jobs completed 2017-11-25 07:46:33 but nothing was printed 2017-11-25 08:13:32 cups-genppd.5.2 -M 2017-11-25 08:13:38 does not give any output 2017-11-25 08:14:01 im using a ppd for my printer extracted from the debian gutenprint package 2017-11-25 08:14:08 but it seems does not work 2017-11-25 08:44:37 in the coniguring printersection of cups i cant select epson as manufacturer, just xerox Generic HP Index Raw Ricoh 2017-11-25 08:45:06 if i specify the ppd ive extracted from debian package directly it adds my printer but thendoes not work when send a job 2017-11-25 08:45:19 its loged as printed without problem 2017-11-25 08:45:25 but ht printer does nothing 2017-11-25 08:45:29 any idea? 2017-11-25 12:17:26 hi, I've set up alpine as a Xen dom0 and want to use a domU as vif backend, how can I get static vif names in the backend domU? (instead of vifDOMID.DEVID) 2017-11-25 12:18:08 in dom0 I can achieve this by using vifname in the xl.cfg(5) 2017-11-25 12:20:04 should I set mac= and bind that somehow in the backend domU? 2017-11-25 12:20:53 (I'll ask in #xen as well) 2017-11-25 13:52:10 omniuwo: yeah, that will better be on #xen , no idea how the backend paths in xenstore are made... 2017-11-25 13:53:09 omniuwo: i once had a storage domU but found if i reboot it all mappings are lost forever (the paths changed and other vms lost access), that kinda left me unhappy 2017-11-25 13:53:21 make sure to test a bit 2017-11-25 14:42:20 hi, do someone have a docker image running von arm architecture ? 2017-11-25 14:50:32 or maybe can help me trying to get a image working 2017-11-25 14:55:30 î alway get the message docker: Error response from daemon: error creating aufs mount to blablaba: invalid argument 2017-11-25 15:06:45 darkfader: I will, thanks =) 2017-11-25 15:17:28 ok just changed aufs to overlay2 2017-11-25 15:25:50 Hi, does alpine remove previous versions? 2017-11-25 15:26:13 Suddenly postgresql-9.6.6-r0 is gone 2017-11-25 15:26:16 I am using alpine 3.6 2017-11-25 15:34:08 yup, i think first time i ran into that i extracted the stuff needed for pg_upgrade from the old package to some place under /tmp 2017-11-25 15:34:19 now i just dumpall since my db's are super tiny 2017-11-25 15:34:38 you can also back out of the update if time is pressing 2017-11-25 15:36:39 <_ikke_> nindustries: http://dl-4.alpinelinux.org/alpine/v3.6/main/x86_64/postgresql-9.6.6-r0.apk 2017-11-25 15:36:42 <_ikke_> It's still there 2017-11-25 15:46:25 wierd _ikke_ 2017-11-25 15:46:29 i'm using edge tough 2017-11-25 15:48:05 <_ikke_> edge has 10.1-r1, yes 2017-11-25 15:48:22 so do versions get rotated out of edge or? 2017-11-25 15:48:46 <_ikke_> yes, only the last version in the repos is there, old versions are not kept 2017-11-25 15:55:24 <_ikke_> nindustries: edge is basically a rolling release 2017-11-25 16:01:09 wait, i'm not using edge 2017-11-25 16:01:13 just main 2017-11-25 16:02:13 postgresql-9.6.5 gives me error on postgresql-9.6.6-r0: 2017-11-25 16:02:49 <_ikke_> sorry? 2017-11-25 16:03:13 My repos are http://dl-cdn.alpinelinux.org/alpine/v3.6/main + community 2017-11-25 16:03:27 And I am trying to install postgresql=9.6.5 2017-11-25 16:03:53 <_ikke_> right, that's not going to work 2017-11-25 16:04:17 <_ikke_> 3.6 has 9.6.6 2017-11-25 16:04:54 <_ikke_> tagged versions will update minor versions for security and bug fixes 2017-11-25 16:04:56 wierd, I had this version previously in 3.6 aswel 2017-11-25 16:05:08 so there is no way to pin versions? 2017-11-25 16:07:16 <_ikke_> Not that I'm aware of 2017-11-25 16:08:59 sigh 2017-11-25 16:09:04 there go my reproducible builds 2017-11-25 16:09:14 <_ikke_> nindustries: you can hae a local cache though 2017-11-25 16:10:50 hm. 2017-11-25 16:12:40 _ikke_: I am using docker, so no local cache 2017-11-25 16:12:53 <_ikke_> nindustries: you can mount volumes 2017-11-25 16:13:05 yeah, but builds can happen anywhere 2017-11-25 16:13:59 <_ikke_> You can create a mirror with the packeges you need 2017-11-25 21:24:58 whenever i invoke abuild to build a package, it re-downloads its dependencies for building 2017-11-25 21:25:11 is there a way to make abuild use some caches? 2017-11-25 21:25:35 -K 2017-11-25 21:26:14 that will keep all the temp files in the build 2017-11-25 21:26:18 so a good idea to clean up when you're totally done 2017-11-25 22:01:48 scv: build okay, install failed, can i resume from that? 2017-11-25 22:02:46 yeah, might not be a great idea to just keep that failed build for final usage but you can fix the issue and finish the install process 2017-11-25 22:02:52 i would still do a full build from scratch at the end 2017-11-25 22:03:47 i'll later put it into an buildbot anyways 2017-11-25 22:04:14 its just that the building itself takes 20 minutes 2017-11-25 22:04:25 yeah better to skip it then 2017-11-25 22:04:41 how? 2017-11-25 22:05:37 just run abuild -h and review the different sub-commands, you'll see how each step of the process is broken down 2017-11-25 22:05:54 so `abuild package` i guess 2017-11-25 22:06:27 rootpkg 2017-11-25 22:07:09 ah, i just found out the hard way 2017-11-25 22:14:52 alright, works as intended 2017-11-25 22:16:26 gonna make an own alpine repo for my custom packages 2017-11-26 01:28:48 liwakura: if you have anything that's not upstream that doesn't have shitty licensing constraints, you should push it upstream 2017-11-26 01:29:15 just try to make sure you're using good practices for your apkbuilds 2017-11-26 01:29:32 such as using static tarball releases and adding a check() section 2017-11-26 08:57:32 mepholic: its an machine-specific custom kernel 2017-11-26 08:57:38 definitely not suited for upstream 2017-11-26 09:00:20 mepholic: https://github.com/ainriver/aports 2017-11-26 09:39:24 mepholic: yeah stuff like that is why i mostly use the solo repo 2017-11-26 11:52:05 which apk can i use of example of git package? 2017-11-26 20:36:09 Hi 2017-11-26 20:36:36 What's the right way to run make menuconfig to configure a new kernel from aports? 2017-11-26 20:36:57 I tried sticking it in build() but t built it twice and bricked my server 2017-11-26 20:38:15 sy: "abuild menuconfig" I guess 2017-11-26 20:38:37 ... 2017-11-26 20:38:40 you're kidding 2017-11-26 20:39:17 no way it's that easy 2017-11-26 20:40:02 should i load the base config then do my changes to that? 2017-11-26 20:40:38 nope not that easy >>> ERROR: linux-hardened: menuconfig failed 2017-11-26 20:41:59 sy: is this why the fucking shell server is down again? 2017-11-26 20:57:14 yeah its not that easy 2017-11-26 20:57:35 sy : should i load the base config then do my changes to that? 2017-11-26 20:57:38 yeah 2017-11-26 22:28:05 lol 2017-11-26 22:28:16 sy, i thought you were somebody else, same nick 2017-11-26 22:28:26 i whois'd you and saw your domain, we're neighbors :D 2017-11-26 22:28:32 i have zl.is 2017-11-26 22:28:32 heh 2017-11-26 23:56:18 lol 2017-11-26 23:56:25 i thought sy was that person soo 2017-11-27 00:59:31 is there a circular filesystem for logs on embedded systems... 2017-11-27 01:00:40 freebsd has fifolog 2017-11-27 05:36:28 # apk update -> OK: 8439 distinct packages available ; / # apk add postgres-client -> ERROR: unsatisfiable constraints: postgres-client (missing) 2017-11-27 05:36:37 can someone explain that 2017-11-27 05:37:28 donpdonp: the package name is postgresql-client not postgres-client 2017-11-27 05:37:42 you're trying to add a nonexistant package 2017-11-27 05:37:47 got it. thx! 2017-11-27 05:38:24 i could have sworn I read it as postgres :) 2017-11-27 07:50:37 yes liwakura 2017-11-27 07:50:40 yes it is 2017-11-27 07:54:05 scv: woah 2017-11-27 07:54:19 you don't have it as your reverse though 2017-11-27 07:55:35 still kinda unsure how to do a menuconfig 2017-11-27 07:55:52 i'm wary to do it manually because i see there's an intermediate patch too 2017-11-27 08:10:07 hello guys ! Does anybody how is it the best way to add a custom exectuble to a standtard initramfs on alpine-linux ? 2017-11-27 08:13:53 btw, freeradius in defult install do not auth android device WPA2-EAP, but in Arch linux freeradius just works 2017-11-27 08:36:51 ilkappe: create a file with '.files' suffix in /etc/mkinitfs/features.d/ , add the name you used to features in /etc/mkinitfs/mkinitfs.conf 2017-11-27 08:40:21 kahiru: ty ! 2017-11-27 08:40:38 np, what do you want to have in initramfs? 2017-11-27 08:41:13 I need the fw_printenv/fw_setenv used to r/w the u-boot variables 2017-11-27 08:42:28 cool, cool 2017-11-27 08:52:37 kahiru: What about if I have to add it directly to the SD card that contains the initramfs archive and hence I can't have access to the fs ? 2017-11-27 08:52:59 maybe it's a simple questi but I am not an expert. 2017-11-27 08:55:11 guess you'd have to spin up another machine (vm, container, whatever), build it there and then retrieve the initramfs from there 2017-11-27 08:55:22 not sure if that's ideal but that's what i would do 2017-11-27 08:56:55 ok ! 