2019-04-01 10:42:22 when run install media (tar ball unpacked to disk in this case) how is the modloop loaded and mounted? i.e. do the initramfs must have kernel modules for disk drive 2019-04-01 10:45:44 where modloop should be mounted during install boot 2019-04-01 11:54:57 hey, does somebody know how the filenames are generated in /etc/ssl/certs? i mean the ones like f081611a.0 2019-04-01 11:56:00 fassl: openssl x509 -hash -noout -in $IN_FILE 2019-04-01 12:00:37 danieli, ty 2019-04-01 12:03:43 hm, wget still does not accept the self signet cert 2019-04-01 13:21:03 aah. thx. 2019-04-01 13:22:28 I am signing up for a linode alpine now. No others offer it that I can see. 2019-04-01 13:22:38 :) 2019-04-01 13:22:57 any kvm (or even openvz) vm will do 2019-04-01 13:23:27 it's just so much easier when you don't have to install and fiddle or hack a lot 2019-04-01 13:23:31 not offical obviously but its pretty easy to replace alpine with just ssh 2019-04-01 13:23:37 replace with* 2019-04-01 13:23:40 yeah 2019-04-01 13:23:42 click a button and bam, you got an alpine box 2019-04-01 13:23:50 but kvm is as easy as dd ; reboot 2019-04-01 13:24:01 openvz is more work 2019-04-01 13:24:13 i'm kinda surprised openvz is still around 2019-04-01 13:24:17 i used to work with it some years ago 2019-04-01 13:24:17 yeah 2019-04-01 13:24:37 its terrible lol 2019-04-01 13:24:42 still 2.6.32 2019-04-01 13:24:51 i'm not surprised 2019-04-01 13:33:48 If the reason for choosing Alpine is an efficient low overhead system, running it under a high overhead bulky wasteful operating system defeats that purpose and removes the incentive to use Alpine in the first place. 2019-04-01 13:34:17 the idea is to replace the running OS with alpine when the hosting provider does not provide you with alpine images 2019-04-01 13:35:33 OK. I can not recommend that to all our members. 2019-04-01 13:39:57 not with linode, but that's what other people are trying to say, I believe 2019-04-01 13:40:06 EarthEngineer: also, if I may ask, who is 'our' in this case? 2019-04-01 13:43:56 Generation Hip Hop 2019-04-01 13:45:03 I am using icecast on Alpine for their Azurecast repeaters. 2019-04-01 13:46:25 http://www.uhhm.org/generation-hip-hop/ 2019-04-01 13:51:39 cool! 2019-04-01 13:57:35 I will ask ECI Networks to provide bare metal instances so we can have UDP off loading enabled. 2019-04-01 14:00:50 jwh: openvz is *still* 2.6.32?! 2019-04-01 14:01:49 was at least about half a year ago when it caused a major problem for a couple of vps's I co-host 2019-04-01 14:02:12 latest testing is el7, kernel 3.10 2019-04-01 14:02:18 latest stable is at 2.6.32 2019-04-01 14:47:25 hey all 2019-04-01 14:48:28 <_ikke_> hey 2019-04-01 14:48:46 so often when i mention alpine, ppl are like "ur security is 5h17 us only gots a small userbase and like 14 developers, traditional distros are like so much better" ... what's your usual response to that strength in numbers argument? (my intuition is that it's bollocks and i'm *feeling* very safe with alpine, but i don't know how to properly reason/argue it) 2019-04-01 14:50:14 less attack surface is one argument 2019-04-01 14:50:48 neither 'small userbase' or 'like 14 developers' are sensible arguments 2019-04-01 14:50:57 "but how do you know they're quick enough at plugging CVEs" 2019-04-01 14:51:00 we have a pretty large userbase, but we don't keep metrics on it 2019-04-01 14:51:29 how do you know we're not? 2019-04-01 14:51:55 that's the good part about that argument: i don't 2019-04-01 14:52:14 do your friends keep actual metrics on it? 2019-04-01 14:52:31 <_ikke_> and just focussing on CVE's is also not productive 2019-04-01 14:52:37 to the best of my knowledge i've never had any intrusion into one of my alpine boxes (contrast that with the debian boxes) 2019-04-01 14:53:01 at the moment, it's pretty much best-effort, and we still don't have a security advisory program, but it's work in progress 2019-04-01 14:53:05 but if i mention that then it's always "that's probably dumb luck or because your alpine boxes are younger" 2019-04-01 14:53:37 it sounds like elitism from your friends, to be honest 2019-04-01 14:53:46 hmm 2019-04-01 14:53:56 several of the arguments you brought up don't really make sense, and some are false 2019-04-01 14:53:58 i would say elitism used as a shield against something new/unknown, perhaps 2019-04-01 14:54:36 it's often a holy war when it comes to operating systems, init systems, programming languages, and so on 2019-04-01 14:55:05 the problem is that we'd like to reinstall our shared VM server and use I'd like to use Alpine as the hypervisor but i'm outgunned rhethorically 2019-04-01 14:55:59 <_ikke_> The small attack surface should be a compelling argument then 2019-04-01 14:56:12 <_ikke_> You want your hypervisor to be as small as possible 2019-04-01 14:56:14 if you want to bring up raw statistics, go with this 2019-04-01 14:56:19 https://www.cvedetails.com/product/36/Debian-Debian-Linux.html 2019-04-01 14:56:19 https://www.cvedetails.com/product/38838/Alpinelinux-Alpine-Linux.htm 2019-04-01 14:56:32 it's not a fair argument, of course, but it'll give them a taste of their own medicine 2019-04-01 14:56:36 yay!! 2019-04-01 14:56:40 thank you both :D 2019-04-01 14:56:46 ACTION loves statistics 2019-04-01 14:57:03 then you should also realize why using those statistics is a bad idea 2019-04-01 14:57:11 but still, statistics is statistics :P 2019-04-01 14:57:27 <_ikke_> statistics can be manipulated to show whatever you want 2019-04-01 14:57:28 well, *rhethorically* they work very well ;-) 2019-04-01 14:57:32 exactly _ikke_ 2019-04-01 14:57:58 just linked to them ... we'll see what happens 2019-04-01 14:58:00 alpine: 2 vulns, debian: 2272 2019-04-01 14:58:10 I'm giggling on the inside 2019-04-01 14:58:22 ur 2019-04-01 14:58:30 for 217: 2 vs 242, right? 2019-04-01 14:58:34 2017* 2019-04-01 14:59:01 just go with the total amount 2019-04-01 14:59:32 since you're manipulating it yourself, you can just as well trust it, then? :) 2019-04-01 15:00:34 whether distros patch CVEs or not, it's a bad idea to blindly trust that everything is in order 2019-04-01 15:00:43 if you're truly security-minded, you'll want to keep track of it yourself 2019-04-01 15:00:58 <_ikke_> One argument from the linux kernel I like a lot: A lot of bug fixes end up fixing security issues without being marked as such. By just focussing on CVE's you miss out a lot 2019-04-01 15:01:18 oh yes, a lot of security issues aren't even considered security issues in the public eye 2019-04-01 15:01:21 they're just another commit 2019-04-01 15:02:42 <_ikke_> I recall some vulns in RHEL because they only backported marked security fixes, while some vulnerability was fixed as a regular bugfix 2019-04-01 15:24:24 guys, how do I get rid of the many many firmware packages I'll never need? "apk del some_firmware_file" gives errors like: "World updated, but the following packages weren't removed due to: ..." 2019-04-01 15:26:23 <_ikke_> silverwillow: Currently linux-firmware is required by linux-vanilla, so they will always be required 2019-04-01 15:26:51 <_ikke_> The only thing I can think of is by making your own custom linux-firmware meta-package that provides linux-firmware and requires the firmware package that you need 2019-04-01 15:27:42 wow. so somebody went to the trouble of splitting up all the firmware stuff... only you then cannot make use of the split-up-ness of it all? 2019-04-01 15:50:47 silverwillow: apk add linux-firmware-youknowneeded, other will be removed 2019-04-01 15:52:54 for example I have only linux-firmware-rockchip on the box where I'm typing right now 2019-04-01 15:53:25 heheh another buddy just remarked "my clients all use centos, that sucks just as much as RHEL but is much cheaper" 2019-04-01 15:56:03 having all these 'backdors' in hardware (spectre, meltdown, PCH) and concetrating only on CVE is strange 2019-04-01 15:57:12 but security industry needs something good for marketing their 'solutions' 2019-04-01 16:30:09 <^ingo^^^^^> I need a special patch in the musl-libc(extend syslog length), can i build may own package, and later if this come from the upstream back in the alpine main package, switch back? So switch between my own, and the main packages. 2019-04-01 17:11:50 <_ikke_> silverwillow: I think they were still looking for a good sollution 2019-04-01 17:12:40 <_ikke_> silverwillow: They want to give people a nice experience by making sure all firmware is availble at installation, but there is no easy way to detect what firmware is actually necessary 2019-04-01 17:23:32 _ikke_: there was few months age idea to list modules and grep needed firmware with modinfo, but no one worked on it 2019-04-01 17:36:43 here is simple method to find needed firmware: modinfo `lsmod | grep -v Module | awk -e '{ print $1}'` | grep firmware 2019-04-01 18:25:38 id like to share my internet connection from this computer (running the latest alpine) to a laptop that has no wifi drivers available--ive done it before using a different linux distribution so i know it is possible however i dont know how to do it without a GUI. does anyone have a good resource for what i need to set and where so that i can internet connection share? thank you 2019-04-01 18:26:49 <_ikke_> th-end: how many ethernet ports do you have? 2019-04-01 18:27:33 one on this pc and one on my laptop. presently speaking to you over wifi. 2019-04-01 18:27:36 should be possible with only one port 2019-04-01 18:27:51 i'd recommend asking ##networking, i don't remember the entire process off the top of my head 2019-04-01 18:28:07 oh so it wouldnt be alpine specific 2019-04-01 18:28:08 needs a few short lines of iptables and some configuration on the network interface 2019-04-01 18:28:09 i get it 2019-04-01 18:28:10 <_ikke_> You need to enable forwarding and create the proper nat rules 2019-04-01 18:28:12 indeed 2019-04-01 18:28:18 it's not alpine specific 2019-04-01 18:28:22 thanks ! 2019-04-01 18:28:47 <_ikke_> # echo 1 >/proc/sys/net/ipv4/conf/all/forwarding 2019-04-01 18:28:57 <_ikke_> that's step 1 2019-04-01 18:29:04 yeah, and some iptables masquerade rules 2019-04-01 18:29:15 ok so google has me headed in the right direction 2019-04-01 18:29:30 however these instructions are for very old ubuntu versions and may have changed? im not certain. 2019-04-01 18:29:40 ill check ##networking 2019-04-01 18:29:44 <_ikke_> iptables is still the same 2019-04-01 18:29:59 <_ikke_> what is distro specific is how to get it to work at boot 2019-04-01 18:30:07 <_ikke_> at runtime, it's the same almost everywhere 2019-04-01 19:22:32 good news. plugged my brain back in and am happy to report that after installing a bunch of dependencies, the latest GNUstep release tarballs compile cleanly on 3.8.4. 2019-04-01 19:24:40 <_ikke_> cool 2019-04-01 19:38:19 Hi. Just install postgresql-client and getting Error relocating /usr/bin/psql: append_history: symbol not found. 2019-04-01 19:47:14 hippo, the package must have failed to record that it needs a newer version of some library 2019-04-01 19:47:40 try upgrading all the libs it depends on or try to figure out which lib is supposed to provide that 2019-04-01 19:50:29 i think I'll just upgrade the entire thing and see what happens 2019-04-01 19:51:36 <_ikke_> hippo: What environment are you getting that error in? 2019-04-01 19:52:53 like my shell? I'm using fish. Running on virtualbox. v3.8 with some edge items for nodejs 2019-04-01 19:53:43 <_ikke_> is edge pinned? 2019-04-01 19:54:22 I don't remember doing so. How can I tell? 2019-04-01 19:54:59 <_ikke_> check /etc/apk/repositories 2019-04-01 19:55:13 <_ikke_> if it's pinned, it should have @ in front 2019-04-01 19:58:41 yes it is pinned. I just changed my repositories to edge, uncommented all the pinned and will run: apk upgrade --update-cache --available 2019-04-01 20:00:30 <_ikke_> you *should* pin, combining stable and edge can lead to these kinds of incompattibilities 2019-04-01 20:01:59 if I'm only using edge, there's nothing to pin right? 2019-04-01 20:02:12 <_ikke_> if you use only edge, then it should be ok 2019-04-01 20:03:05 yep. that my plan. upgrading right now 2019-04-01 20:03:59 <_ikke_> Though edge itself might be broken from time to time 2019-04-01 20:07:37 I understand 2019-04-01 21:18:47 where are the init.d scripts for dhcrelay and dhcpd for IPv6? 2019-04-01 23:09:48 hi! is there a telnet like command in alpine? if yes in which package? 2019-04-01 23:10:29 busybox-extras? 2019-04-01 23:18:55 thx! :) 2019-04-02 08:02:07 hello 2019-04-02 08:02:34 if I understand correctly, util-linux/coreutils package will replace some parts of busybox? (as the busybox use post-install script to create symlinks) 2019-04-02 08:02:50 <_ikke_> yes 2019-04-02 08:06:00 and if I remove util-linux after that, will busybox symlinks be reinstalled? 2019-04-02 08:06:50 <_ikke_> should be, yes 2019-04-02 08:06:58 okay understood :) 2019-04-02 08:08:18 thanks 2019-04-02 08:08:19 no and yes 2019-04-02 08:09:04 its not post-install that recreates them, thats the trigger script 2019-04-02 08:09:16 <_ikke_> right 2019-04-02 08:10:04 oh wait i think im wrong :) 2019-04-02 08:10:28 i think its path related 2019-04-02 08:11:03 ACTION hides back into a corner 2019-04-02 08:12:39 ok so it does trigger on any change in /bin:/usr/bin:/sbin:/usr/sbin:/lib/modules/ 2019-04-02 10:38:15 hi, anyone tried to do a netboot with RPi + Alpine ? 2019-04-02 10:39:46 regardless what I do I aslways end up with "nfs could not be mounted ..." - raspbian etc are working without a problem, I also played around with "alpine_dev=" etc based on https://wiki.alpinelinux.org/wiki/PXE_boot 2019-04-02 10:42:57 asgoo: yes i am able to netboot in the past 2019-04-02 10:43:10 but the default image does not support it out of the box 2019-04-02 10:45:23 asgoo: https://dev.alpinelinux.org/~clandmeter/other/rpi-netboot.mp4 2019-04-02 10:47:12 what needs changing ? compiling etc is no problem 2019-04-02 10:47:52 check the root of aports 2019-04-02 10:47:59 we have scripts to generate releases 2019-04-02 10:48:24 initramfs needs network support, maybe also nfs if you need it. never used nfs though. 2019-04-02 10:48:55 https://github.com/alpinelinux/aports/blob/master/scripts/mkimg.netboot.sh ? 2019-04-02 10:49:15 thats a profile 2019-04-02 10:49:18 btw no nfs ? from where did you get your data ftp/http ? 2019-04-02 10:49:29 btw sry never build alpine my self ;) 2019-04-02 10:49:36 https://github.com/alpinelinux/aports/blob/master/scripts/mkimage.sh 2019-04-02 10:49:46 alpine boots to tmpfs by default 2019-04-02 10:49:58 with network you can use any repo to build your rootfs 2019-04-02 10:50:15 i think but default it used our cdn 2019-04-02 10:50:42 look for features 2019-04-02 10:50:47 network is one of them 2019-04-02 10:51:06 i think we could enable network by default on rpi 2019-04-02 10:52:15 we have 100ish RPi so getting data from your CDN doesn't sounds reasonable :D 2019-04-02 10:53:35 asgoo: why wouldnt that be reasonable? 2019-04-02 10:53:48 from your side or ours? 2019-04-02 10:54:17 our 2019-04-02 10:54:30 you can setup your own if you like 2019-04-02 10:54:53 but i think nfs is also possible 2019-04-02 10:55:20 both have its advantages 2019-04-02 10:55:33 btw do you have an example what needs changing ? can't find the missing parts (obviously my fault ... but I am not really familar with the buildsys structure) 2019-04-02 10:55:50 features 2019-04-02 10:55:59 check the arm profile 2019-04-02 10:56:05 it should have an rpi section 2019-04-02 10:56:35 https://github.com/alpinelinux/aports/blob/master/scripts/mkimg.arm.sh#L82 2019-04-02 10:57:45 uhh "network" would be nice at features ;) 2019-04-02 10:58:06 guess I try :) tx 2019-04-02 10:58:22 you can finetune it to your liking and gen a new image 2019-04-02 10:59:21 I basically want to stay as vanilla as possible to avoid trickery in future. 2019-04-02 10:59:44 these are all vanilla options 2019-04-02 11:00:02 yea I mean vanilla to the release image 2019-04-02 11:00:11 we support netboot out of the box 2019-04-02 11:00:14 just not for rpi 2019-04-02 11:00:19 not much request for it 2019-04-02 11:01:34 I guess it works similar awesome to the "normal" SD card usage so it would be perfect for my usecase here. 2019-04-02 11:01:55 its a tad slower 2019-04-02 11:02:13 if you dont mind the boot delay then its ok. 2019-04-02 11:02:51 ofc supporting it at default image would be nice too ;) but not my business to decide if you want this or not (no idea about the drawback sizewise) 2019-04-02 11:02:55 as my video shows, it takes time for rpi to fetch the bootloader 2019-04-02 11:03:25 we have netboot directory http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/aarch64/netboot/ 2019-04-02 11:03:27 yea thats okay - I toyed around with raspbian because I was pretty unsure if my setup works or not 2019-04-02 11:03:43 but it doesnt have the rpi kernel included 2019-04-02 11:04:51 so in that case it makes sense to generate your own image and copy the contents to some local webserver. 2019-04-02 11:05:09 you would also have a local apk repo with the packages you included. 2019-04-02 11:05:29 yea thats reasonable, for raspbian its rather easy "root=/dev/nfs nfsroot=10.20.20.20:/tftp/root,tcp,v3 vers=3 rw ip=dhcp rootwait elevator=deadline" , done 2019-04-02 11:06:03 alpine is not made to be easy 2019-04-02 11:07:07 :) its much easier then trying to strip down debian* 2019-04-02 11:07:29 ofc I feel a bit like misusing it for my usecase but it works like a charm at sd card 2019-04-02 11:09:17 +1 tx for help 2019-04-02 11:09:26 no problem, glad to help 2019-04-02 12:06:26 Hello, are you aware of this issue with alpine and dns requests timing out. https://github.com/kubernetes/kubernetes/issues/56903#issuecomment-409603030 2019-04-02 12:07:25 <_ikke_> ronneke2019: I've heard about some musl regression in dns resolving, don't know the details anymore though 2019-04-02 12:09:16 Is the work around that I linked recommended? Or don't you know about it? 2019-04-02 12:12:01 <_ikke_> ronneke2019: https://bugs.alpinelinux.org/issues/10063 2019-04-02 12:51:57 __iki__ : Thanks for the link. For now we're going to do the dns requests with Python, as we are running an alpine container with a python application. And Python seems to do the dns requests correctly 2019-04-02 12:52:22 _ikki_* sorry for misspelling 2019-04-02 12:54:46 use tab completion :) 2019-04-02 13:11:29 @clandmeter sry I am again, maybe I am missing the obvious solution but I can't find a way (google tells me no docs available) how I build http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/armhf/alpine-rpi-3.9.2-armhf.tar.gz 2019-04-02 13:12:16 what are the steps you did now? 2019-04-02 13:15:20 git clone alpinelinux/aports then checkout to 3.9 and change the needed stuff, but then I am a bit out of ideas how I actually "build" it , besides some 3rd party outdated stuff google and the docs werent helpful (not saying that I didn't missed something obvious) 2019-04-02 13:16:12 i think you need to be on the same arch as you build the image 2019-04-02 13:16:39 no crosscompiling ? 2019-04-02 13:16:53 you dont need to compile anything 2019-04-02 13:17:01 and no we do not do cross compiling at all 2019-04-02 13:17:38 except initial toolchain bootstrap when new arch is introduced. 2019-04-02 13:18:12 mkimage.sh --help 2019-04-02 13:20:00 "abuild not found" :) 2019-04-02 13:20:17 apk add alpine-sdk 2019-04-02 13:20:38 you need to be on alpine to build it i think 2019-04-02 13:20:55 oh, that would explain it 2019-04-02 13:22:05 btw I am happy to RTFW but I didn't found anything besides https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage - ofc makes a bit more sense if you need alpine to build alpine 2019-04-02 13:22:57 sry I thought its more an regular crosscompile toolchain + packages, guess I was completly off 2019-04-02 13:22:58 i think that doc is kind of okish 2019-04-02 13:23:05 didnt read the details though 2019-04-02 13:23:19 ofc for rpi it would be an tar.gz image 2019-04-02 13:25:25 btw not sure if this is intended but there is no network support but raid and scsi at initramfs for RPi, not sure this makes sense ;) 2019-04-02 13:26:38 asgoo: as clandmeter has said, I believe netboot (by download kernel + initramfs) from dl-cdn.alpinelinux.org is pretty straightforward and ease of mind 2019-04-02 13:26:51 clandmeter : rpi netboot images dont have network support ? 2019-04-02 13:27:06 we do not have rpi kernel for netboot 2019-04-02 13:27:23 it only has vanilla 2019-04-02 13:28:07 and raid+scsi ;) 2019-04-02 13:29:11 doesnt scsi work over gpio? 2019-04-02 13:29:14 :p 2019-04-02 13:31:17 not if you have an raid controller attatched 2019-04-02 13:32:02 raid is for software raid, i think you could technically do raid over sd and usb 2019-04-02 13:33:02 asgoo: as mentioned, mkimage.sh and mkimg.arm.sh (profile_rpi() ) are 2 files you need to build your custom images 2019-04-02 13:33:26 <- creating new image for rpi to try 2019-04-02 13:34:28 initfs_features 2019-04-02 13:34:36 yes line #82 2019-04-02 13:37:16 I came across sometimes ago, not sure if it helps : https://pi3g.com/2019/01/10/alpine-boot-process-on-the-raspberry-pi/ 2019-04-02 13:37:59 i think it would have been a bit more clear if the profiles would be have been moved to a profile.d dir 2019-04-02 14:33:16 back again, https://pastebin.com/02TfE44E "update-kernel: unrecognized option: modloopfw" any ideas ? 