2017-11-27 08:57:49 I thought there is a way to extract the files from the initramfs archive add it and repack 2017-11-27 09:03:23 well, the initramfs is a gzipped cpio archive, so you could probably do that, but I never tried it 2017-11-27 09:07:28 this could work https://paste.pound-python.org/show/FGoCC7MCO3WBPfepDlqQ/ 2017-11-27 09:13:02 kahiru: ty. I'll try it 2017-11-27 09:20:46 sy ^^ ;) 2017-11-27 09:21:22 heh 2017-11-27 09:21:26 nice 2017-11-27 09:21:52 but is yours an email and an xmpp address too 2017-11-27 09:21:56 i dont think so 2017-11-27 09:22:06 nah i use astr.al for that 2017-11-27 09:22:18 might put xmpp on uz.is tho 2017-11-27 09:22:52 lol bring your own punctuation 2017-11-27 09:22:53 lol 2017-11-27 09:23:06 uzis :s 2017-11-27 09:23:11 why didnt i think of that 2017-11-27 09:23:21 it was regged until a few days ago :D 2017-11-27 09:24:08 :s 2017-11-27 09:24:17 what else do you have? :P 2017-11-27 09:29:40 meh a bunch 2017-11-27 09:29:43 but those are the good ones 2017-11-27 09:29:49 rest are pretty mundane 2017-11-27 09:30:04 pretty hard to make words ending in is, eh? 2017-11-27 09:30:08 i gave up 2017-11-27 09:30:22 haha yeah but im a sucker for short zones 2017-11-27 09:30:45 ZoMes 2017-11-27 09:30:53 also use $ORIGIN 2017-11-27 09:31:41 was gonna recommend the rest of the free .is 2letters 2017-11-27 09:31:46 but they all got snapped up 2017-11-27 09:32:05 not all of them... 2017-11-27 09:32:08 surely 2017-11-27 09:32:14 just checked 2017-11-27 09:32:17 100% registered 2017-11-27 09:32:23 cq gy nq qb qj qs uz vq yh yj yk zj 2017-11-27 09:32:30 was all that was free when i got uzis 2017-11-27 09:32:31 the other day 2017-11-27 09:32:34 all gone now too 2017-11-27 09:35:14 i should prob put some of my domains up for auction im never going to use them all 2017-11-27 09:35:24 such a bad habit 2017-11-27 09:36:08 :ss 2017-11-27 09:36:16 i was going to buy more 2 letters 2017-11-27 09:36:33 .is had the best price :( 2017-11-27 09:39:51 yeah 2017-11-27 09:45:02 and they actually care i think 2017-11-27 09:45:20 no preregistered domains for a premium price 2017-11-27 09:45:45 nice simple interface, actually validates dns before setup, prices are reasonable 2017-11-27 09:45:57 how a registrar should be! 2017-11-27 09:46:40 back when I was interested in .is domains they weren't available for people outside of the country 2017-11-27 09:46:58 which basically ruined my "Penpals Iceland" concept (you figure the rest of it out) 2017-11-27 09:47:17 oh sy you're in luck 2017-11-27 09:47:23 theres exactly one free right now 2017-11-27 09:47:26 lol 2017-11-27 09:47:31 % No entries found for query "yh.is". 2017-11-27 09:56:52 scv: nah 2017-11-27 09:56:55 looks ugly 2017-11-27 09:57:24 most of them pretty bad 2017-11-27 10:12:08 Hi, anyone keeping their versions fixed in Docker? 2017-11-27 10:15:19 scv: its not the length, its what you do with it 2017-11-27 10:41:05 sy: hehe, so for most of them nothing :D 2017-11-27 10:41:31 scv: fun fact: the shortest domain is also a tld 2017-11-27 10:41:35 http://ai 2017-11-27 10:41:37 yep 2017-11-27 10:41:43 int also resolves 2017-11-27 10:41:47 mx records iirc 2017-11-27 10:42:00 you used to be able to AXFR a lot of the tlds too 2017-11-27 10:42:05 not so much anymore 2017-11-27 10:42:30 int has no mx record anymore :( 2017-11-27 10:43:38 :( 2017-11-27 10:44:07 its pretty hard to run a tld 2017-11-27 10:45:08 i imagine its the administrative overhead 2017-11-27 10:45:37 naw mostly fixing web things when they break 2017-11-27 10:45:54 and removing spam domains 2017-11-27 10:46:08 running a pile of resolvers can't be that difficult even with security requirements in place 2017-11-27 10:46:20 yeah i'd consider "removing spam domains" administrative overhead for it 2017-11-27 10:46:28 like day-to-day admin tasks i meant 2017-11-27 10:46:46 well its less so for openni TLDs though :) 2017-11-27 10:46:49 opennic* 2017-11-28 00:04:29 Anyone else getting this? PAX: RAP hash violation for function pointer: expired_send_keepalive+0x0/0xe0 [wireguard]. Edge kernel and wireguard. 2017-11-28 00:10:03 sol56: bug in the wireguard driver(?) 2017-11-28 00:10:09 didn't think wireguard used a kernel module of it s own 2017-11-28 00:15:47 Maybe it actually works fine until I load my config into it and it tries sending keepalives, I assume based on the message, even though I don't have that set in my config. Is there a preferred pastebin for a more full log? 2017-11-28 00:26:16 any bin is fine 2017-11-28 00:50:06 https://ptpb.pw/_d41 2017-11-28 00:50:17 Doesn't actually seem all that useful to me. 2017-11-28 00:52:54 I think the first time I noticed it happen was 11/1 with 4.9.59 and wg 20171101 2017-11-28 00:53:38 its a bug with wg 2017-11-28 00:54:18 Known bug? 2017-11-28 00:54:35 doesn't appear to be 2017-11-28 00:54:43 it wouldn't be an issue normally 2017-11-28 00:54:46 they're calling a function pointer with an incorrect type 2017-11-28 00:54:55 and pax is catching that/crashing on it 2017-11-28 01:00:59 huh 2017-11-28 01:02:11 I linked it to #wireguard. 2017-11-28 01:26:30 sol56: little more info - the function is being triggered from a timer, so it'd be an issue where they setup the timer from the looks of it 2017-11-28 07:08:35 An aarch64 version of Alpine seems to exist, and I wonder if it works on rpi3? 2017-11-28 08:02:04 Looks like there is a fix here https://git.zx2c4.com/WireGuard/commit/?id=e4bf02b833f99f4dcc2ab685d92517ccf8cc4766. Anyway I could ask a newer version of wireguard either 1122 or 1127 be pushed to the edge repos? 2017-11-28 08:37:42 sol56: xz2c4 already does that automatically https://github.com/alpinelinux/aports/pull/2833 2017-11-28 08:40:50 Gotcha 2017-11-28 09:22:16 When reading https://www.alpinelinux.org/posts/Alpine-3.5.0-released.html it says "Support for ZFS as root". 2017-11-28 09:22:36 I tried to do this by manually creating the datasets and mounting under /mnt 2017-11-28 09:22:57 "setup-disk /mnt" did not accept ZFS 2017-11-28 09:23:43 What is the official way to install ZFS root? 2017-11-28 10:00:19 hello guys ! I'm trying to add a custom binary executable file to the alpine-linux image for rpi. So far I've extracted the initramfs archive, added my executable to the sbin folder and repacked the whole fs. By the way, after booting the image, I can't see my executable under the current /sbin folder 2017-11-28 10:00:26 Any siggestions? 2017-11-28 10:00:47 well 2017-11-28 10:00:53 the file's in the initramfs 2017-11-28 10:00:54 thats why 2017-11-28 10:02:33 uh ? 2017-11-28 10:03:03 scv: Can you be more explicit ? 2017-11-28 10:03:41 the initramfs is just for the startup process 2017-11-28 10:03:45 it isn't the runtime filesystem 2017-11-28 10:05:56 scv: ty. I'm a novice 2017-11-28 10:06:30 How can I add a new executables on a non running image ? 2017-11-28 10:06:39 you'd need to edit the actual image 2017-11-28 10:06:53 no idea where that's stored for the rpi variant 2017-11-28 10:09:46 scv: so I can't used the .tar.gz ready to be downloaded.. correct ? 2017-11-28 10:10:19 no idea, i have not used the rpi distro. i can look at the structure 2017-11-28 10:10:20 sec 2017-11-28 10:15:53 ty 2017-11-28 10:17:11 i'm fairly sure that's a "data" install 2017-11-28 10:18:16 or like, diskless mode 2017-11-28 10:18:18 not sure 2017-11-28 10:18:30 pretty sure the install is a read only volume 2017-11-28 10:19:57 basically there is the dtb, initramfs vmlinuz and the ovl pkg 2017-11-28 10:20:37 I hoped that including directly my binary into the shipped initramfs would have been sufficient 2017-11-28 10:22:11 ilkappe: after the initramfs, it pivotmounts the / directory to the actual install image 2017-11-28 10:23:03 i'm downloading the tgz right now 2017-11-28 10:23:06 i'll take a look 2017-11-28 10:24:00 mepholic: ty for help 2017-11-28 10:28:05 ilkappe: it seems likely that the initramfs is just installing the apks in the apk directory into the rootfs 2017-11-28 10:28:10 which i'm guessing is in ram? 2017-11-28 10:28:21 yes 2017-11-28 10:28:46 you are guessing right 2017-11-28 10:28:48 so build your binary into a package and dump it in the apks directory 2017-11-28 10:29:50 you'll probably need to add your apk pubkey to the initramfs 2017-11-28 10:30:16 mepholic: ok ! So each apk package has a related pubkey ? 2017-11-28 10:30:53 maybe it's a simple question but I really don't know how alpine-linux works. I'm looking at it basically right now for the first time 2017-11-28 10:32:52 ilkappe: nah 2017-11-28 10:33:14 most of them are just signed by the two alpine keys 2017-11-28 10:33:16 alpine-devel@lists.alpinelinux.org-524d27bb.rsa.pub alpine-devel@lists.alpinelinux.org-58199dcc.rsa.pub 2017-11-28 10:34:01 just add your pubkey to /etc/apk/keys in the initramfs 2017-11-28 10:34:14 by the way it seems to be a little more complicated than I thought.. 2017-11-28 10:35:04 # copy keys so apk finds them. apk looks for stuff relative --root 2017-11-28 10:35:05 mkdir -p $sysroot/etc/apk/keys/ 2017-11-28 10:35:07 cp -a /etc/apk/keys $sysroot/etc/apk 2017-11-28 10:35:12 that happens in the initramfs 2017-11-28 10:35:34 ilkappe: this is just if you want to build an image 2017-11-28 10:35:51 you can probably come up with a way of doing a standard alpine install 2017-11-28 10:36:01 onto an sdcard or something 2017-11-28 10:36:20 I need only to add a binary to the original image ! 2017-11-28 10:37:13 mepholic: by the way thank you. I have an idea on how to do that now 2017-11-28 10:37:28 no problem 2017-11-28 10:43:08 ilkappe: if you just want a traditional persistant installation 2017-11-28 10:43:10 try this 2017-11-28 10:43:12 https://wiki.alpinelinux.org/wiki/Raspberry_Pi#Traditional_disk-based_.28sys.29_installation 2017-11-28 11:20:34 mepholic: ty. I'll look at that asap. I'm very busy now 2017-11-28 11:28:26 is alpine linux suitable for daily usage of desktop user? 2017-11-28 11:28:42 with some work yes 2017-11-28 11:28:45 mepholic: ? 