2019-04-02 14:34:55 thats is pure git clone of aports, same error if I build edge instead of 3.9 2019-04-02 14:36:17 hello, I've got a question about php7-pecl-imagick package: https://pkgs.alpinelinux.org/package/v3.9/community/x86_64/php7-pecl-imagick `php -i` with that package shows weird stuff: `Imagick compiled with ImageMagick version => ImageMagick 7.0.8-19 Q16 x86_64 2018-12-20 ` `Imagick using ImageMagick library version => ImageMagick 7.0.8-23 Q16 x86_64 2019-01-10` 2019-04-02 14:36:33 where did that 7.0.8-19 Q16 x86_64 2018-12-20 get taken from? 2019-04-02 14:37:13 asgoo: aports version need to match alpine version 2019-04-02 14:37:53 aports is master and i also tried edge as version, same error 2019-04-02 14:38:02 did you apk upgrade -a? 2019-04-02 14:39:23 ups (hide) 2019-04-02 14:41:55 hmm works ... sry my bad - now just my broken sd partition strikes again, any idea how to setup alpine that it actually boots ? I cleaned the sd card, created a fat 32 partition with boot flag at ubuntu (gparted and disk utility), copied files over -> rpi is not booting :( 2019-04-02 14:42:31 btw I used the official 3.9.2 files for booting 2019-04-02 14:50:11 ive read some ppl complain about booting 2019-04-02 14:50:13 @clandmeter alpine-rpi-3.9.0 80mb , alpine-rpi-3.9.2 45mb - at a quick look I would say there are no RPi2/3 bootfiles anymore :) 2019-04-02 14:50:20 i wonder if the bootloader is wrong 2019-04-02 14:50:25 which arch did you use? 2019-04-02 14:50:47 http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/armhf/alpine-rpi-3.9.2-armhf.tar.gz 2019-04-02 14:50:53 http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/armhf/alpine-rpi-3.9.0-armhf.tar.gz 2019-04-02 14:51:46 i would go for armv7 or aarch64 2019-04-02 14:52:01 for rpi 2/3 ? 2019-04-02 14:52:10 for 2 no 2019-04-02 14:52:33 i think 2 is armv6 with hardfloat 2019-04-02 14:53:44 im not sure we will support armhf for new releases. 2019-04-02 14:54:41 hmm, i think 2 should boot armv7, not sure though. 2019-04-02 14:54:58 LBlaboon: https://alpinelinux.org/sponsors/ :) 2019-04-02 14:55:17 looking good 2019-04-02 14:55:28 aarch64 is useless at rpi atm, armv6 is rpi1, v7 rpi2 and v8 rpi3 2019-04-02 14:55:33 oops 2019-04-02 14:55:37 this was the wrong channel 2019-04-02 14:55:45 oh well 2019-04-02 14:55:58 thanks danieli! \o/ 2019-04-02 14:56:06 I can use v7 as well, good pointer if you drop it in the future :9 2019-04-02 14:56:17 thank YOU \o/ 2019-04-02 14:56:24 aarch64 useless? 2019-04-02 14:57:12 no benefits, not at active developement 2019-04-02 14:58:27 it runs in 64bit mode, so i think you can find some benefits, but its probably not much. 2019-04-02 14:58:59 me (I am a LibreELEC dev) was told from an RPi dev thats basically not worth the hassle 2019-04-02 14:59:26 from video point of view it has no benefitrs 2019-04-02 14:59:31 only drawbacks 2019-04-02 14:59:48 btw "releases/armhf/alpine-rpi-3.9.0-armhf.tar.gz" works, "releases/armhf/alpine-rpi-3.9.2-armhf.tar.gz" does not boot at RPi3 2019-04-02 15:00:05 thats good to know 2019-04-02 15:00:11 can you report it on bugs.a.o? 2019-04-02 15:00:16 i think there is an issue already 2019-04-02 15:00:23 its not just the the video, its overall 2019-04-02 15:00:35 ofc no harm in offering an aarch64 2019-04-02 15:01:53 i never did any tests, so i dont know. 2019-04-02 15:02:20 but it sounds sane to have a kernel and userspace targeted at what the CPU offers. 2019-04-02 15:03:01 but, i did mention it so you dont run out of support. 2019-04-02 15:03:29 i guess you dont like upgrading 100+ rpi's when we discontinue it. 2019-04-02 15:03:36 :D 2019-04-02 15:04:05 I have currently 20+ versions of some raspbian floating around 2019-04-02 15:04:26 im going to add rpi netboot support 2019-04-02 15:04:38 so that means next version will have netboot images on our mirror 2019-04-02 15:05:00 v3.9/releases/armv7/alpine-rpi-3.9.2-armv7.tar.gz works at RPi3 , so just 3.9.1 and 3.9.2 are broken for armhf 2019-04-02 15:05:50 %10155 2019-04-02 15:05:54 oh 2019-04-02 15:05:57 #10155 2019-04-02 15:06:44 asgoo: is it the bootloader or the kernel? 2019-04-02 15:07:28 can't tell, no screen at all and the serial connection stays silent 2019-04-02 15:07:48 would be nice if you could find out. 2019-04-02 15:07:59 just replace the kernel initramfs 2019-04-02 15:08:07 and keeping the bootloader 2019-04-02 15:08:19 ah sry ... 2019-04-02 15:08:24 the files are just missing 2019-04-02 15:08:24 then ill make sure its fixed on next release 2019-04-02 15:08:33 files? 2019-04-02 15:09:03 at 3.9.0 you have boot/config-rpi boot/config-rpi2 .... at 3.9.2 just boot/config-rpi ... 2019-04-02 15:09:50 huh? 2019-04-02 15:10:11 can you do a tree paste to some service? 2019-04-02 15:10:21 oh ill just dl the release :) 2019-04-02 15:10:52 ah i see it 2019-04-02 15:11:04 the rlesaes are suddenly half the size 2019-04-02 15:11:16 i think ncopa tried to minimize it even more ;-) 2019-04-02 15:11:44 :D 2019-04-02 15:12:51 Do scripts in /etc/sysctl.d get automatically applied ? 2019-04-02 15:13:42 there is a 00-alpine.conf script there but if i check some of the values it should have set they are not there 2019-04-02 15:14:04 kernel.panic is set to 120 but if i type sysctl kernel.panic its set to 0 2019-04-02 15:14:53 asgoo: but i do see a kernel 2019-04-02 15:15:24 alpine-rpi-3.9.2-armhf.tar.gz right? 2019-04-02 15:16:25 yes 2019-04-02 15:16:43 looks like i have a kernel and initramfs 2019-04-02 15:16:48 in boot 2019-04-02 15:18:05 idk what actually changed I just saw that all rpi2 files are missing from 3.9.0 -> 3.9.2 2019-04-02 15:18:14 at a fresh install 2019-04-02 15:18:20 update worked 2019-04-02 15:18:30 yes, i think we ship a single kernel 2019-04-02 15:18:39 i remember some change 2019-04-02 15:19:05 ¯\_(ツ)_/¯ 2019-04-02 15:19:09 ah so it tries to boot the rpi2 kernel 2019-04-02 15:19:14 which is not available 2019-04-02 15:19:22 ncopa: ping 2019-04-02 15:23:08 asgoo: can you try the armv7 image? 2019-04-02 15:23:18 @clandmeter tx for help, looking forward for network at initramfs :P (trying tomorrow to build my own again) 2019-04-02 15:23:22 v7 is working 2019-04-02 15:23:28 right 2019-04-02 15:23:29 at least for my rpi3 2019-04-02 15:23:38 currently no 1-2 for testing 2019-04-02 15:25:29 i will fix it. 2019-04-02 15:25:33 not sure when though 2019-04-02 15:25:40 :) tx 2019-04-02 15:25:40 and we have to wait for release 2019-04-02 16:19:00 hi there :D 2019-04-02 16:20:46 is there a dead easy tutorial/distribution to use Alpine as my router firmware? 2019-04-02 16:21:08 something simple as ipfire and openwrt :D 2019-04-02 16:22:29 ShalokShalom: we don't have releases for musl, so beware 2019-04-02 16:22:41 oho.. ok 2019-04-02 16:22:49 whats up? 2019-04-02 16:22:50 it'll need a fair amount of tinkering to have it run on various embedded devices 2019-04-02 16:22:59 uh.. what do you mean? 2019-04-02 16:23:05 thought you are always on musl? 2019-04-02 16:23:25 musl is the libc, yes? 2019-04-02 16:23:33 I'm confused 2019-04-02 16:24:44 https://en.wikipedia.org/wiki/Alpine_Linux 2019-04-02 16:25:01 I've read the page before, but I'm not sure what you mean 2019-04-02 16:25:19 oh! I wrote the wrong thing 2019-04-02 16:25:22 <_ikke_> :D 2019-04-02 16:25:28 I meant we don't have releases for mips 2019-04-02 16:25:38 it was a brainfart, my apologies :) 2019-04-02 16:26:15 I was working on making something work with musl and I slipped up 2019-04-02 16:26:19 i see 2019-04-02 16:26:22 thy 2019-04-02 16:26:24 arm 2019-04-02 16:26:37 is my router 2019-04-02 16:27:14 you could try running alpine on it with u-boot, but you will probably need to at the very least adjust some things to make it work 2019-04-02 16:27:30 <_ikke_> Maybe something like this? https://wiki.alpinelinux.org/wiki/Linux_Router_with_VPN_on_a_Raspberry_Pi 2019-04-02 16:27:40 copying parts of the openwrt kernel config for that target could help 2019-04-02 16:28:08 IPQ4019 2019-04-02 16:28:15 thx, yeah i already saw that 2019-04-02 16:28:29 seems a bit confusing, why is there no simple script? 2019-04-02 16:29:12 alpine comes with a pretty basic setup, you'll have to configure it to your liking once you get it running 2019-04-02 16:29:47 clandmeter: pong 2019-04-02 16:29:49 https://wikidevi.com/wiki/AVM_FRITZ!Box_7530 2019-04-02 16:29:55 i see 2019-04-02 16:29:57 thy 2019-04-02 16:30:30 we don't have any builds specifically tested on / targeted at routers 2019-04-02 16:30:59 you're probably better off using openwrt if you want it to be easy to install, use, and set up, where just about everything works out of the box 2019-04-02 16:31:16 <_ikke_> and have a web interface for configuring 2019-04-02 16:31:29 that too, LuCI is neat 2019-04-02 16:32:37 openwrt is too boring :P 2019-04-02 16:32:56 want something lean, flexible 2019-04-02 16:33:07 <_ikke_> Then you need to put in some effort yourself 2019-04-02 16:33:20 hnn 2019-04-02 16:33:22 ok, thanks 2019-04-02 16:33:33 how to redistribute Alpine? 2019-04-02 16:33:35 well, the tradeoff is that we don't target or tes on router devices 2019-04-02 16:33:39 what do you mean? 2019-04-02 16:33:54 make my own spin 2019-04-02 16:34:18 so I can share my own work with the community 2019-04-02 16:34:31 the very first thing would be to change the name and branding 2019-04-02 16:34:39 is that process fluid? 2019-04-02 16:34:45 ofc 2019-04-02 16:35:11 <_ikke_> It would be easier to add something to Alpine itself 2019-04-02 16:35:16 honestly, not really, we have a lot of infrastructure for CI/CD and various other things 2019-04-02 16:35:21 it would be easier to contribute upstream to us 2019-04-02 16:36:00 in which form, in this case? 2019-04-02 16:36:07 script? 2019-04-02 16:36:10 eh? 2019-04-02 16:36:10 <_ikke_> depends on what it is 2019-04-02 16:36:14 <_ikke_> It could be an aport 2019-04-02 16:36:18 i prefer an own ISO 2019-04-02 16:36:19 <_ikke_> (a package) 2019-04-02 16:36:26 hnn 2019-04-02 16:36:36 ok i will see 2019-04-02 16:36:36 <_ikke_> There are scripts to build your own iso 2019-04-02 16:36:37 thy 2019-04-02 16:36:51 if you want an ISO of your custom work, you could indeed use scripts for that 2019-04-02 16:38:10 is Valery Kartel (php7-pecl-imagick maintainer) here? 2019-04-02 16:41:32 perfect 2019-04-02 16:41:37 thanks a lot 2019-04-02 16:52:47 ncopa: rpi armhf is missing a kernel 2019-04-02 16:52:53 I'll fix it 2019-04-02 23:04:04 I cannot boot a pvh domU from an iso, what am I missing? 2019-04-02 23:21:52 what is the fully sanctioned way to autostar wireguard on boot? I can't find a wireguard-openrc package or the like 2019-04-02 23:27:43 Xe: wiki https://wiki.alpinelinux.org/wiki/Configure_a_Wireguard_interface_(wg) 2019-04-02 23:27:47 you can do it from the interfaces file 2019-04-02 23:32:18 I had some time /etc/local.d/wireguard.start script, before I switched to /etc/network/interface 2019-04-02 23:34:05 maldridge: how do you do it from the interfaces file? 2019-04-02 23:34:55 with PostUp commands 2019-04-02 23:36:25 Xe: did you looked to wiki url I posted? If you put it there it will be started at boot automatically 2019-04-02 23:42:29 mps: ah i see 2019-04-02 23:42:45 i have multiple routes, do I need to add all those in postup commands manually? 2019-04-02 23:43:27 you still need /etc/wireguard/wg0.conf 2019-04-02 23:44:37 you can add more routes with post-up in interface file 2019-04-02 23:45:37 post-up ip route add x.x.x.x/24 dev wg0 2019-04-02 23:48:13 late is here, good night all 2019-04-03 06:57:11 hi, I am again, "sh mkimage.sh --tag 3.9 \ --outdir /home/build/builds \ --arch armhf \ --repository http://dl-cdn.alpinelinux.org/alpine/v3.9/main \ --profile rpi" results in 2019-04-03 06:57:12 WARNING: no repository set 2019-04-03 06:57:32 any ideas ? apk update/upgrade etc was done 2019-04-03 07:43:47 https://pastebin.com/HS5VsW1L (log with some output) <- WARNING: no repository set, ERROR: unsatisfiable constraints - any ideas ? 2019-04-03 08:38:56 --tag is correct? 2019-04-03 08:39:04 JOIN 2019-04-03 08:39:23 3.9 i guess ? tried with 3.9.2, v3.9.2 etc too 2019-04-03 08:39:29 v3.9? 2019-04-03 08:39:33 not sure though 2019-04-03 08:41:24 tried it at master with "edge" too 2019-04-03 08:43:27 asgoo: try running it with sh -x and see where it fails 2019-04-03 08:45:57 asgoo: is that rpi running armhf? 2019-04-03 08:46:25 any reason you still want to use armhf? 2019-04-03 08:46:29 https://pastebin.com/EbCGycHi 2019-04-03 08:46:54 rpi is running armv7, just tried if armhf is working 2019-04-03 08:47:05 and copied "wrong" cmd 2019-04-03 08:47:17 ./mkimage.sh --profile netboot --outdir /home/clandmeter/netboot --repository http://dl-cdn.alpinelinux.org/alpine/v3.9/main/ 2019-04-03 08:47:20 this works for me 2019-04-03 08:47:32 sec i try 2019-04-03 08:49:10 damn it ... the \ before --repository .... 2019-04-03 08:49:11 Do scripts in /etc/sysctl.d get automatically applied ? there is a 00-alpine.conf script there but if i check some of the values it should have set they are not there 2019-04-03 08:50:57 adaro: it should, if you didn't add something in /etc/sysctl.conf and forgot 'source' there 2019-04-03 08:51:06 adaro: what does rc-status boot say? 2019-04-03 08:51:29 it should mention sysctl 2019-04-03 08:52:59 @clandmeter works now :D so the howto at the wiki made it worse xD 2019-04-03 08:53:01 iirc, start sysctl is added during install 2019-04-03 08:53:04 buuuut "ERROR: Failed to create lib/libdevmapper.so.1.02: No space left on device" 2019-04-03 08:53:20 tmpfs is full 2019-04-03 08:53:23 yep 2019-04-03 08:53:29 any idea how to expand it ? 2019-04-03 08:53:32 you need to mount it to another place 2019-04-03 08:53:53 should I symlink python to python3 executable? Because if You install python3, there is no python alias created ... 2019-04-03 08:55:13 @clandmeter if you say mount it somewhere else, currently its mounted at / - so ... 2019-04-03 08:55:41 thats not whatr i mean 2019-04-03 08:55:54 it tries to write to some dir 2019-04-03 08:56:02 mount that dir somwhere else, sd probably 2019-04-03 08:56:12 maybe its tmp dir or dest dir 2019-04-03 08:56:42 ah the "--outdir /home/clandmeter/netboot" you mean ? 2019-04-03 08:57:06 yes 2019-04-03 08:57:14 not sure thats the problem 2019-04-03 08:57:19 :D 2019-04-03 08:57:20 i think its tmp thats filling up 2019-04-03 08:57:25 but i could be wrong 2019-04-03 08:57:27 yep its full 2019-04-03 08:57:37 lbu commit and reboot ? 2019-04-03 08:57:42 say what? 2019-04-03 08:57:55 i think you can mount it somewhere else? 2019-04-03 08:58:04 create a dir on sd 2019-04-03 08:58:06 yea i try it first 2019-04-03 08:58:11 and bind mount it 2019-04-03 08:58:26 its slow, but should work 2019-04-03 08:58:32 or use some usb stick 2019-04-03 09:00:32 mps: the sysctl.conf file just has one entry that says # content of this file will override /etc/sysctl.d/* 2019-04-03 09:01:14 clandmeter: i don't think it comes with rc-status by default ? i dont seem to have it on the path 2019-04-03 09:02:03 adaro: then it should work without problem, I mean files from /etc/sysctl.d/ if sysctl is started during boot 2019-04-03 09:03:08 mps: what is responsible for loading it ? 2019-04-03 09:03:29 install scripts will enable it 2019-04-03 09:03:44 so only if you have some special setup it could be missing 2019-04-03 09:04:08 what do you mean by install scripts ? 2019-04-03 09:06:06 @clandmeter nope ... :( 2019-04-03 09:06:25 I guess I should buy the 2gb RPi :D 2019-04-03 09:06:27 adaro: how did you install alpine? 2019-04-03 09:06:40 asgoo: nope what? 2019-04-03 09:07:01 mounted usb stick to my build dir -> still out of storage 2019-04-03 09:07:02 clandmeter: i use a docker image that extends from alpine:3.9 2019-04-03 09:07:16 tmpfs is at 100% 2019-04-03 09:07:28 asgoo: did you mount /tmp somewhere else? 2019-04-03 09:07:48 and make sure its cleaned up 2019-04-03 09:07:49 and use https://github.com/krallin/tini for init 2019-04-03 09:08:05 adaro: i dont think that works on docker? 2019-04-03 09:08:44 clandmeter: what do you mean ? 2019-04-03 09:08:57 can you manually set systctl settings? 2019-04-03 09:09:35 @clandmeter there is no tmp just tmpfs at / 2019-04-03 09:09:52 sigh 2019-04-03 09:09:57 I can't simply unmount it i guess 2019-04-03 09:10:20 asgoo: you system has a /tmp directory? 2019-04-03 09:10:23 your... 2019-04-03 09:10:29 no 2019-04-03 09:10:31 ah 2019-04-03 09:10:32 sec 2019-04-03 09:10:34 no? 2019-04-03 09:10:40 $ sysctl -p /etc/sysctl.d/00-alpine.conf sysctl: error setting key 'net.ipv4.tcp_syncookies': Read-only file system 2019-04-03 09:10:42 hmmm 2019-04-03 09:10:47 sry brainbug 2019-04-03 09:10:55 adaro: you need to do that from the host. 2019-04-03 09:11:07 its a container which is restricted 2019-04-03 09:11:25 right 2019-04-03 09:14:21 clandmeter: so how does this work, does it automatically have the settings of the host system ? 2019-04-03 09:14:51 adaro: probably, you will need to read the docker documentation. 2019-04-03 09:15:11 i think i'll have to :-) 2019-04-03 09:15:19 not that familiar with it 2019-04-03 09:15:29 and we are not #docker :p 2019-04-03 09:16:20 but we are handsome, just to compensate. 2019-04-03 09:24:09 @clandmeter looks working 2019-04-03 09:30:49 @clandmeter Signing: /tmp/update-kernel.KiJJAl/boot/modloop-rpi2 Can't open for reading, No such file or directory 2019-04-03 09:31:14 "unable to load key file" 2019-04-03 09:31:45 does your system have sign keys? 2019-04-03 09:31:50 abuild keys 2019-04-03 09:34:15 hmm I redone it and now it ask for installing it there -.- ... not sure why this wasn't working, maybe because user build was not at the sudoer file before 2019-04-03 09:39:41 user abuild, not build 2019-04-03 09:40:11 and user abuild should not need to be in sudo group 2019-04-03 09:40:33 your user needs to be in abuild group 2019-04-03 09:49:33 the setup script will create its own repo to be included in the release 2019-04-03 09:49:38 so it needs to sign the new apkindex 2019-04-03 09:49:52 that need to be done with your own abuilld key 2019-04-03 09:50:06 and the pubkey will be included in the release 2019-04-03 10:19:29 idk tbh :) followed https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage and it wasn't working 2019-04-03 10:24:42 @clandmeter I saw your last commits to alpine/aports and did you missed adding "network" here https://github.com/alpinelinux/aports/blob/master/scripts/mkimg.arm.sh#L82 or does it work without it too ? 2019-04-03 10:25:15 asgoo: if we have netboot you dont need those 2019-04-03 10:25:48 but i guess we can clean it up a bit. 2019-04-03 10:25:58 but that needs more deeping thinking. 2019-04-03 10:54:53 hey all, does busybox/ash allow for ANSI colouring of the shell prompt? 2019-04-03 10:56:01 from my experiements it seems it doesn't. but then I may be doing it all wrong. Again. :) 2019-04-03 10:57:16 silverwillow: I have it set up, but I didn't do anything specific to get it 2019-04-03 10:57:57 hmm okay. i must be doing something wrong then. will try again. thnx. 2019-04-03 10:59:37 what's your $TERM set to? 2019-04-03 11:01:06 'linux' 2019-04-03 11:03:09 ok - my bad it works fine. 2019-04-03 12:48:14 anyone can tell me how the fedora build packaging system is called, i.e. fedora 'APKBUILD' 2019-04-03 13:01:41 rpm something 2019-04-03 13:02:05 i think their build server system is called koji or similar 2019-04-03 13:02:36 their build scripts are *.spec files 2019-04-03 13:02:46 rpm is resulting pkg, 2019-04-03 13:02:47 which IMHO is pretty nice 2019-04-03 13:03:04 i look here how they do things: https://src.fedoraproject.org/cgit/rpms/ 2019-04-03 13:03:48 searched for it but all results points to stack{overflow,exchange} and similar 2019-04-03 13:04:10 for example, if i want look how they build boost, i just append "boost" to the above url: https://src.fedoraproject.org/cgit/rpms/boost 2019-04-03 13:04:15 and cgit will filter 2019-04-03 13:04:20 cgit is nice 2019-04-03 13:05:24 yes, really, better than github/gitlab, imo 2019-04-03 13:06:01 ok, thanks. found how they are packaging vim 2019-04-03 13:06:05 cgit is sort of readonly 2019-04-03 13:06:26 so we can still keep a cgit interface if we switch to gitlab 2019-04-03 13:07:53 most of distro have vim-basic or vim-minimal, and vim-nox or similar, vim-gtk or gvim-gtk, not unique naming scheme between distros but similar 2019-04-03 13:08:26 and, vim-common 2019-04-03 13:09:10 some have every plugin and syntax files in subpackages, which is a mess imo 2019-04-03 13:12:17 i guess vim-common and gvim subpackage makes sense 2019-04-03 13:13:41 agree, but what you think about splitting vim and gvim in separate aports 2019-04-03 13:14:22 just thinking about it, not a proposal 2019-04-03 13:16:26 dont like that. will increase maintenance burden 2019-04-03 13:16:40 and will make it messy to have a vim-common 2019-04-03 13:16:59 and thinking about name it gvim-gtk3 in case someone post request to have gvim-gnome or gvim-gtk2 (mate people) 2019-04-03 13:17:31 vim-gtk3 may work 2019-04-03 13:17:37 or just "gvim" 2019-04-03 13:18:25 for now gvim also ok in hope vim-gnome will not arrive 2019-04-03 13:18:44 i wouldnt worry about that 2019-04-03 13:19:29 understand, it will arrive sooner or later :) 2019-04-03 13:19:51 and then we have to fix it 2019-04-03 13:21:57 btw, I will remove vimdiff subpkg, in hope Alpine 'power users' will know that they have to install diffutils if they need vimdiff 2019-04-03 13:55:38 Hi, I'm currently trying to setup an alpine VM with the setup-alpine utility. When using a config file for setting up the system, the setup fails as the mirror list cannot be fetched. This is due to the random number generator not ready yet. The rng is needed by the script as it fetches the mirror list via https... Is this a known problem or intended behavior? 