2017-11-28 11:29:11 yes 2017-11-28 11:29:18 you'll probably have to build some packages 2017-11-28 11:29:23 but that's easy on alpine 2017-11-28 11:29:29 i'm currently using arch and have a plan moving to alpine 2017-11-28 11:29:39 i run alpine on my desktop and laptop 2017-11-28 11:29:57 and osx at work, because the only linux we're allowed to run is ubuntu 2017-11-28 11:31:02 why there are 2 version of firefox in alpine repo, one is version 38 and the other is 57? 2017-11-28 11:32:29 one is the ESR version 2017-11-28 11:35:22 nvhhnvn: what window manager/desktop environment are you planning on using/ 2017-11-28 11:35:43 i'm currently using i3-gaps 2017-11-28 11:35:57 gotcha 2017-11-28 11:38:18 and i3wm-gaps is in edge brand :) 2017-11-28 11:46:13 there are many unix malwares in wild (mac malware). i'm thinking of switching to more secure distro. alpine seem much more interesting than arch or gentoo. i have just discovered alpine since last week. 2017-11-28 11:46:50 why alpine is so silent on the media? 2017-11-28 11:48:30 Alpine is younger 2017-11-28 11:52:47 i use zathura-djvu to open .djvu on arch. is there anything similar to that on alpine? 2017-11-28 11:55:58 <_ikke_> nvhhnvn: something can be packaged probably 2017-11-28 12:09:37 mkinitfs seems to include the kernel module, but it seems to not make the initramfs to automatically load it. How do I get it to load? 2017-11-28 12:10:38 Now it will fail, and I get dropped to a prompt where I can load the module, mount root as /sysroot and then continue to boot. 2017-11-28 12:11:45 In my case it is the zfs module 2017-11-28 14:06:48 ACTION created a bug report https://bugs.alpinelinux.org/issues/8211 2017-11-28 14:12:49 ok (1/1) Upgrading firefox (55.0.2-r0 -> 57.0-r1) and nothing is rendered anymore in firefox 2017-11-28 14:25:34 how can i make a git based apk? 2017-11-28 14:29:03 git based apk? 2017-11-28 14:30:37 an apk which does not install a particular pogram version, but uses git to download sources 2017-11-28 14:31:16 i trying to make a package for uzbl 2017-11-28 14:31:32 it seems does not have releases just a github repository 2017-11-28 14:31:54 not 2017-11-28 14:32:18 we typically use a date-based version tied to a git hash for those programs 2017-11-28 14:32:22 e.g. 20171128 as version 2017-11-28 14:36:10 can you give me some example 2017-11-28 14:36:19 so i can watch the code of the apkbuild 2017-11-28 14:44:09 im also trying to ake a package for gftp i put in makedepends="glib-dev" but when abuild -r build it does not install glib and configure checks faisl 2017-11-28 14:44:22 configure: error: gFTP needs GLIB 1.2.3 or higher 2017-11-28 14:44:32 dont know why is not intalling missing dependencies 2017-11-28 15:20:39 maybe the check is wrong 2017-11-28 15:21:16 is it possible to configure a network interface by its mac address? i.e. match an iface block to an iface with a specific mac address 2017-11-28 15:23:44 oh, https://wiki.alpinelinux.org/wiki/User:Mattx86/qwn:Renaming_a_network_interface 2017-11-28 15:37:22 what check? i put glib-dev in makedepends, then i make abuild -r build, shouldnt it install glib-dev preivious trying to build my package? 2017-11-28 15:38:57 well as i see abuild has commands deps and undeps, im gonna use this 2017-11-28 15:39:07 but abuild -r build fails.. 2017-11-28 15:42:15 make[1]: Entering directory '/home/h/devel/aports/testing/gftp/src/gftp-2.0.19' 2017-11-28 15:42:17 Making all in docs intl lib po src 2017-11-28 15:42:19 zsh:cd:18: no such file or directory: docs intl lib po src 2017-11-28 15:42:21 does not work 2017-11-28 15:42:54 all this dirs exists so mayb is taking all the string as a single name? 2017-11-28 15:43:00 dont know how to solve this 2017-11-28 15:43:06 think needs a patch 2017-11-28 15:44:54 that's what the error message indicates, trying to cd to the whole string 2017-11-28 15:45:21 so either the cd itself has wrong quoting or whatever feeds it the directory names is wrongly quoted 2017-11-28 15:46:43 and how to solve this? patching makefiles ...? 2017-11-28 15:47:07 idont know a lot of autotool 2017-11-28 15:47:09 s 2017-11-28 15:47:57 but this time i think is gonna be easy Makefile.am ... SUBDIRS= docs intl lib po src 2017-11-28 15:48:08 i suppose is a matter of wuote thi 2017-11-28 15:48:10 s 2017-11-28 15:48:12 quote* 2017-11-28 15:48:16 and then there's the zsh issue 2017-11-28 15:48:58 SoHTaGa: maybe you could post some code? 2017-11-28 15:49:13 afaik the aports dont have gftp yet 2017-11-28 15:49:40 my 5 minute ago pulled local doesn't, at least 2017-11-28 15:49:44 which code? im just trying to make a package for gftp is simply a ./configure && make 2017-11-28 15:50:47 my instinct is often wrong, but this time it suggests this is zsh related 2017-11-28 15:51:15 TBB: you think the guilty is zsh? 2017-11-28 15:51:55 ok trying with sh 2017-11-28 15:52:24 it gives same error 2017-11-28 15:52:35 maybe an quoting issue 2017-11-28 15:52:39 is abuild using zsh instead of the sh shell i have just opened? 2017-11-28 15:52:44 SoHTaGa: you habe an issue with your code. 2017-11-28 15:52:48 We dont know your code. 2017-11-28 15:52:58 s/habe/have/ 2017-11-28 15:53:28 heck, i dont even know if your APKBUILD is broken or if gftp is just borked 2017-11-28 15:53:31 well this time in Makefile.am i have sustituted the previuos string by a "new one" 2017-11-28 15:53:43 maybe i sholud left as original and compile with sh 2017-11-28 15:53:47 but still zsh error 2017-11-28 15:53:50 dont know why 2017-11-28 15:53:57 is abuild using zsh? 2017-11-28 15:54:12 ok gonna post the apkbuild 2017-11-28 15:55:45 https://gist.github.com/aGaTHoS/fac7d07260f80287e9098bd14083c41e 2017-11-28 15:57:44 alright 2017-11-28 15:57:58 in an hour i'll be at the hackspace, i'll put the file into my aports then and look into it 2017-11-28 17:49:00 SoHTaGa: are you building in a nice clean environment? typically it's best to build on absolutely thinnest alpine possible, like an LXC container with only alpine-sdk installed 2017-11-28 17:49:44 im building in my pc 2017-11-28 17:50:00 with thi things i use that are not much is still lightweight 2017-11-28 17:50:25 even so, not even your preferred shell should be installed 2017-11-28 17:52:05 here are the pckages installed on my build lxc alpine-base 2017-11-28 17:52:07 alpine-sdk 2017-11-28 17:52:09 dev86 2017-11-28 17:52:11 diffutils 2017-11-28 17:52:13 dropbear 2017-11-28 17:52:15 dropbear-dbclient 2017-11-28 17:52:17 dropbear-scp 2017-11-28 17:52:19 libexecinfo-dev 2017-11-28 17:52:21 msmtp 2017-11-28 17:52:23 sed tar 2017-11-28 17:53:35 dependencies will be pulled in (as you may have noticed) 2017-11-28 17:55:48 SoHTaGa: sorry, i got quite bus 2017-11-28 17:55:50 *busy 2017-11-28 17:55:55 do you still have your issues? 2017-11-28 17:58:37 but no yes 2017-11-28 17:58:41 yes* 2017-11-28 18:00:02 and waht is supposed for me to do ScrumpyJack ? use another pc? a virtual machine? or a jail 2017-11-28 18:03:16 if you are on linux, I would use LXC 2017-11-28 18:03:44 a build of gftp with your file fails with http://tpaste.us/YnKr 2017-11-28 18:04:03 ScrumpyJack: isn't abuild using an fakeroot on its own already? 2017-11-28 18:10:21 then if its gonna fail anyway.. 2017-11-28 18:10:57 dont know how to solve this 2017-11-28 18:17:31 fix the code? 2017-11-28 18:17:45 its a typical glibc assumption 2017-11-28 18:21:45 maybe this is related: https://github.com/voidlinux/void-packages/blob/master/srcpkgs/gftp/patches/musl.patch 2017-11-28 18:23:18 i just tested roughly the asme thing 2017-11-28 18:26:00 https://git.brinstar.org/scv/alpine-packages/src/master/gftp/APKBUILD 2017-11-28 18:26:09 SoHTaGa: can see the missing subpkgs you'll need etc ^ 2017-11-28 18:26:23 cappy 2017-11-28 18:26:40 ? 2017-11-28 18:26:56 it builds fine if you just patch that one missing include 2017-11-28 18:27:20 it is a glibc-specific header but musl provides a compatibility shim for it 2017-11-28 18:28:35 ok 2017-11-28 18:29:48 are you gonna merge this apkbuild 2017-11-28 18:29:50 ? 2017-11-28 23:03:07 SoHTaGa: im not an alpine dev thats just a personal repo 2017-11-29 00:50:23 scv: aren't we going to merge some of those packages at some point though? 2017-11-29 00:50:37 the ones that are relevant yeah 2017-11-29 00:50:40 we should probably do it sooner rather than later 2017-11-29 00:50:53 i personally would like to find some path to get those galera packages in 2017-11-29 00:51:08 it did require a fair amount of work to get them comfy on base alpine but they work like a charm 2017-11-29 00:51:18 i maintain packages at https://github.com/Xe/aports as well, we should work with the apkfission guys on putting these somewhere tbh 2017-11-29 00:51:34 https://git.brinstar.org/scv/alpine-packages is our repo 2017-11-29 00:51:47 a lot of that is personal cruft though 2017-11-29 00:51:51 i should split that stuff out 2017-11-29 00:52:21 most of those perl packages are deps for stuff that i wouldn't be submitting for example 2017-11-29 00:52:52 yay gogs 2017-11-29 00:52:57 <3 gogs 2017-11-29 00:53:11 so damn easy to use 2017-11-29 00:53:22 it's also lighter weight than it has any reason to be 2017-11-29 00:53:40 light weight is good 2017-11-29 00:56:14 until i broke it 2017-11-29 00:56:18 by pushing a commit 2017-11-29 00:56:24 force* 2017-11-29 00:57:07 tbh it was out of date 2017-11-29 00:57:12 and was fixed already 2017-11-29 00:57:18 kek 2017-11-29 00:57:31 im just a lazy bastard^W^Wtoo busy 2017-11-29 00:57:32 :^) 2017-11-29 00:57:56 apathy is a perfectly reasonable justification for things breaking 2017-11-29 00:57:58 :P 2017-11-29 05:38:15 ncopa: You seems to be the biggest contributor to mkinitfs. Did you see the bug report I made yesterday? https://bugs.alpinelinux.org/issues/8211 2017-11-29 05:38:56 ncopa: I could test and help if you work on this bug, as I have a ready VM to test on. 2017-11-29 05:41:00 ncopa: Basically, what I need to do at the emergency prompt to get it to continue the boot is to 'modprobe zfs', 'zpool import -R /sysroot -N rpool', 'zfs mount rpool/ROOT/alpine' 2017-11-29 06:05:03 I have a problem - I need to run uwsgi in alpine linux (in docker, incidentially), but on port 80. 