2019-04-03 13:57:39 fin_ger: https://bugs.alpinelinux.org/issues/9960 2019-04-03 13:59:08 it is not intended but a result of the kernel developers decision to move that problem to distributors 2019-04-03 13:59:53 hmm is there a possibility to just wait until the rng is ready? It does not matter for me how long this takes. I just need a stable VM setup... 2019-04-03 14:02:18 you can use virtio-rng kernel module which should solve your problem in VM, or add haveged pkg and run it early in boot 2019-04-03 14:03:23 run haveged in sysinit runlevel 2019-04-03 14:03:59 mps: modprobe virtio-rng ? 2019-04-03 14:05:41 yes, but better is to put it in /etc/modules to be loaded on boot 2019-04-03 14:06:38 mps: as I'm booting the installation ISO, putting it into /etc/modules will not work 2019-04-03 14:07:08 mps: it's the installation that is failing for me. Also adding new packages will not work, as I don't have access to the apk mirrors 2019-04-03 14:09:39 oh, sorry, then yes, by hand 2019-04-03 14:10:45 btw, you can 'remaster' install media with config and packages you need if you will use this media more than one time 2019-04-03 14:11:45 there is wiki page about that, https://wiki.alpinelinux.org/wiki/Main_Page 2019-04-03 14:16:29 mps: I just tested doing a `find /` before running `setup-alpine`... This increased the entropy of /dev/random enough to fetch the mirror list... I'm still not quite satisfied with this solution 😅 2019-04-03 14:17:22 btw the VM is not security critical. It is for local testing only 2019-04-03 14:22:58 the problem is discussed for years on different mailing lists, irc channes, web forums and where not, but without conclusion or best practice agreements 2019-04-03 14:23:55 at the end it is not technical problem but social one, i.e. problem of trust 2019-04-03 15:30:25 I really like our git.a.o and its cgit interface. I use it all the time. I visit github interface for pr + issue 2019-04-03 15:30:43 git.a.o simple, and very fast. 2019-04-03 15:35:41 cgit is SOOOO much better than github's NIH'd crap 2019-04-03 20:14:48 Hello, when I run `apk upgrade`, nothing happens (no upgrade) but the output show `18 errors`. How can I find details about those errors? 2019-04-03 20:15:21 try: apk fix 2019-04-03 20:15:41 Oh ! Thank you ! 2019-04-03 20:16:39 I now remember that I ran a `apk upgrade` when I chrooted in the fs, and couldn’t run post-upgrade scripts… (noob cross-arch chroot here) 2019-04-03 20:16:52 I always love how apk is soooo quick 2019-04-03 22:08:56 dalias: is musl explicit_bzero incompatible with bsd implementation 2019-04-03 22:09:51 mps, why? 2019-04-03 22:16:28 dalias: sorry, connection was lost to freenode. if you answered my question please could you repeat 2019-04-03 22:16:58 i asked "why?" - what do you mean by "incompatible with bsd implementation" ? 2019-04-03 22:18:30 aha, in musl /usr/include/string.h it is defined as 'void explicit_bzero (void *, size_t);' 2019-04-03 22:19:04 in iwd source, it is 'static void explicit_bzero(void *s, size_t n)' 2019-04-03 22:20:03 here is problematic file from iwd source http://tpaste.us/roJ1 2019-04-03 22:23:59 why is HAVE_EXPLICIT_BZERO not getting detected ? 2019-04-03 22:24:42 that source file is supposed to be providing its own definition if the system doesn't have one 2019-04-03 22:26:00 true, this file doesn't have #include so I presume it is included from other file 2019-04-03 22:26:44 which file I should add to it in #include to detect explicit_bzero in musl 2019-04-03 22:33:59 if I comment out this function in their missing.h iwd builds and passes tests, but not sure if this is correct solution 2019-04-03 22:36:36 mps, HAVE_* macros probably come from autoconf or similar 2019-04-03 22:37:10 my guess is they detect it but failed to ensure their config.h was included here 2019-04-03 22:38:04 configure.ac have 'AC_CHECK_FUNCS(explicit_bzero)' 2019-04-03 22:38:59 config.h have '#define HAVE_EXPLICIT_BZERO 1' 2019-04-03 22:39:20 does adding #include "config.h" fix it? 2019-04-03 22:39:37 i think this is just a simple oversight 2019-04-03 22:41:12 yes, looks like. will talk with iwd developers tomorrow, although they told me that they build it in glibc without problem 2019-04-03 22:42:56 I asked you because wanted to know if it could be musl incompatibility, I have seen that explicit_bzere relatively new in musl 2019-04-03 22:43:40 about half year, or little more 2019-04-03 22:51:24 not sure if glibc has explicit_bzero yet (at least versions they tested with) or if so what profiles it's exposed under 2019-04-03 22:51:28 dalias: yes, simply adding '#include "../config.h"' in missing.h solved the problem 2019-04-03 22:51:41 but this looks like clearly a bug where they're not using the result of their configure test 2019-04-03 22:52:16 will talk with them tomorrow about this 2019-04-03 22:52:32 thank you for help 2019-04-03 23:42:56 > I always love how apk is soooo quick 2019-04-03 23:42:57 same, makes me wish gentoo's package manager wasn't written in python 2019-04-04 02:40:54 I don't think portage being python is its biggest problem... ;D 2019-04-04 08:29:30 perhaps not, I never messed with the source code 2019-04-04 08:58:57 Hi, I can build the kernel in Alpine using abuild -kR (https://wiki.alpinelinux.org/wiki/Custom_Kernel), but what's the recommended way to rebuild after a change without rebuilding the entire tree? 2019-04-04 09:41:14 dalias: I talked with IWD developers and they confirmed that it is a bug because they didn't included config.h in all .c files. thanks again for help 2019-04-04 10:14:42 there is option to copy pkg.apk-new config over old one, but can't remember. any hint 2019-04-04 13:20:49 is Dan Theisen (current tcsh maintainer) here or some alpine channels? 2019-04-04 14:49:05 hmmm, musl implements NLS, so I can't understand the purpose of libintl? 2019-04-04 14:54:35 markand, it's a legacy thing that needs some work still to rip out iirc 2019-04-04 14:55:10 okay, because I've seen in the efibootmgr alpine pacakge that a patch explicitly links to libintl 2019-04-04 14:55:47 but IMHO I think gettext things should not be in a C library :P 2019-04-04 14:57:47 but it looks like all C libraries implement gettext (except *BSD) 2019-04-04 14:58:00 i'm sympathetic but i'd rather have 3k in libc than 300k of some buggy gnu crap :) 2019-04-04 14:58:44 I say that because I usually disable NLS everywhere so this does not count for me but I agree with you in that case :) 2019-04-04 14:59:09 i might not have done it if i realized there was going to be significant gettext code on top of the core mo file lookup code which is trivial and already used by libc (~350 bytes) 2019-04-04 14:59:43 (resolving paths from selected locale, etc.) 2019-04-04 15:00:12 but overall i think it's a better outcome 2019-04-04 15:00:37 it's the localization framework all real-world translators want to use 2019-04-04 15:00:51 so effectively it's a dependency of basically everything 2019-04-04 15:00:55 yes, on a desktop it's pretty impossible to get rid of it 2019-04-04 15:00:59 and having a separare lib for that is just inefficient 2019-04-04 15:01:08 some packages are unconditionally configurable 2019-04-04 18:00:30 I need to doo a uefi install to nvme ssd of alpine linux ... I am booted ito a pxe image of 3.8 but am lost from here 2019-04-04 18:00:39 can anyone guide mme? 2019-04-04 18:00:43 me 2019-04-04 18:45:15 vizo: did you run setup-alpine? 2019-04-04 19:00:36 clandmeter: yes I did 2019-04-04 19:00:55 and it didnt work? 2019-04-04 19:01:11 it install but is not a efi install it seems 2019-04-04 19:01:27 but I could be missing something never used efi befor 2019-04-04 19:02:27 i guess your system didnt boot up in efi 2019-04-04 19:02:37 but you can force the installer to use efi 2019-04-04 19:02:52 oh how? 2019-04-04 19:03:04 it doesnt mention it in the help? 2019-04-04 19:03:34 I didnt see it and I looked before I cam to irc 2019-04-04 19:05:08 you need setup-disk 2019-04-04 19:06:55 i htink you can do USE_EFI=1 setup-alpine 2019-04-04 19:10:47 will try thanks 2019-04-04 19:56:45 I wonder why scaleway doesn't provide the Alpine image for it's ARM servers ? it seems they run Cavium ThunderX ARMv8... I can't find relevant info, if anyone as a clue or a link ? 2019-04-04 20:02:07 I have found some info here: https://dev.alpinelinux.org/~clandmeter/other/forum.alpinelinux.org/forum/general-discussion/alpine-linux-scaleway-arm64-vps-armv8-ssd-cloud-servers.html 2019-04-04 20:59:35 clandmeter: yes !!! that worked 2019-04-04 22:22:49 Procedural question: GitHub seems to have Pull Request activity, but https://wiki.alpinelinux.org/wiki/Creating_patches says to send patches. 2019-04-04 22:23:00 Which should I use for a package update? 2019-04-04 22:23:06 gdh: either one works :) 2019-04-04 22:23:24 GitHub PR or a patchwork patch (email -> alpine-aports@lists.alpinelinux.org) 2019-04-04 22:23:44 danieli, perfect. Thank you. :) 2019-04-05 00:10:24 Sorry for asking, but why didn't uswgi-php made it into Alpine? https://bugs.alpinelinux.org/issues/6602 2019-04-05 00:19:22 looks like it's because of some dependency issues 2019-04-05 00:19:40 I'll have a look tomorrow 2019-04-05 00:20:03 thx 2019-04-05 00:22:41 Hi all. I have what seems like a noob question but can't seem to find it addressed on the wiki. If you want to run Alpine in diskless or data mode and you use the installation CD as the boot medium, does it have any means of automatic discovery and loading of an lbu apkovl or a /var partition, or do you have to bring them up manually? 2019-04-05 00:51:08 hmm how do I add auser in alpine? 2019-04-05 00:52:14 coming from gentoo so the commands I know are not working 2019-04-05 00:52:55 adduser? 2019-04-05 00:53:20 you are correct sire 2019-04-05 04:09:15 Separate question - I installed the zfs package, and because it still didn't come up on startup, I added zfs to /etc/modules. The module is loaded on startup but the zpools I imported do not automatically import on the next boot... 2019-04-05 04:09:52 But they can be manually imported with no problem. 2019-04-05 04:10:54 I read something that suggested that this might be because the block devices are not guaranteed to come up at any particular time or in any particular order, but I'm not certain this is what's happening. Anyone have any suggestions how I could address this? 2019-04-05 08:30:56 Has been wireguard removed? https://pastebin.com/zeukXp1W 2019-04-05 08:32:37 <_ikke_> MICROburst: It's still in testing, so you need to make sure you have the testing repo on edge 2019-04-05 08:33:05 OK. thanks for the hint... 2019-04-05 08:33:27 <_ikke_> https://pkgs.alpinelinux.org/packages?name=wireguard-tools&branch=edge&arch=x86_64 2019-04-05 08:38:33 @clandmeter hi, me again :) is it intention that mesa-dri-vc4 is just available for armhf and not for armv7 ? so basically I can't use opengl at RPi, right ? 