2017-11-29 06:05:15 Yet, whenever I try, I get this error: bind(): Permission denied [core/socket.c line 769] 2017-11-29 06:05:21 How do I resolve this? 2017-11-29 06:22:21 Arahael: it's failing to bind to port 80? 2017-11-29 06:22:33 is it running (or at least starting) as root? 2017-11-29 06:22:55 or has CAP_NET_BIND_SERVICE ? 2017-11-29 06:26:22 scv: I don't really know. I believe that error means that it's failing to bind to port 80. It's starting as root, but is told to drop privs to a specific uid and gid. 2017-11-29 06:26:35 As for CAP_NET_BIND_SERVICE, I don't know what that means? 2017-11-29 06:26:44 i'm not ultra familiar with docker so there might be some missing piece here i'm unaware of 2017-11-29 06:26:51 but you are correct on starting as root 2017-11-29 06:27:14 CAP_NET_BIND_SERVICE is a capability flag that can be set on a binary to allow it to bind to ports under 1024 without being root 2017-11-29 06:27:28 the command for it is on the caddy readme 2017-11-29 06:27:46 scv: As a hacky, hacky workaround, I've gotten docker to do the port binding and then I connect to the host's IP address using netstat -nr | grep '^0\.0\.0\.0' | awk '{print $2}' to find the ip. 2017-11-29 06:27:54 Xe: What's "caddy"? 2017-11-29 06:28:00 sudo setcap cap_net_bind_service=+ep /path/to/binary 2017-11-29 06:28:06 caddy is a webserver written in Go 2017-11-29 06:28:11 Ah, I'm not using caddy. 2017-11-29 06:28:25 yeah 2017-11-29 06:28:31 but the setcap command is in caddy's readme 2017-11-29 06:28:32 I did try that, actually - made no difference. 2017-11-29 06:28:35 Arahael: i was under the impression that docker required you to setup NAT/PAT always? 2017-11-29 06:28:40 scv: not always 2017-11-29 06:28:53 i know kubernetes uses a bridged model 2017-11-29 06:29:02 but i thought standard docker was routed/NAT 2017-11-29 06:29:05 guess not 2017-11-29 06:29:10 scv: I'm dealing with a fairly stupid webserver that needs to know which "production domain" it runs as, *and* it queries that same domain for crap. 2017-11-29 06:29:19 ow 2017-11-29 06:29:21 that sounds lovely 2017-11-29 06:29:25 if you are doing things other ways you can omit port forwarding entirely 2017-11-29 06:29:37 Xe: I'm port-forwarding as a workaround. :/ 2017-11-29 06:29:58 scv: It's indeed lovely. :( 2017-11-29 06:30:22 Xe: yeah i would consider a traditional bridged interface setup to be not only easier to configure but less problematic overall 2017-11-29 06:30:31 not sure why the port translation thing is so popular 2017-11-29 06:30:44 scv: Docker sets up a virtual network. 2017-11-29 06:31:10 scv: tbh i've been working on a solution that omits the need to do any kind of port forwarding, instead of making the load balancer talk to the backend, i'm making the backend talk to the load balancer and reusing an existing connection to relay http streams 2017-11-29 06:31:28 Xe: Do you document that somewhere? 2017-11-29 06:31:29 Xe: a la fd_pass sort of thing? 2017-11-29 06:31:46 how does that work with ssl? 2017-11-29 06:31:56 gotta encapsulate at the lb i assume? 2017-11-29 06:32:04 scv: the load balancer terminates SSL 2017-11-29 06:32:10 yeah figured as much 2017-11-29 06:32:27 also the agent literally makes TCP connections to the backend from a poo b/c i haven't found a more efficient design 2017-11-29 06:32:29 pool* 2017-11-29 06:32:45 i mean, seems pretty sane to me 2017-11-29 06:33:02 Arahael: i have code for it, but my implementation so far is incredibly bad 2017-11-29 06:33:07 assuming you can scale multiple backends at least 2017-11-29 06:33:14 yeah 2017-11-29 06:33:21 in my testing i can scale to `n` backends 2017-11-29 06:33:31 where n is large enough (300) that it doesn't matter 2017-11-29 06:34:05 jeez, i can't imagine what kind of application is gonna be so heavy it'll need 300 backends on a single cluster :< 2017-11-29 06:34:16 Xe: Heh, that's OK, lets leave it then. :) My stuff is always incredibly bad, too. :( 2017-11-29 06:34:21 that's why 300 is the number i use for "hysterical load balancing" 2017-11-29 06:34:28 I've not done "scaling" yet... Thats' on my todo. 2017-11-29 06:34:40 its always a fun time :D 2017-11-29 06:34:47 Actually kind of hoping I don't need to scale, but if this was real production, I should at least scale to 2. 2017-11-29 06:34:55 Arahael: https://git.xeserv.us/xena/route, undocumented at best, you are mostly on your own using it. stability is "who knows man" 2017-11-29 06:35:12 it does do let's encrypt for you tho 2017-11-29 06:35:28 and has an API-driven interface using grpc 2017-11-29 06:35:49 Xe: Ah, I see. That could be really intriguing if documented, but it isn't useful for me, sadly. 2017-11-29 06:35:51 i plan to make it support arbitrary TCP at some point 2017-11-29 06:35:59 Xe: I'm 100% amazon, here. :/ 2017-11-29 06:36:03 Arahael: https://backplane.io is probably a more immediately usable version of this 2017-11-29 06:36:16 i have an alpine package to make an in-container agent for it 2017-11-29 06:36:36 Xe: how do you like grpc? i'm seriously considering it for replacing an in-house solution 2017-11-29 06:36:36 (i really need to clean that up) 2017-11-29 06:36:54 scv: it's able to codegen C++, python, java, Go and Swift 2017-11-29 06:36:55 afraid it might not meet all my current capabilities though 2017-11-29 06:36:59 that's more than you can do 2017-11-29 06:36:59 scv: I tend to prefer message passing, rather than rpc. 2017-11-29 06:37:37 scv: i like gRPC better than other options because gRPC is at least statically typed 2017-11-29 06:37:51 and you're not locked into an ecosystem using it 2017-11-29 06:38:30 Arahael: yeah - my current solution is truly just a message passing system, but it is mostly used for rpc-like calls 2017-11-29 06:38:35 but not strictly limited to it 2017-11-29 06:39:01 scv: I might be tempted to just use http requests, but that's my golang influence showing. 2017-11-29 06:39:17 scv: If you're using rpc-like all the time, then perhaps codifying it as actual rpc is a good idea. 2017-11-29 06:39:26 scv: What language? 2017-11-29 06:39:41 Arahael: with HTTP at the least you have a guarantee to get at least one response from a request 2017-11-29 06:39:53 Xe: Indeed. 2017-11-29 06:40:08 the main thing that's been a holdup for me finding a good solution is that we currently have a sort of broadcast message type that can be filtered to certain worker classes, but it seems most alternatives don't have similar functionality at the protocol level 2017-11-29 06:40:30 Arahael: it's just perl, although i have a prototype version i did in go some time ago 2017-11-29 06:40:36 scv: tbh use grpc and a message queue 2017-11-29 06:40:48 two different things need two different approaches 2017-11-29 06:41:18 for extra credit, embed a STOMP broker or spin up some irc server 2017-11-29 06:41:44 IRC is basically a message queue with a funny framing format 2017-11-29 06:42:32 they're two different things at that level but at a more abstract view both the rpc and broadcast query types are used together in most cases, thats why i wanted to keep them on the same transport 2017-11-29 06:43:22 i never really thought about using IRC since it doesn't have any message persistency, a lot of writing i've read about these types of systems people seem to use XMPP which i found amusing, yet somewhat suitable 2017-11-29 06:44:04 IRC is a poor broker. 2017-11-29 06:44:48 it is 2017-11-29 06:45:03 but it makes for interesting reactions when people find out you are really using IRC as a message broker 2017-11-29 06:45:43 the other caveat is that worker-type clients can issue standalone broadcast-type messages without an associated request id, api-type clients can bind to these message types and often use them as triggers to start a new task 2017-11-29 06:46:04 Xe: It's easy to understand, at least. 2017-11-29 06:46:16 Xe: Though, relatively few people would know about IRC's broadcast. 2017-11-29 06:46:54 Arahael: it's good for one -> many delivery with at least once end client delivery 2017-11-29 06:47:07 or one -> one communication 2017-11-29 06:47:16 Or one -> all 2017-11-29 06:47:16 other than that, it's like trying to get blood from a stone 2017-11-29 06:47:25 Arahael: one -> many == one -> all 2017-11-29 06:47:38 Sure. 2017-11-29 06:47:40 example use case - we have a worker-type client on a router or switch, and if it detects a high packet rate flow it'll issue a broadcast. an rpc-type worker can pick up one of these broadcasts and search for the device that is receiving the abusive flow, retrieve information about the target and then issue another request to take action on it 2017-11-29 06:47:51 But one thing it misses, is that there's no confirmation or queuing, whatsoever. 2017-11-29 06:47:55 whether it be blackholing, or disabling the affected service 2017-11-29 06:47:56 etc 2017-11-29 06:48:00 scv: i'd look at something like nats for that? 2017-11-29 06:48:10 nats? 2017-11-29 06:48:24 http://nats.io/ 2017-11-29 06:48:37 it has a kafka-like streaming server mode that may be interesting to you 2017-11-29 06:48:53 hrm this looks promising 2017-11-29 06:48:57 (or you could just try kafka, but it deps both jvm and zookeeper) 2017-11-29 06:49:22 i'd rather avoid running jvms everywhere heh 2017-11-29 06:49:38 it's worth trying kafka at least once tho 2017-11-29 06:49:46 kafka is a jvm meme but it's a hell of a drug 2017-11-29 06:49:56 this is extremely promising Xe 2017-11-29 06:49:59 thank you for the link 2017-11-29 06:50:03 np 2017-11-29 06:50:06 i'm going to try this out some time son 2017-11-29 06:50:07 soon 2017-11-29 06:50:34 all the things i've done with nats have been beautifully simple, an earlier version of my HTTP proxy used nats even 2017-11-29 06:51:12 i did try porting parts of this stack to using various *MQ systems, and they all proved to either be slow and hard to scale or excessively complex compared to the interface we have now 2017-11-29 06:51:18 scv: idea: combine worker groups and wildcards? 2017-11-29 06:51:41 kinda one of the nice parts of in-housing it, i was able to abstract tons of stuff away to fit our model more tightly 2017-11-29 06:51:46 Xe: we do to some extent 2017-11-29 06:51:56 no i mean in nats 2017-11-29 06:51:58 oh 2017-11-29 06:52:26 that'd be essentially what we need for the specific routing parts 2017-11-29 06:52:45 can't find the nats.io docs on it but: https://github.com/nats-io/go-nats#wildcard-subscriptions 2017-11-29 06:53:04 yeah that looks essentially like my current interface lol 2017-11-29 06:53:08 xD 2017-11-29 06:53:53 when i started writing the mp/rpc codebase there were very few alternatives, so it was pretty suitable back then 2017-11-29 06:54:48 it sat mothballed for a few years with the code frozen in prod, now i'm reimplementing these services/refactoring just about everything and i think it's time to ditch the in-house solution 2017-11-29 06:55:05 message passing got popular again which makes me happy :D 2017-11-29 06:55:20 honestly, you got far enough with your self-made solution that you might wanna at least write a blogpost on it 2017-11-29 06:55:34 at the least it could make for an interesting read how you came to that decision 2017-11-29 06:55:42 eh i have about two dozen half written blogposts i need to finish as it is :D 2017-11-29 06:55:46 maybe sooner or later 2017-11-29 06:56:03 i started working on this codebase in like 2008 or so 2017-11-29 06:56:08 it's rather long in the tooth now 2017-11-29 06:56:13 why else would i suggest writing something else you can forget to finish writing 2017-11-29 06:56:15 :D 2017-11-29 06:56:16 :^) 2017-11-29 06:57:11 the last major adjustment i made was to port the network layer to use zmq which was a nice improvement 2017-11-29 06:57:25 reimplemented the broker with much less code, and it helped reduce latency across the board 2017-11-29 06:58:56 I hear mixed dthings about zmq. The go guys seem to think you can just use tcp directly? 2017-11-29 06:59:02 zmq's interface made it pretty easy to bolt it onto the existing socket layer 2017-11-29 06:59:03 (I haven't used zmq myself, but it's on my todo) 2017-11-29 06:59:15 Arahael: it really depends on what you do tbh 2017-11-29 06:59:17 Arahael: mmmm, it really 2017-11-29 06:59:19 yah. lol 2017-11-29 06:59:29 Xe: "Message passing". :) 2017-11-29 06:59:50 it's quite useful if you need MP primitives at the socket level 2017-11-29 06:59:50 Arahael: io.copy passes every message! 2017-11-29 07:00:00 Xe: Indeed! 2017-11-29 07:00:41 man Xe you got me excited to try out nats, if it truly is as good as it looks it'll save me so much time 2017-11-29 07:00:56 scv: worst case you still have a production-ready impl 2017-11-29 07:01:11 i don't really want to be working on a whole MP/RPC suite, i have more interesting code to be writing 2017-11-29 07:01:20 yeah, it's in dire need of refactoring though 2017-11-29 07:01:28 but it works 2017-11-29 07:01:29 my personal code standards have evolved quite a lot since 2012 2017-11-29 07:01:34 that is more than many people can say 2017-11-29 07:02:00 mhm, and it's prod-tested so at least i know what i need to replace it with 2017-11-29 07:02:16 successfully powered three different LOB applications so far 2017-11-29 07:03:52 i'm really digging these example snippets on the go bindings. it's extremely close to the interface i'm familiar with 2017-11-29 07:04:21 it even has proper authentication and autodiscovery layers which were missing from a lot of the other alternatives i've found 2017-11-29 07:18:43 it lets you cluster nats servers too kinda like ircd iirc 2017-11-29 07:22:10 yeah i see the clustering in the docs, that was one big issue i ran into, the way i implemented message reliability made my broker clustering very difficult to implement properly 2017-11-29 07:24:19 yeah 2017-11-29 07:24:23 i'd say TIAS tbh 2017-11-29 07:24:29 the annoying part is then you need a config file lol 2017-11-29 10:15:33 Hi guys, how can i do to change the language in Xfce? 2017-11-29 10:23:41 elc79: apk add lang 2017-11-29 10:23:44 to add language files 2017-11-29 10:26:43 and then? 2017-11-29 10:27:33 then i dont know more :) 2017-11-29 10:27:39 maybe set env vars 2017-11-29 10:27:51 maybe pick language in your login manager 2017-11-29 10:28:19 maybe you can find language selector in some configs dialog 2017-11-29 10:29:37 maybe ask in #xfce 2017-11-29 11:50:19 ncopa: I changed the kernel line similar to what you have, according to the bug thread and it boots and works, but prints an extra error: "cannot import 'rpool': pool already exists" 2017-11-29 11:50:46 hm 2017-11-29 11:50:58 Everything is mounted right though 2017-11-29 11:51:13 and it boots? 2017-11-29 11:51:16 Yes 2017-11-29 11:51:28 ok, i have seen some similar message 2017-11-29 11:51:34 but never figured out why it happens 2017-11-29 11:53:45 ncopa: So now I have a fully ZFS system, without any separate boot. 2017-11-29 11:53:58 thats awesome 2017-11-29 11:54:05 i suppose i could do the same on my lappy 2017-11-29 11:55:08 Yeah. But grub is needed for that, as grub is the only bootloader able to read the kernel from a ZFS filesystem. 2017-11-29 11:56:30 It would be good to add zfs to supported filesystems in setup-disk... Now I edited it by hand before installing to the ready mounted zfs filesystem 2017-11-29 12:03:04 ncopa: Would you add zfs to setup-disk if I would add a page to the wiki about installing a zfs system? 2017-11-29 12:03:41 hm 2017-11-29 12:03:47 not sure i have time for that for 3.7 release :-/ 2017-11-29 12:03:54 also 2017-11-29 12:04:11 we dont ship the zfs modules with the release boot media 2017-11-29 12:05:54 I know. I had to apk them. The line in setup-disk I changed was 198 for edge (local supported=".....) 2017-11-29 12:08:08 ncopa: Out of curiousity, when will you update to some more reasent kernels? I am mostly asking because I am interested in aarch64 and rpi3. Recent kernels should support it out of box. 2017-11-29 12:09:09 ah 2017-11-29 12:09:25 i think our lixnu-rpi kernel should support rpi3 2017-11-29 12:09:35 yes, we will upgrade the kernel after v3.7 release 2017-11-29 12:09:37 to 4.14 2017-11-29 12:10:15 Unless I am mistaken, you do not have 64-bit rpi3 images... only 32-bit 2017-11-29 12:10:43 4.14 sounds good :) 2017-11-29 12:14:27 I like different dialects of lisp, and SBCL only gives native threads on arm in 64 bit mode. 2017-11-29 12:14:45 32 bit means I am stuck using a single core. 2017-11-29 13:01:46 Is there a way to get something like /dev/disk/by-id/ ? 2017-11-29 13:03:28 Nevermind, I found I can install udev... 2017-11-29 13:29:20 engblom: doesn't UUID=something good enough for mounting stuff? 2017-11-29 13:29:44 s/doesn/isn/ 2017-11-29 13:33:07 Not with ZFS 2017-11-29 13:39:15 By the way, grub is broken. It does not understand /dev/disk/by-up/whatever 2017-11-29 13:39:49 Probably other distros like Ubuntu & Co got a patched grub... 2017-11-29 13:40:59 I am able to boot directly from ZFS (without boot) if I use /dev/sd* and not /dev/disk/by-id/* when creating the zfs array 2017-11-29 13:41:49 s/by-up/by-id/ 2017-11-29 13:59:56 my guess is, it's not a grub problem 2017-11-29 14:00:30 it's more an udev/mdev problem, as some of those paths are created by udev and contain symlinks to the /dev/sd* device nodes 2017-11-29 14:01:02 ... and Alpine defaults to mdev 2017-11-29 15:25:53 When I do /etc/init.d/docker stop, openrc does not give enough time to docker to stop all containers before it kills the service and leaves docker in an unclean state. Can I increase this grace-period-before-sigkill somehow? 