2019-04-05 08:38:59 hi 2019-04-05 08:39:07 i dont think its intentional 2019-04-05 08:39:16 which aport is that? 2019-04-05 08:39:28 or is that the actual aport name? 2019-04-05 08:40:17 https://pkgs.alpinelinux.org/package/edge/main/armhf/mesa-dri-vc4 2019-04-05 08:40:35 arch="all" 2019-04-05 08:41:04 ah 2019-04-05 08:41:21 good catch :) 2019-04-05 08:41:29 will fix asap 2019-04-05 08:41:30 just trying to use it ;) 2019-04-05 08:41:43 https://git.alpinelinux.org/aports/tree/main/mesa/APKBUILD#n87 2019-04-05 08:42:30 Tashtari: if you boot from cdrom as boot medium, it will look for apkovl, on all found block devices it can mount 2019-04-05 08:42:34 is there a timeframe when those changes are available at the repo ? (idk how long you take from commit -> repo) 2019-04-05 08:43:15 Tashtari: you need zfs modules on the boot cdrom which means that you need to use the extended iso 2019-04-05 08:44:10 btw netboot is still not working, it is not even booting with my selfmade files, sadly not time yet to find the problem :D 2019-04-05 08:45:22 also do you know if I can start chromium without X ? 2019-04-05 08:45:28 also do you know if I can start chromium without X at Rpi? 2019-04-05 08:45:46 i think we have a headless chromium 2019-04-05 08:45:53 uh ;) slack habbits 2019-04-05 08:46:14 if I start it it tells no display found 2019-04-05 08:46:28 if I instll xbase it works with xinit chrome .... 2019-04-05 08:47:03 ofc I like to avoid X 2019-04-05 08:48:44 "xrandr --listmonitors" -> "Can't open display" 2019-04-05 09:04:37 asgoo: fix is pushed, you need to wait till build finishes and uploaded to mirrors. 2019-04-05 09:05:00 oh so basically ~1-2h or so then its public ? :) 2019-04-05 09:05:06 3.9 ? 2019-04-05 09:05:14 no its in edge 2019-04-05 09:05:24 i think we can backport it 2019-04-05 13:24:58 ncopa: I see, thanks for that information. Regarding ZFS, though, this is with a sys installation, I installed zfs and added it to /etc/modules and it isn't importing pools on startup... 2019-04-05 13:44:20 Tashtari: do you have your root fs on zfs? 2019-04-05 13:44:39 you need load it from initramfs if you do 2019-04-05 13:44:41 No, root fs is on ext4. 2019-04-05 16:39:17 hello, where does one report issues with alpine-linux, the hybridiso images, to be more precise? 2019-04-05 16:41:09 namely, by writing those images to usb sticks, i've wrecked two of them (all with alcor controllers), because apparently they trigger the write protect in the controller somehow. I had to use the manufacturers software to reflash the controllers, it was only available in chinese, and only worked on XP. (but at least my drives work again). ... 2019-04-05 16:41:59 (to be more precise, no linux tool works with the device once the alpine iso has been written with dd (hdparm,fdisk,dd, etc, all report errors)) 2019-04-05 16:43:14 i assume it's due to the iso9660 filesystem, but i'm not sure 2019-04-05 17:05:32 Hi there. Did anyone had time to figure out, why uswgi-php didn't make it into Alpine? https://bugs.alpinelinux.org/issues/6602 2019-04-05 19:25:08 howdy folks, it appears that confluent has done a big 1.0 release of librdkafka and confluent-kafka-python. however librdkafka 1.0 isn't available in the edge repositories yet. is there a recommended pattern for downloading and building from source within a dockerfile? 2019-04-05 22:23:37 Hi, i am trying to get kivy to work on alpine and raspberry pi - for now it seems i run into memory trouble when i try to compile the kivy :( 2019-04-05 22:23:53 follows some of this: https://kivy.org/doc/stable/installation/installation-rpi.html#kivypie-distribution 2019-04-05 22:24:11 anyone have tryed kivy on this system? 2019-04-06 07:57:31 Hi, I am trying to compile Kivy on raspberry, but i runs out of ram - what is the best temporary way to extend the space during compile, and afterwards still have a RAM system= 2019-04-06 08:01:02 use some zram for swap and if that's still not enough use disk based swap. not much else you can do 2019-04-06 08:01:22 well execept for cross-compiling 2019-04-06 08:02:37 i think i am not expert enough for cross compiling :( 2019-04-06 08:03:15 i guess try the other options first ^^ 2019-04-06 08:06:06 if zram based swap is not enough, i'd suggest attaching some hard disk or ssd via usb and use that for swap instead of the sd card. 2019-04-06 08:09:07 Do you know a guide for zram? - sorry i am quite new :( 2019-04-06 08:41:50 got some 511mb swap free - lets hope that makes it work :) 2019-04-06 09:19:11 reppy: https://wiki.gentoo.org/wiki/Zram just load the module, set the device size, create swap, enable swap 2019-04-06 09:41:08 I found this site: http://it-offshore.co.uk/component/tags/tag/3-alpine-linux 2019-04-06 09:41:23 trying to compile kivy now - fingers crossed 2019-04-06 09:41:54 and hopefully the graphic back-end will not give to big problems 2019-04-06 12:28:00 Anyone has some clues about the window provider for kivy? 2019-04-06 12:28:20 seems egl_rpi or pygame or x11 can be used somehow 2019-04-06 16:01:55 Hi there. Did anyone had time to figure out, why uswgi-php didn't make it into Alpine? https://bugs.alpinelinux.org/issues/6602 2019-04-06 16:03:56 <_ikke_> Not sure, but this seems to be the biggest roadblock: "php is in community and uwsgi is in main." 2019-04-06 16:11:58 how to save a setup there has been installed by PIP, between reboots? 2019-04-06 16:12:15 (module) 2019-04-06 16:54:31 _ikke_: But I mean, where's the problem? Is it a license problem? 2019-04-06 16:54:47 _ikke_: I'm not a native alpine user... just using it from time to time. 2019-04-06 16:54:56 I mean, can't we push unity-php to community? 2019-04-06 16:55:17 unit-php* 2019-04-06 17:01:19 <_ikke_> velix_home: It's more a policy issue. uwsgi is in main. Packages in main can only depend on packages in main (main in alpine are the core packages) 2019-04-06 17:01:59 _ikke_: oh dear. worse than debian :D 2019-04-06 19:45:01 Okay, so I got my ZFS issue resolved by adding `rc-update add zfs-import boot` and `rc-update add zfs-mount boot`. 2019-04-06 19:45:34 One followup question, though, after installing the ZFS package, should I be loading the kernel module by adding it to /etc/modules, or should I be doing that some other way? 2019-04-06 19:48:57 Ah, it seems like zfs-import will modprobe the module itself if it isn't done already. 2019-04-07 03:55:29 hello 2019-04-07 03:56:56 never used IRC before dont know what im doing 2019-04-07 03:58:54 anyways if people can see this how come lbu is not saving state of packages in live-boot? 2019-04-07 04:00:02 and I get a kernal panic when I try to install from cache 2019-04-07 11:10:02 polkit is a incredible piece of bloat 2019-04-07 11:17:45 true, I built slim to not have polkit on my X machines 2019-04-07 11:20:33 I knew there was a Javascript engine built in but could not believe it was spidermonkey 2019-04-07 11:21:37 yeah, javascript for basic utilities is insane, imo 2019-04-07 11:24:30 ACTION wondering why forth vanished from programmers 'radar' 2019-04-07 12:10:59 are you saying that because FreeBSD's bootloader switches from forth to lua? :) 2019-04-07 12:28:27 markand: didn't know that FreeBSD uses forth bootloader, heh interesting will look at it. thanks for hint 2019-04-07 12:30:10 I worked with Sun workstation which used forth in their 'bios' (to call it as such) 2019-04-07 13:17:58 is there a way to list all packages that were installed from unofficial sources? at some point I built my own nfs utils and some dependencies to fix a bug, but now it crashes, I want to go back to the official build and I don't know which package is left over. 2019-04-07 13:29:56 jnt: apk fix -r reinstalls stuff. 2019-04-07 13:30:30 if you have removed the unofficial apks then it should be ok 2019-04-07 13:38:29 p4Wv1qn095FW: Okay, that just says "OK: 1608 MiB in 438 packages" So that means I already found all packages? 2019-04-07 13:38:46 Sadly exportfs still segfaults. 2019-04-07 13:39:36 jnt: apk fix -u 2019-04-07 13:39:57 and, -d maybe 2019-04-07 13:41:20 I always have to find right combination of parameters by try and error till get good result 2019-04-07 13:47:09 I did sudo apk fix -rd nfs-utils, it reinstalled 27 packages, but it still segfaults =( 2019-04-07 13:50:37 what says 'apk policy nfs-utils' 2019-04-07 13:55:26 nfs-utils policy: 2019-04-07 13:55:31 2.3.2-r1: 2019-04-07 13:55:36 lib/apk/db/installed 2019-04-07 13:55:42 etc/apk/cache 2019-04-07 13:55:49 http://dl-cdn.alpinelinux.org/alpine/edge/main 2019-04-07 14:00:57 this is from edge, maybe it is not tested enough 2019-04-07 14:19:47 https://www.openwall.com/lists/musl/2019/02/03/3 apparently it's this. 2019-04-07 14:24:51 nfs-utils has a fix on master, but no new release yet. 2019-04-07 15:02:10 Is there a version of alpine that would work with --> https://tinkerboarding.co.uk/wiki/index.php/Hardware ? 2019-04-07 15:14:10 Kingsy: thinkerboard is Asus with rk3288, i.e. arm 32bit? 2019-04-07 15:14:38 armv7 Alpine should work if you have bootloader for the board 2019-04-07 15:22:30 mps: bootloader for the board? 2019-04-07 15:22:39 this wouldf be something that does not come with alpine? 2019-04-07 15:23:52 I didn't seen bootloader for rk3288 in Alpine but I could be wrong 2019-04-07 15:25:42 bootloader could be easily built from u-boot source for most rk3288 boards 2019-04-07 15:28:46 that is over my head I am afraid 2019-04-07 15:32:03 Kingsy: you can take bootloader from some other distro, Raspbian for example if they support your board 2019-04-07 15:32:34 oh ok, tbh unless there is a guide I have absolutely no idea how I wouild do that. 2019-04-07 15:34:20 Is this a difficult task? 2019-04-07 15:37:24 well, if want to play with SBC's and ready to learn it is not difficult. Anyway if you plan to work with SBC's be prepared to learn a lot of new 'things' 2019-04-07 17:18:53 has anyone installed opennebula on alpine? 2019-04-07 18:05:40 hi, is there a way to get Ruby 2.6 on Alpine? 