2017-11-29 15:39:24 The good news of the day: https://github.com/nbs-system/snuffleupagus/blob/master/APKBUILD 2017-11-29 15:39:51 it'll land as a MR as soon as we do make a release :) 2017-11-29 15:53:55 snuffleupagus is an interesting name 2017-11-29 15:54:10 since i'm on a train right now, may i ask what it does? 2017-11-29 15:55:09 kind of suhosin for php7 2017-11-29 15:55:22 cool! 2017-11-29 15:55:27 but with php-level virtual patching 2017-11-29 15:55:38 and as simple to spell too :) 2017-11-29 15:55:40 and a grsec-like approach, aka "kill bug classes in a generic way" 2017-11-29 15:55:44 haha 2017-11-29 15:56:13 then very much thank you for making the apkbuild and anything else you did for it 2017-11-29 15:56:23 and for the explanation 2017-11-29 15:57:14 we do like Alpine at work, so we thought that we should "contribute" back :) 2017-11-29 16:11:31 that's our approach as well; our product uses Alpine as the base platform, so IMO we should contribute back. Unfortunately, the kind of project I'm working on has quite strict policies regarding anything like that, so I need to get a permission for each and every contribution, and it can be close to impossible at times :/ 2017-11-29 16:13:09 I said that we liked it, not that we're using it in production :D 2017-11-29 17:42:32 TBB: what product is this :P 2017-11-29 17:44:59 can't really be more specific about it than to say that it basically runs a single application 2017-11-29 17:45:08 boring 2017-11-29 17:50:28 oh, there are interesting details to it, some of which have already landed in Alpine too :) 2017-11-29 17:52:13 jvoisin: why no use alpine in prod? 2017-11-29 17:53:52 <_ikke_> jvoisin: that APKBUILD has some issues though (which you might be aware of) 2017-11-29 18:36:11 ncopa: we've got more than 5k machines in production on Debian, and we didn't find a reason good enough™ to move 2017-11-29 18:36:23 _ikke_: do feel free to tell :) 2017-11-29 18:41:11 I don't really think distros matter that much anymore, and will matter even less in the near future. The future belongs to minimal host OS's running containerized apps, and with that minimalism comes the end of a lot of redundant work that is maintaining a distribution 2017-11-29 18:42:13 even currently the base distro doesn't matter anymore, they're basically all the same in the big picture - some advantages, some disadvantages 2017-11-29 18:42:44 That doesn't sound like a very nice world 2017-11-29 18:44:31 on the contrary 2017-11-29 18:45:00 application isolation is a great idea, and application packaging moves away from distro packages 2017-11-29 18:45:41 sure, every solution to a problem brings with it a new problem, but is there really any way to prevent that... 2017-11-29 18:46:06 containers don't offer much in the way of isolation 2017-11-29 18:46:46 Their aim is not isolation but separating components into instances. 2017-11-29 18:52:32 I feel I've had this conversation before 2017-11-29 20:13:27 how do i compress a manpage when install'ing it from an APKBUILD file? 2017-11-29 20:15:10 <_ikke_> jvoisin: mostly that it does not build a specific version 2017-11-29 20:16:13 sure 2017-11-29 20:16:18 but we didn't do a release yet :) 2017-11-29 20:17:11 <_ikke_> right 2017-11-29 20:18:05 soon™ 2017-11-29 20:19:09 liwakura: that should be done automatically by abuild 2017-11-29 20:19:26 nmeum: how do i trigger that? by including the *-doc image? 2017-11-29 20:19:42 yes, by defining a -doc subpackage 2017-11-29 20:20:01 subpackages="$pkgname-doc" should be sufficient 2017-11-29 20:20:42 ah, accidentally solved that fuckup, but dont know how exactly 2017-11-29 20:21:09 lets just hope my crap never ends up in stable 2017-11-29 20:23:47 hi all 2017-11-29 20:24:10 just wanna ask something 2017-11-29 20:25:48 i have tried 2 times alpine linux inside Virtualbox win7(host) i downloaded the VIRTUAL iso file, and both times i was not able to run an X server , is there anything broken ? or do i made something wrong, i have read the wiki enough 2017-11-29 20:26:39 is that maybe the wrong iso? should i go for the Standard iso? 2017-11-29 20:27:09 adilix: "not able to do", did a wizard jump by and disabled you from doing it? what did happen? 2017-11-29 20:27:22 an error when using startx? 2017-11-29 20:27:28 or were you unable to install the xorg stuff? 2017-11-29 20:28:07 after i installed the x server and run startx it got broken, slike server not found or , i had installed firstly vesa drivers and then the VBox additions 2017-11-29 20:28:20 yes an error starting x server 2017-11-29 20:28:39 both didn't worked 2017-11-29 20:28:56 i remember normally with vesa drivers it should work 2017-11-29 20:29:54 i got no screen found if i remember right 2017-11-29 20:30:11 should i have to configure X like the old days 2017-11-29 20:30:21 in xorg.conf.d or so 2017-11-29 20:31:23 adilix: how did you install xorg? 2017-11-29 20:31:36 with the setup-xorg-base script? 2017-11-29 20:31:54 the first time 1 by 1 package the second time i used the setup script like it says in the wiki 2017-11-29 20:33:27 i have to say , the first time i was able to start the x server but i got only a black screen 2017-11-29 20:33:48 i wasn't able to use the WM (TWM) or any mouse or keyboard 2017-11-29 20:34:42 i'd look into the Xorg.log then 2017-11-29 20:34:54 hmm don't have it anymore theVM 2017-11-29 20:35:25 but it's ok when i get the VIRTUAL iso for VBox? right 2017-11-29 20:35:38 i mean the X server works there 2017-11-29 20:40:11 is there any number of how many packages you provide? 2017-11-29 20:40:23 i am on the packages site but can't see any number 2017-11-29 20:46:48 adilix: FYI im just another alpine user 2017-11-29 20:47:13 how do you mean ? 2017-11-29 21:02:45 adilix: "i" dont provide any packages 2017-11-29 21:02:51 because im not maintainer/distributor 2017-11-29 21:06:21 aaa 2017-11-29 21:06:36 i asked how many poackages provide alpine linux not you :) 2017-11-29 21:09:07 <_ikke_> adilix: I count 5550 packages in main x86_64 2017-11-29 21:09:16 <_ikke_> (3.6) 2017-11-29 21:09:16 ok thank you 2017-11-29 21:09:36 <_ikke_> 2882 in community 2017-11-29 21:09:45 <_ikke_> (I just did a grep on one of the mirrors) 2017-11-29 21:09:47 very good 2017-11-29 21:10:05 so... around 10000 pkg's 2017-11-29 21:10:22 <_ikke_> in that orther of magnitude 2017-11-29 21:10:23 it's ok will check what kind of software is around 2017-11-29 21:30:07 last question for today, what is the edge repo? testing? 2017-11-29 21:31:49 <_ikke_> adilix: edge is a rolling release 2017-11-29 21:32:01 <_ikke_> releases are cut off edge 2017-11-29 21:32:12 hmm what does that mean, do i have to download another iso? for that 2017-11-29 21:32:29 or just enable it in the repo file 2017-11-29 21:32:34 <_ikke_> the latter 2017-11-29 21:32:45 <_ikke_> it's just another repository 2017-11-29 21:32:52 i should also use the number , like 3.6 2017-11-29 21:32:53 <_ikke_> but it's less stable 2017-11-29 21:33:10 <_ikke_> no, it's instead of that version 2017-11-29 21:33:24 adn which repo comes forward ? i mean when i have enable it with the 3.6 repo 2017-11-29 21:33:36 <_ikke_> adilix: you define 2 things: 2017-11-29 21:33:47 <_ikke_> / 2017-11-29 21:34:06 yes the main and the community with a number at last 2017-11-29 21:34:25 <_ikke_> edge is a 'branch' 2017-11-29 21:34:32 <_ikke_> edge/main or edge/community 2017-11-29 21:34:35 <_ikke_> or edge/testing 2017-11-29 21:35:13 ok that means i should only go for 3.6/main and 3.6/community 2017-11-29 21:36:12 <_ikke_> right 2017-11-29 21:36:31 ok thx 2017-11-29 21:36:52 will give another try tomorrow, this time i will take the STANDARD iso 2017-11-29 21:45:27 thx for now and cya 2017-11-29 21:52:58 how does apk deal with manually changed config files on upgrades? 2017-11-29 21:53:08 i cant find documentation for this 2017-11-29 21:57:28 yes, it does. if the configuration changed on upgrade the new config file is placed in '${oldconfigfilename}.apk-new' the old one is left untouched 2017-11-29 21:58:21 is this written down somewhere? 2017-11-29 22:26:41 https://wiki.alpinelinux.org/wiki/User:Ncopa/Install_Handbook 2017-11-29 22:26:43 sort of right there 2017-11-29 22:27:20 also right here 2017-11-29 22:27:23 https://wiki.alpinelinux.org/wiki/ISP_Mail_Server_Upgrade_2.x#Upgrading_Alpine_Linux 2017-11-29 22:27:33 but I'd call that poorly documented 2017-11-29 22:33:47 How can I resolve an 'unsatisfiable constraints' issue? https://gist.github.com/sabrehagen/73b13580c790fa2a8c92cf12dd5f144c 2017-11-29 22:41:18 Have you tried running the command again ? 2017-11-29 22:42:24 From the error message it looks like it has to do with busybox being updated 2017-11-29 22:43:11 wait, I misread.. 2017-11-29 22:45:07 Sorry, I don't know that much either. Can't really help. It does look to me as if it probably has to do with the update of the apk-tools. 2017-11-29 22:45:58 And I would (and perhaps others too) like to know what happens if you now run just : 2017-11-29 22:46:08 ``` 2017-11-29 22:46:09 apk add mongodb@testing 2017-11-29 22:46:10 ``` 2017-11-29 22:47:33 (since "apk upgrade --update" seems to have completed sucessfully) 2017-11-29 22:48:23 PS - is this from within a script ? ("Step 2/5 : RUN echo [...] "). If so what script is that ? 2017-11-29 22:49:06 oh - sabrehagen :) ! ^^^^^^^^^^ 2017-11-29 22:57:43 hi there 2017-11-29 22:58:16 do you know how to set ulimit in initscript 2017-11-29 22:58:19 ? 