2019-04-07 18:07:24 <_ikke_> not at the moment, unless you build it yourself 2019-04-07 18:08:53 vizo: looks like jirutka has something related to opennebula on his github, you can check his user-aports to see if it's there 2019-04-07 18:09:11 it doesn't look like we have opennebula in our repos, and i haven't heard any mention of it 2019-04-07 18:10:52 _ikke_: I don't know if I'am wrong but, to compile different version of package is not sufficient to download the script to build apk a change version inside? (i never try) 2019-04-07 18:11:28 <_ikke_> vinniec: In theory, in practice, more could be involved 2019-04-07 18:11:35 unless we have some patches that were made obsolete, in which they have to be rebased or removed 2019-04-07 18:11:45 <_ikke_> and new patches may be required 2019-04-07 18:11:55 oky 2019-04-07 18:12:11 there could also be modules built against the older version of ruby that break 2019-04-07 18:13:42 132/5000let's say the script is a good base from which to compile an old package, if you're lucky it works without modification 2019-04-07 19:12:06 danieli: ah thanks thats for using docker with opennebula ... openebula has some ready made alpine linux images in their marketplace.I uwas suprised when it wasnt a apk 2019-04-07 19:12:27 its what led me totry alpine in the first place 2019-04-07 19:12:35 right, I'm not familiar with that, looks like there are no opennebula packages in our repositories for sure 2019-04-07 19:59:20 I am using Alpine Linux in a docker compose setup - how would I use apk to specify a php version - currently even after the upgrade from 3.8 to 3.9 still 7.2.14 is being added 2019-04-07 20:00:44 like so https://gist.github.com/inoas/a16fb7f608d96cc29e3412e16fed311e 2019-04-07 20:01:07 <_ikke_> What version do you expect? 2019-04-07 20:02:26 7.2.13 in 3.8, 7.2.14 in 3.9 2019-04-07 20:02:33 7.2.17 in edge 2019-04-07 20:02:39 <_ikke_> nod 2019-04-07 20:02:43 7.3 obviously 2019-04-07 20:02:54 <_ikke_> well then 2019-04-07 20:02:55 7.3.x - do not care which 2019-04-07 20:03:04 you'll have to compile it yourself, unfortunately :) 2019-04-07 20:03:05 <_ikke_> there is no 7.3 available yet 2019-04-07 20:03:37 Is there a roadmap for when it will be about to be added? 2019-04-07 20:03:51 and will you directly add ^7.3.4 then? 2019-04-07 20:03:51 whenever we add it to edge, it'll be in the subsequent releases 2019-04-07 20:04:48 edge is good enough for a docker box I suppose 2019-04-07 20:05:15 if you *must* have 7.3.x as soon as possible, you must compile it 2019-04-07 20:05:37 well no I am just updating my docker compose env and don't want to repeat doing that any time soon 2019-04-07 20:06:16 inoas: check the PR queue 2019-04-07 20:06:17 7.3 is out and stable for a while though I had problems with composer on osx, 7.4 is currently deemd the last 7.x release while 8.x is underway which will take some time I suppose 2019-04-07 20:06:22 i guess there are refs to 7.3 2019-04-07 20:06:32 you can monitor the progress 2019-04-07 20:06:54 about alpine in general, is it close to how debian works in terms of scripts and defaults or more redhat or something else? 2019-04-07 20:07:12 andypost does most of the php stuff 2019-04-07 20:07:30 PRs on which repo https://github.com/alpinelinux ? 2019-04-07 20:07:33 7.2 -> 7.3 will most likely require us to test and rebuild some packages that depend on it 2019-04-07 20:07:39 https://github.com/alpinelinux/aports 2019-04-07 20:08:50 I had no previous experience with alpine or docker and both worked very well together 2019-04-07 20:08:53 thank you :) 2019-04-07 20:09:07 but like danieli mentioned, you can always compile it yourself. 2019-04-07 20:09:39 many docker images based on alpine are custom compiles, ie nginx. 2019-04-07 20:14:10 custom compiles / nginx? yes I run an nginx a php-cli and a php-fpm image 2019-04-07 20:14:12 How are these applied? https://github.com/inoas/aports/blob/master/community/php7/APKBUILD#L183 2019-04-07 20:15:45 Be back later! - night train 2019-04-07 20:15:57 the version of that package seems to be 7.2.17, those comments just mean that those CVEs were patched in that version 2019-04-07 20:16:02 that was quick 2019-04-07 20:39:32 re 2019-04-08 01:18:54 Just did a sys install of Alpine, and for some reason, when sshd is in runlevel default, there is a several-minute-long pause in startup, after "Starting busybox crond ... [ ok ]" and before "Starting sshd ..." appears. When "Starting sshd ..." appears, it goes to "[ ok ]" basically immediately, though. 2019-04-08 01:19:18 But when I remove sshd from runlevel default, there's no pause, and if I immediately log in and rc-service sshd start, it starts immediately. 2019-04-08 01:22:21 your kernel lacks working entropy sources 2019-04-08 01:22:44 by the time you login there's a massive amount of entropy from whatever the source of your login was 2019-04-08 01:22:57 what hardware is this on? 2019-04-08 01:23:36 It's a Sandy Bridge i7. 2019-04-08 01:23:56 weird 2019-04-08 01:24:04 on the metal not virtualized or anything? 2019-04-08 01:24:11 On the metal, yeah. 2019-04-08 01:24:35 c0oo0 2019-04-08 13:02:01 anyone having issues with acme-client? 2019-04-08 13:02:37 its deprecated, so its bound to have issues. 2019-04-08 13:02:44 :( 2019-04-08 13:03:00 i liked it - it worked well for me :) 2019-04-08 13:03:13 most of liked it. 2019-04-08 13:03:49 what do you use as a replacement? 2019-04-08 13:03:55 complain at openbsd :) 2019-04-08 13:04:21 we have special setup here 2019-04-08 13:04:31 btw, im not saying acme-client is broken 2019-04-08 13:04:48 we use official client to generate wildcard cert 2019-04-08 13:05:17 and distribute that to which ever service needs it. 2019-04-08 13:05:52 i think an lightweight alternative is acme.sh (if thats the right name) 2019-04-08 13:06:37 recommended client is certbot, which is python :( 2019-04-08 13:06:48 https://letsencrypt.org/docs/client-options/ 2019-04-08 13:07:27 yeah, it's a bit clunky 2019-04-08 13:07:35 tons of features though 2019-04-08 13:07:42 and i'd say it's pretty well tested 2019-04-08 13:08:16 acme.sh is well maintained afaik 2019-04-08 13:08:38 yeah, i think so 2019-04-08 13:08:40 needs only posix shell and i think curl 2019-04-08 13:08:43 maybe even wget 2019-04-08 13:08:50 short link to repo is https://acme.sh/ btw :) 2019-04-08 13:10:06 acme.sh is quite nice. dehydrated is also a nice one 2019-04-08 13:10:36 looks like dehydrated is zsh-compatible bash 2019-04-08 13:10:38 I use acme.sh personally on a bunch of alpine VMs 2019-04-08 13:10:40 while acme.sh supports ash 2019-04-08 13:10:47 s/ash/other shells/ 2019-04-08 13:11:03 I particularly like the fact that acme.sh has a nice and easy DNS api extension mechanism 2019-04-08 13:12:10 i'll look at both those shell scripts 2019-04-08 13:12:17 thank you 2019-04-08 13:12:21 true, acme.sh need curl and socat is good to have 2019-04-08 13:12:34 acme.sh needs curl or wget 2019-04-08 13:12:46 but use curl, because they've had a few wget only bugs 2019-04-08 13:12:59 TomJepp: on Alpine it need curl 2019-04-08 13:13:20 mps: I install it straight from upstream instead 2019-04-08 13:13:56 also I did untill it is packaged for Alpine 2019-04-08 16:14:34 Hi all, the package amavisd-milter is currently in the testing repository. We have several productive mail servers using this package without any issue. Would it be possible to move this package to the main repo? 2019-04-08 17:05:56 @midasi: Do you mind openening an issue at https://bugs.alpinelinux.org/projects/alpine/issues ? 2019-04-08 17:07:15 This package currently has no maintainer 2019-04-08 17:08:10 sure, will do... thanks 2019-04-08 17:23:28 anyone running Alpine X window on ARM rockchip rk3399 board 2019-04-08 17:53:18 Hello! Im using alpine in a docker image and the applications requires /lib/lsb/init-functions. How can I wrap this in alpine? 2019-04-08 18:02:10 pqatsi: consider patching the software to using /etc/init.d/functions.sh instead 2019-04-08 19:04:24 How can I handle LD_PATH with Alpine? I'm trying to run https://pkgs.alpinelinux.org/contents?branch=edge&name=raspberrypi-libs&arch=armhf&repo=main but I see something is missing to binary reach libs in /opt/vc/lib 2019-04-08 19:17:39 Hi, I'm curious about networking support in extlinux. I can't seem to find much documentation, other than that relating to PXELINUX. 2019-04-08 19:18:34 I've seen "network" listed as a feature in mkinitfs, which adds PXE support but I can't tell if this forces PXE or would actually allow networking to take place using tools installed on the /boot volume 2019-04-08 19:21:09 specifically I'm trying to build an alpine system that accepts a key to decrypt the / volume over SSH, similar in function to Ubuntu's dropbear scripts. anyone messed around with this before? 2019-04-08 19:44:08 i just played around with getting networking to work in the initfs 2019-04-08 19:45:39 but i've used bios/uefi pxe boot function 2019-04-08 20:10:36 no static IP/route configuration, I suppose 2019-04-08 20:27:44 na with dhcp 2019-04-08 20:31:17 dhcp server provided ip and kernel+initfs, initfs had udev and dropped you to a shell with drivers loaded for your hardware 2019-04-08 20:31:29 something like that, can't remember all the details, was some weeks ago 2019-04-08 20:54:01 but i still have the scripts in case you might need and want it to achieve what you need or want, the initfs thingy was built off an alpine docker container, so everything what was in there would end up in the initfs iirc 2019-04-08 21:04:48 yeah, I'm interested. I still have a lot to learn about alpine's initfs 2019-04-08 21:05:21 so now that is really out 2019-04-08 21:09:49 it's alive muhahahaha http://p.xqz.ca/f89ca 2019-04-08 21:10:19 smooth upgrade from 3.9.2 2019-04-08 21:11:03 rpi booted up and upgraded 2019-04-08 21:12:26 hardware is an old OptiPlex GX520 :P 2019-04-08 21:16:07 ajin: that was quick :) 2019-04-08 22:23:11 is there an easy way to dual boot when windows 10 is already installed? I already have close to 500gb of free space. 2019-04-08 22:45:06 machman, on uefi? 2019-04-08 22:53:10 I am not too sure. I know I had to set it to legacy mode in bios so I could boot from usb. 2019-04-08 22:53:44 I just bought the laptop at a pawnshop a week ago, 2019-04-08 22:54:06 Does Alpine Linux do a rolling update? I.e., I did a "sys" install with release 3.9.2, if I do `apk update` and `apk upgrade`, will I have the equivalent of release 3.9.3? 2019-04-08 22:54:29 rickhg12hs: right 2019-04-08 22:54:45 Cool! 2019-04-08 22:55:25 even with main releases, i.e. 3.9 to 3.10 2019-04-08 22:56:15 Was syslinux upgraded? In 3.9.2, syslinux-6.04_pre1 has a compile issue that prevents hdt from working. 2019-04-08 22:56:45 hdt? what is it 2019-04-08 22:56:47 pretty sure syslinux-6.04_pre3 fixes it. 2019-04-08 22:57:29 New compilers "optimize" out `pow()` and replace with `exp()`. There's no `exp()`. 2019-04-08 22:58:16 https://www.syslinux.org/archives/2018-August/026159.html 2019-04-08 22:58:17 didn't know for this, never had any issue with syslinux 2019-04-08 22:59:12 https://wiki.syslinux.org/wiki/index.php?title=Hdt_(Hardware_Detection_Tool) 2019-04-08 23:01:00 fedora uses glibc so maybe this is not issue with musl 2019-04-08 23:01:48 anyway, thanks for info, good to know if this is appears on alpine 2019-04-08 23:02:11 Maybe it's a different cause, but I have the same symptoms. When I boot and select "Hardware Info", I get the "Undef Symbol FAIL: exp" message. 