2017-11-29 22:58:43 looks like rc_limit is ignored 2017-11-29 22:58:47 https://github.com/OpenRC/openrc/blob/master/etc/rc.conf#L128 2017-11-29 23:14:13 it's rc_ulimit, not rc_limit Mo0O 2017-11-29 23:19:40 Shiz: oops sorry, typo error here, but not in service file 2017-11-29 23:19:53 but it still doesn't work 2017-11-29 23:19:56 :D 2017-11-30 03:36:36 hey, i try to chroot into a fs dir with mounts from the hosts /proc and /dev and get failed to run command ‘/bin/sh’: No such file or directory but there is bin/sh in the dir... 2017-11-30 03:37:09 what am i missing here :/? 2017-11-30 03:45:15 yo 2017-11-30 03:45:26 claws-mail depends on libenchant but doesn't install enchant by default 2017-11-30 03:45:35 ie missing dependency 2017-11-30 04:11:05 the answer to the /bin/sh thing is to do `apk upgrade --update --available` 2017-11-30 04:11:30 it needs a newer apk-tools to properly solve the dependency 2017-11-30 06:07:33 mecasou, probably /lib 2017-11-30 06:08:05 ENOENT can mean the PT_INTERP (dynamic linker, /lib/ld-*) for the program doesn't exist 2017-11-30 06:28:45 The number of packages missing is slowly shrinking. Now there are 5 packages (and possibly their dependents) holding me back from switching my primary computers to Alpine. 2017-11-30 06:33:32 i can't seem to get any DNS server running on alpine. have been banging my head against a wall for ~48 hours now, any help appreciated! https://forum.alpinelinux.org/forum/networking/no-dns-server-working-alpine-362 2017-11-30 06:41:32 edited that post to add a tcpdump in case it's helpful 2017-11-30 06:48:58 sugarpuff: stupid question, but have you made sure it's not a firewall (either on the node itself or upstream) ? 2017-11-30 06:50:21 detha: on another IRC channel that's what someone else also thinks it might be, that my hosting providing is either blocking communication with DNS roots, or they're blocking packets from my host 2017-11-30 06:50:39 i'm now trying to figure how i can tell either unbound or powerdns to not query the roots itself.. if that's possible? 2017-11-30 06:52:01 not really - unless you somehow manage to axfr a full root zone from somewhere 2017-11-30 06:52:46 pick one of the root servers, and try 'dig @root.server.ip.addr com ns' 2017-11-30 06:56:11 detha: yeah i tried with the d root server and it timed out 2017-11-30 06:57:36 that is....interesting. and 'dig @9.9.9.9 com ns' ? 2017-11-30 06:57:38 at least, what i think is the d root server's ip 2017-11-30 06:57:59 detha: that worked 2017-11-30 06:58:25 said stuff like "com. 55284 IN NS a.gtld-servers.net." etc 2017-11-30 06:59:10 strange. ok, try 'dig @192.43.172.30 com ns' 2017-11-30 06:59:19 (that's another root) 2017-11-30 06:59:42 detha: nope. :( 2017-11-30 07:00:00 seems like i should contact my hosting providing to find out what's going on at this point 2017-11-30 07:00:18 and 'dig +tcp @192.43.172.30 com ns' ? 2017-11-30 07:00:35 detha: that worked! 2017-11-30 07:01:40 aha. so something specifically blocks UDP/53 to root servers. You could investigate further with traceroute,mtr and friends, or ask whoever provides connectivity to that machine 2017-11-30 07:03:07 Out of curiosity, who is that? So I can avoid them. 2017-11-30 07:04:47 detha: lol, it's RamNode, and they're pretty great actually 2017-11-30 07:05:15 detha: i figured out a workaround based on your +tcp suggestion! unbound.conf has a setting: tcp-upstream: yes 2017-11-30 07:05:20 i set that and it worked!!! :D 2017-11-30 07:06:11 detha: thanks so much for helping me troubleshoot this! 2017-11-30 07:06:19 yeah, that will do nicely as a work-around until you can figure out what is actually going on. Small performance penalty, but at least it works. 2017-11-30 07:06:27 yep :) 2017-11-30 07:06:40 i'll send them an email, they're usually pretty good with support (so far at least) 2017-11-30 07:49:06 hello guys, does any of you tried to get edac running on alpine linux? i did not find it in repo and when i compiled it it didn't work 2017-11-30 10:32:37 bitchecker: nice nick 2017-11-30 10:32:43 i still remember that story. 2017-11-30 12:54:16 engblom: hey, any updates I missed on the zfs root thing you're doing? 2017-11-30 14:25:36 kahiru: Probably not... I have zfs to boot directly withot a separate boot partition. The only thing not working so far is /dev/disk/by-id/*. Instead I have to use the less desired /dev/sd* 2017-11-30 14:25:51 without* 2017-11-30 14:26:46 still pretty cool. Have you tried the same setup but on top of dm-crypt container? 2017-11-30 14:27:28 No 2017-11-30 14:28:37 Now that ZFS works, I will consider to move my home server to Alpine... 2017-11-30 14:29:21 makes sense 2017-11-30 14:29:39 My desktop migration will have to be postponed until Alpine has the last packages I am missing. 2017-11-30 14:29:47 what are you missing? 2017-11-30 14:30:32 did you see my comment on that problem yesterday engblom ? 2017-11-30 14:31:10 There is not even one package for 3D design/printing: Blender, OpenSCAD, Slic3r. Then I am missing wxMaxima (I often use it when I calculate stuff for the 3D models). 2017-11-30 14:31:25 And also actively maintained SBCL. 2017-11-30 14:31:32 TBB: I think I missed it 2017-11-30 14:32:26 engblom, /dev/disk/by-id/ is probably something that udev creates, but Alpine defaults to mdev which doesn't necessarily create those structures 2017-11-30 14:32:37 that was my theory for an explanation 2017-11-30 14:33:22 after all, contents in that directory are just symlinks to /dev/sd* most of the time 2017-11-30 14:34:00 TBB: The same it is in Ubuntu. They earlier had buggy Grub, and the problem disappeared with Grub updates in newer versions of Ubuntu. 2017-11-30 14:34:37 TBB: When installing Grub, it tries to find the disks directly under /dev/ in Alpine. 2017-11-30 14:39:06 I do not know if I need to be granted rights at the wiki, otherwise I could put up a wiki on how to currently install to a ZFS-only disk. 2017-11-30 16:56:27 hi all 2017-11-30 16:56:28 again 2017-11-30 16:56:48 i just installed the STANDARD in MY VBox but same problem again, no xserver 2017-11-30 16:57:09 i get (EE) no screens found (EE) 2017-11-30 16:58:03 do i have to write a xorg conf file like the old days? 2017-11-30 17:12:30 Did you set it up using setup-xorg-base? 2017-11-30 17:13:01 yes 2017-11-30 17:13:26 installe dth VM modules enabled it in etc/modules but nothing 2017-11-30 17:13:37 this is really weird now 2017-11-30 17:14:32 just installed the vesa driver and rebooted and now it starts but black screen 2017-11-30 17:14:40 the resolution is 1024x768 2017-11-30 17:15:04 no mouse no keyboard 2017-11-30 17:15:47 should i install the vbox additions from the iso ? 2017-11-30 17:16:05 And if you exit back to the console? Ctrl-Alt-F1, if that works in Alpine. I'm afraid I haven't really used XOrg very much ontop of Alpine. 2017-11-30 17:16:23 I don't think the iso contains any files that are usable by Alpine. 2017-11-30 17:16:54 i am back now with ctrl alt f2 2017-11-30 17:17:10 i think there is a BUG 2017-11-30 17:17:23 found that shit https://bugs.alpinelinux.org/issues/7406 2017-11-30 17:18:01 Might be that there's not really anything for Xorg to start. 2017-11-30 17:18:12 Unless you've also installed xinit and some or other WM. 2017-11-30 17:18:24 it's installed 2017-11-30 17:18:31 i am not new in the linux world 2017-11-30 17:19:02 Have to ask. A lot of the distros install basically everything along with Xorg after all. 2017-11-30 17:21:34 even the vboxvideo module isnt there, WTF? 2017-11-30 17:24:10 Oracle would have to provide it, I believe. 2017-11-30 17:24:35 oracle has nothing to do , i have installed the virtualbox-modules from the repo 2017-11-30 17:26:21 i think the maintainer made i fault 2017-11-30 17:26:40 that vboxvideo.ko must be in the kernel modules folder normally 2017-11-30 17:26:40 Does that contain the Xorg drivers normally though? Or just the kernel modules? 2017-11-30 17:26:54 yes thats the video driver 2017-11-30 17:27:06 to get 3d accelaration 2017-11-30 17:27:26 but it doesnt get installed with installing the modules 2017-11-30 17:27:40 https://wiki.alpinelinux.org/wiki/VirtualBox_shared_folders 2017-11-30 17:27:53 https://wiki.alpinelinux.org/wiki/VirtualBox_guest_additions 2017-11-30 17:28:44 look at the second link where it say echo vboxdrv >> /etc/modules, this driver afaik loads the vboxvideo.ko for the video drivers 2017-11-30 17:28:54 and then you have 3d accel in the VM 2017-11-30 17:29:10 there is something wrong in the package 2017-11-30 17:33:25 https://pkgs.alpinelinux.org/contents?branch=v3.6&name=virtualbox-guest-modules-hardened&arch=x86_64&repo=community 2017-11-30 17:34:04 look at the contents of the package its inside misc but when i load the module with modprobe i get the message module not availabe 2017-11-30 17:34:51 and the documenatation in the wiki for Virtualization should be changed 2017-11-30 17:35:09 especially this one https://wiki.alpinelinux.org/wiki/VirtualBox_guest_additions 2017-11-30 17:36:34 anyway thx for listening maybe if the distro gets better will try it again cya 2017-11-30 17:56:22 if an APKBUILD dumps core during the build, where can I find the core file? 2017-11-30 18:04:48 on hardened is doesn't dump by default. 2017-11-30 18:05:16 ulimit -c is zero by default even on vanilla, unless abuild sets it ? not sure 2017-11-30 18:05:30 running vanilla 2017-11-30 18:06:05 did it explicitly say (core dumped) or just a segfault, abort or w/e triggered it to crash? 2017-11-30 18:06:46 it said core dumped 2017-11-30 18:07:19 if you kept the build files around during the process it might well be somewhere in the tree then 2017-11-30 18:07:33 I just did abuild -r 2017-11-30 18:07:37 is there an option to keep build files? 2017-11-30 18:07:47 -K 2017-11-30 18:07:50 ah, thanks 2017-11-30 18:08:11 yes 2017-11-30 18:08:14 K 2017-11-30 18:09:05 SirCmpwn: just for curiousity, sth you are up to? 2017-11-30 18:09:15 building openjdk7 2017-11-30 18:09:17 if it occurs again while running with -K it should be in the cwd of whereever the build process was at the time 2017-11-30 18:09:20 first at all, then with patches 2017-11-30 18:09:23 alright 2017-11-30 18:09:27 SirCmpwn: I wish you luck :p 2017-11-30 18:09:35 thanks, I'll need it 2017-11-30 18:09:38 might be forced to resort to like `find . -name core` 2017-11-30 18:09:40 lol 2017-11-30 18:09:42 building massive projects like that always scares me 2017-11-30 18:09:48 yeah good luck SirCmpwn 2017-11-30 18:09:58 java is a difficult beast to tame 2017-11-30 18:10:01 Elizabeth: im currently building an kernel :P 2017-11-30 18:10:14 machine at 100% for half an hour... 2017-11-30 18:10:16 I mailed the package maintainer as well to ask if they can still build it on a fresh box 2017-11-30 18:10:17 SirCmpwn: what does kernel.core_pattern look like 2017-11-30 18:10:38 uh 2017-11-30 18:10:43 scary 2017-11-30 18:10:46 lol 2017-11-30 18:10:48 /usr/lib/systemd/systemd-coredump %P %u %g %s %t %c %e 2017-11-30 18:10:52 oh no 2017-11-30 18:10:53 systemd 2017-11-30 18:10:53 I'm running in a docker though, so that makes sense 2017-11-30 18:10:54 oh god 2017-11-30 18:11:04 wonder where systemd throws it then 2017-11-30 18:11:07 maybe systemd on the host has the core lying around 2017-11-30 18:11:08 there's coredumpctl 2017-11-30 18:11:10 it's not going to be in your tree though 2017-11-30 18:11:11 its in uh 2017-11-30 18:11:14 ffs 2017-11-30 18:11:15 i forget 2017-11-30 18:11:16 sec 2017-11-30 18:11:24 yeah coredumpctl has it 2017-11-30 18:11:28 /var/crash/ 2017-11-30 18:11:31 i think? 2017-11-30 18:11:37 no 2017-11-30 18:11:41 timestamped folder then in there 2017-11-30 18:11:42 you do coredumpctl list | grep whatever 2017-11-30 18:11:43 ah 2017-11-30 18:11:55 then coredumpctl dump [pid] > dump.core 2017-11-30 18:12:01 /var/crash seems to be kernel dumps anyway 2017-11-30 18:12:17 SirCmpwn: well, that solved about an hour of running around then at least 2017-11-30 18:12:41 apparently systemd just parks them in /var/lib/systemd/coredump 2017-11-30 18:12:46 aye, thanks 2017-11-30 18:12:48 surprising, i was expecting some fancy database format and everything 2017-11-30 18:12:57 I'm willing to bet it's exhaustion of resouces, though, honestly 2017-11-30 18:13:12 me too 2017-11-30 18:13:15 but it'd be nice to be sure 2017-11-30 18:13:24 I'll be trying it with full virt next 2017-11-30 18:13:35 is it an apkbuild from 3.6 ? 2017-11-30 18:13:40 no, from HEAD 2017-11-30 18:13:44 I should check out 3.6, good call 2017-11-30 18:13:54 lemme take a whack at it on my build host and see if it also bombs out 2017-11-30 18:13:58 yeah it got SIGABRT 2017-11-30 18:14:41 community/openjdk7 on HEAD, let's give this a shot 2017-11-30 18:14:56 SirCmpwn: moaaaar memmmorrrry! 2017-11-30 18:15:10 (would've been worth checking dmesg first then I guess) 2017-11-30 18:15:19 yeah fair 2017-11-30 18:15:41 hmm, dmesg doesn't have anything to say 2017-11-30 18:15:44 my debugging process is pretty much always checking dmesg first, saves a lot of guesswork 2017-11-30 18:15:56 huh, interesting, if it was OOMing it should've printed something there 2017-11-30 18:16:00 yeah 2017-11-30 18:16:06 unless .. containers 2017-11-30 18:16:11 yeah, unless containers 2017-11-30 18:16:14 gonna spin up a VM 2017-11-30 18:16:15 I wonder if the host dmesg 2017-11-30 18:16:15 containers~ 2017-11-30 18:17:53 is there an easier way of making usb drive with alpine with zfs than creating a custom iso with zfs, booting the iso and then doing setup-bootable? 2017-11-30 18:19:02 https://astr.al/u/f4414607_1036x139.png eargh i feel dirty 2017-11-30 18:23:05 scv: just be thankful it isn't your job 2017-11-30 18:23:19 what, building alpine pkgs or java ;) 2017-11-30 18:24:05 I wish my job involved _more_ building of alpine packages 2017-11-30 18:24:34 convince them to move :D 2017-11-30 18:57:56 hi there 2017-11-30 18:58:48 i'm trying install of alpine, i'm running live ( from usb with standard x86 iso ) 2017-11-30 18:59:10 but i get a strange thing with setup-interfaces & my wifi 2017-11-30 18:59:17 everything looks fine 2017-11-30 18:59:26 but i don't get an ip 2017-11-30 19:00:52 it show my SSID, i type password 2017-11-30 19:00:57 looks good 2017-11-30 19:01:26 and udhcpc wlan0 says me Network is down 2017-11-30 19:02:41 ok it missed the -i option 2017-11-30 19:03:08 now it's working 2017-11-30 19:03:15 glad we could help 2017-11-30 19:03:23 lol SirCmpwn 2017-11-30 19:04:09 but the setup-interfaces was not working 2017-11-30 19:07:27 how can i inspect why ? 2017-11-30 19:08:58 module iwl3945, but i'm not sure it changes something 2017-11-30 19:32:09 http://wiki.alpinelinux.org/wiki/LVM_on_LUKS#Creating_the_Partition_Layout <= isn't that 100M rather than 100m ? (in the creation of the boot partition) 2017-11-30 19:39:53 woo, it built in a VM 2017-11-30 19:40:00 now to backport this patch 2017-11-30 19:43:02 jeanjack: "Enter n → p → 1 → 1 → 100m to create a new 100 MB primary partition."? Seems to work both, though fdisk says "+size{K,M,G,T,P}" 2017-11-30 19:55:43 How reliable is Alpine's NFS support? Tempted to switch my home server from FreeNAS to alpine as I want to be able to run Docker (not in a VM) for a couple things... 2017-11-30 19:56:19 about the same as anywhere else i'd imagine 2017-11-30 19:56:22 other thing: docker host on hardened kernel has some issues 2017-11-30 19:56:25 it's implemented in-kernel in linux 2017-11-30 19:56:28 Not NFS, ZFS* 2017-11-30 19:56:30 oh 2017-11-30 19:56:30 My bad 2017-11-30 19:56:36 can't tell you that one 2017-11-30 19:57:16 I use it in Arch and it's not amazing but I think that comes down to it not being a package in Arch itself... so you can end up being offered a kernel without a compatible ZFS module 2017-11-30 19:59:59 jomat: maybe need a + so. I tried and it makes me a higher volume 2017-11-30 20:00:22 used 100 sectors if i remember well 2017-11-30 20:03:30 i'm following the rest of the wiki but after PV creation i just get a 8M size 2017-11-30 20:05:14 cryptsetup open make the device available and uncrypted in /dev/mapper if i understand well, but it looks like it isn't well mapped :s 2017-11-30 20:29:23 Hm, it'd be nice if lbu supported ZFS snapshots... 2017-11-30 20:35:45 Also, another q. If I install in diskless or data mode, can I have it automatically restore the latest apkovl when power is restored? 2017-11-30 20:43:25 yes 2017-11-30 20:44:01 if will use any *.apkovl.tar.gz it finds during boot 2017-11-30 20:48:49 ncopa: Where does it search? 2017-11-30 20:49:55 Wondering if I could modify the Alpine install on the USB stick to add ZFS support and store my apkovl on my ZFS pool? 2017-11-30 20:55:06 ok. i finally removed the partition, created a new one without linux lvm type and now looks good. i can't get why it did fail first 2017-11-30 21:34:54 Hi all, newbie here. I was wondering if anyone has any tips for getting gnome working - I can't seem to get the gnome_setup from the wiki working properly, and I've tried a number of other dm's besides slim and lxdm and nothing is working right. I'm intrigued by Alpine though, so I'm not going to give up on it. 2017-11-30 21:37:54 slim is a login manager aint it? 2017-11-30 21:37:55 hmm anyway 2017-11-30 21:38:01 xfce should def work, but kaniini is the local gnome expert 2017-11-30 21:38:26 bpye: it searches all found blockdevices it manages to mount 2017-11-30 21:38:49 I believe you're correct on slim. I was able to get xfce working but don't really like using it. I also tried out awesome, but again, not my style. 2017-11-30 21:40:02 should add that I'm currently running Alpine in a virtualbox vm to try it out 2017-11-30 21:46:32 russ_: gnome_setup sounds like something old though -- have you checked the date of the wiki page? 2017-11-30 21:46:36 sadly a lot on the wiki page is outdated 2017-11-30 21:47:32 Shiz, it says it was last edited by Ttrask on 25 Mar 2016. So yeah, a little dated. 2017-11-30 21:47:50 https://wiki.alpinelinux.org/wiki/Gnome_Setup 2017-11-30 21:48:07 i think kaniini completely threw around the gnome packages earlier this year 2017-11-30 21:49:20 I mean, correct me if I'm wrong, but from the pkgs.alpinelinux.org page it doesn't even look like there's a package named "gnome-base" anymore. 2017-11-30 21:52:03 yes, that page is hopelessly outdated 2017-11-30 21:52:16 I believe the correct package to install is gnome-shell 2017-11-30 21:52:33 noted that it's in testing 2017-11-30 21:53:41 once you get it working, russ_, update the wiki 2017-11-30 21:54:39 If I do, I absolutely will. 2017-11-30 22:43:58 3.7.0 is out 2017-11-30 22:46:18 yay~ 2017-11-30 22:46:24 time to go push some updates 2017-11-30 23:38:20 Hey, so 3.7 - is considered as stable now ? 2017-11-30 23:40:37 yes 2017-11-30 23:42:15 Oop. Well I'm glad I didnt install yet today 2017-11-30 23:42:47 cool :) cheers and congrats 2017-11-30 23:42:52 ncopa: Hm okay, I have no idea how that plays with ZFS, guess I'll have to try 2017-11-30 23:45:17 nice, i just installed 3.6 today ^^