2019-04-08 23:03:37 aha, ok, this is interesting, but boot normal system works fine 2019-04-08 23:04:06 Yeah, normal boot works fine. 2019-04-08 23:04:07 nice thing it could be tested with qemu 2019-04-08 23:04:29 I'm using VirtualBox. :-) 2019-04-08 23:05:12 btw, you can use grub also if need/want 2019-04-08 23:05:49 True, and I might try grub too. Thought you'd want to know about the possible syslinux issue too. 2019-04-08 23:06:36 yes, if I find some time I will at that 2019-04-08 23:07:15 I have a lot of work last days, and little time for Alpine :( 2019-04-08 23:09:27 I'm wondering if all it takes is a variable change in the syslinux APKBUILD file. Currently it has "pkgver=6.04_pre1" 2019-04-08 23:10:50 don't have here aports so not sure, but probably answer is yes 2019-04-08 23:10:58 ... and the answer is NO. :-( 2019-04-08 23:11:18 will check this tomorrow 2019-04-08 23:11:25 Just checked the website the source is pulled from and the updates aren't there. :-( Weird. 2019-04-08 23:13:51 6.04 is not released as I see 2019-04-08 23:15:09 No, but Alpine is current using 6.04_pre1, there's a _pre3 now. 2019-04-08 23:16:30 do you have download url, I found only _pre1 2019-04-08 23:17:23 ah, ok found it 2019-04-08 23:21:22 https://repo.or.cz/syslinux.git/snapshot/458a54133ecdf1685c02294d812cb562fe7bf4c3.tar.gz 2019-04-08 23:21:29 ok, built apk of this new version, will try it tomorrow 2019-04-08 23:21:43 Awesome! Thanks! 2019-04-08 23:24:08 if it is not late here I would test it right now, but I had all day work to solve some problems which is made by some 'admins' on network providers for my customer 2019-04-08 23:25:34 I had a professor would taught networking technology, and he could not stop himself from swearing when "admins" were mentioned. :-) 2019-04-08 23:29:05 yeah, today people think they could learn networks by getting certificate from some companies academy (hey cisco, microsoft) 2019-04-08 23:42:19 cisco's is surprisingly general purpose 2019-04-08 23:55:39 We used Cisco's PacketTracer as part of a class. It was fun. 2019-04-09 00:12:21 I personally prefer gns3 2019-04-09 00:12:28 but yeah, I've touched by PT 2019-04-09 04:09:33 hello, I' 2019-04-09 04:11:02 I'm trying to boot alpine inside of a docker container, I've took the miniroot tar and imported it, now when I attach to it I get can't open /dev/tty[1-6] errors and it seems that no /sbin/openrc exists. how can I fix it? 2019-04-09 04:42:35 <_ikke_> dagg_: what do you mean with 'booting'? THe miniroot system is not setup to boot at all (and hence, does not contain openrc). Usually in docker a single process is started that can then if it wants start other processes. 2019-04-09 05:07:15 dagg_: try lxc/lxd for the kind of experience you describe... there are alpine images for lxc already 2019-04-09 05:13:07 mmmm, I want to migrate a alpine vm to container, initially I thougt of lxc but as my system is intel based, I was recommended to install clearlinux and run containers, am I missing something? 2019-04-09 05:14:03 docker containers aren't a 1:1 for VMs as much as lxc is 2019-04-09 05:15:17 so I'd better go with lxc 2019-04-09 05:15:23 but clearlinux+docker has the same kind of isolation as a VM 2019-04-09 05:15:33 i.e. separate kernels 2019-04-09 05:17:27 I see, thanks for the info. I'll try google. 2019-04-09 09:34:38 lolz 2019-04-09 09:34:56 From: 2019-04-09 09:34:59 I made a video showing how you satisfy yourself in the left half of the screen, and in the right half you see the video that you watched. 2019-04-09 09:35:01 With one click of the mouse, I can send this video to all your emails and contacts on social networks. 2019-04-09 09:35:03 I can also post access to all your e-mail correspondence and messengers that you use. 2019-04-09 09:35:19 (that old one) 2019-04-09 09:35:49 that spammer sure gets around 2019-04-09 09:38:19 tya99: this started to appears few month ago 2019-04-09 09:40:31 i saw a variant of it in the past that included a password, as I use unique passwords everywhere i figured out which site got hacked 2019-04-09 09:40:55 because it would have only been in that database, seems the spammer is too poor to buy database dumps now and just resorts to sending it to everyone without a password 2019-04-09 09:54:36 "you have a video of me compiling a kernel?" :) 2019-04-09 09:56:00 yeah man 2019-04-09 09:56:02 hot shit! 2019-04-09 09:56:24 koollman: he must have a video of me installing alpinelinux 2019-04-09 09:56:37 because that's what i am doing in the right half of my screen right now ;) 2019-04-09 09:57:23 > I can send this video to all your emails and contacts on social networks 2019-04-09 09:57:31 so he's going to send this video to all the email accounts I own? 2019-04-09 09:57:42 and I'm not on social networks so that's not going to be an issue 2019-04-09 10:28:00 hmm when i start X in my virtualbox of alpinelinux the mouse/keyboard freezes 2019-04-09 10:28:09 i forget the solution for that i have to install something 2019-04-09 10:29:03 apk add xorg-server i3wm dmenu dunst i3status@testing rxvt-unicode firefox-esr ttf-dejavu virtualbox-guest-additions virtualbox-guest-modules-vanilla sudo xf86-video-vesa xf86-input-libinput eudev 2019-04-09 10:29:07 i thought was all i needd 2019-04-09 10:30:17 apk add xf86-input-mouse xf86-input-keyboard 2019-04-09 10:30:21 hm that might be a good idea 2019-04-09 10:40:36 and kbd didn't help either 2019-04-09 10:45:06 I'm having some issues with updating alpine 2019-04-09 10:45:24 "Bad file descriptor", what can that come from? 2019-04-09 10:47:31 okay so gpm works, so it must something X related 2019-04-09 11:36:33 tya99: is eudev started? There is a script setup-xorg-base or similar that can replace the mdev with eudev for you 2019-04-09 12:11:03 oh yeah that's what might be the issue :) 2019-04-09 12:11:24 hmm yeah i didn't run setup-xosrg 2019-04-09 12:11:38 i did install eudev though lol 2019-04-09 12:12:25 ncopa: thanks for reminding me, most of the time i never use X on alpinelinux :) 2019-04-09 12:33:18 does anyone else have the issue where they are used to typing "systemctl start " and then when they use openrc they're trying to do "rc-service start service" instead of "rc-service service start" 2019-04-09 12:33:48 yes, once. 2019-04-09 13:18:12 I have that the other way round; the openrc way is the way it's been ever since systemd, but systemd naturally turned the command word and the target around 2019-04-09 13:18:26 because why not :D 2019-04-09 13:21:37 because we can :D 2019-04-09 13:34:27 hmm it's not possible to use virtualbox-guest-modules-vanilla in diskless mode is it 2019-04-09 13:34:48 asit wants to write, /lib/modules/4.19.34-0-vanilla/misc/vboxvideo.ko, /lib/modules/4.19.34-0-vanilla/misc/vboxguest.ko, /lib/modules/4.19.34-0-vanilla/misc/vboxsf.ko and that's read-only 2019-04-09 13:35:06 is there a workaround for that? 2019-04-09 13:38:55 by the sounds of it that will only work in sys mode 2019-04-09 13:49:59 For some reason I can't get mongodb to accept connections (installed via apk) 2019-04-09 13:56:36 After install, there is no mongodb.conf in /etc. When I create one, it is ignored 2019-04-09 14:56:57 i'm trying to get xe-guest-utilities installed using `apk add xe-guest-utilities` but i'm getting the error that it is missing. feeling a little confused, trying to get an Alpine template set up on my xen host 2019-04-09 14:59:02 desnudopenguino: which architecture are you on? 2019-04-09 14:59:24 x86_64 2019-04-09 14:59:56 do you have the community repository enabled in /etc/apk/repositories? 2019-04-09 15:00:13 https://pkgs.alpinelinux.org/packages?name=xe-guest-utilities&branch=edge&arch=x86_64 2019-04-09 15:00:14 it *is* there 2019-04-09 15:00:30 i see that the package is on the edge branch, though i'm not necessarily sure what that means, if i have to do anything with apk to use it? 2019-04-09 15:00:43 it's in earlier versions as well 2019-04-09 15:01:10 it used to be in main, but is now in community by the looks of it 2019-04-09 15:04:04 hmm. ok, so i just looked up how to add the community repository, if i do that and update apk, it should be good to go i'm assuming 2019-04-09 15:04:52 it should be, yes :) 2019-04-09 15:05:08 just add an entry identical to the 'main' one in /etc/apk/repositories with 'main' replaced with 'community' 2019-04-09 15:05:29 i.e. duplicate the line and replace "main" -> "community" 2019-04-09 15:05:50 yeah, there's one in there commented out already. so just uncommented and updated and it installed. 2019-04-09 15:05:57 that works :) 2019-04-09 15:09:38 i wonder how hard it would be to set up an alpine xen host... though i do like out-of-the-box tools, because i actually have to be productive during work haha 2019-04-09 15:13:03 desnudopenguino: trivial -- there's a pre-configured Xen host on the official downloads page 2019-04-09 15:13:14 I just got through setting it up myself 2019-04-09 15:14:02 oh nice! 2019-04-09 15:16:10 bit of a learning curve for me but that's more becos I don't know Xen well, as yet. 2019-04-09 15:21:26 i'm running xcp-ng currently but the Alpine ecosystem is growing on me. so i may give it a shot. probably virtualizing another xen host to start. lol 2019-04-09 15:23:09 if the alpine dom0 has the proper api stuff, you might be able to hook it up through xen orchestra or xen center/xcp center 2019-04-09 15:24:27 Alpine has gotten a lot of love recently due to the Docker community 2019-04-09 15:24:45 I've been using it for about 5 years now, even pre-Docker 2019-04-09 15:26:31 Not sure what you mean by "hook it up"? I'm relatively new to Xen -- forgive me if I'm asking about the obvious 2019-04-09 15:39:38 plsuh: xen center and xen orchestra are citrix GUI tools to interact with the xen api on the host. they make things easier if you don't want to type xen commands all day 2019-04-09 15:41:33 looks doable -- https://github.com/yobasystems/alpine-xen-orchestra/blob/master/alpine-xen-orchestra-amd64/Dockerfile 2019-04-09 15:41:47 nice! 2019-04-09 15:42:52 While I've got you -- is there a standard way to share a directory from the dom0 to a domU? There are tantalizing hints about p9 in the xl.cfg man page but no one seems to actually do it 2019-04-09 15:43:26 you'd need a shared filesystem 2019-04-09 15:43:38 i'm not sure xenfs got anywhere 2019-04-09 15:44:04 it'd probably be easier to go for NFSv4 2019-04-09 15:49:28 plsuh: i'm not really sure an nfs share would probably be my best guess. besides my man xen orchestra guest, all my other vms are on my NAS 2019-04-09 15:49:45 Grr. thx, at least now I know not to chase that black hole 2019-04-09 15:50:34 there might be a way, but it is probably a black hole. you could set up a "disk" and might be able to mount it between guests, though maybe not at the same time, via xen 2019-04-09 15:50:50 sounds like the best alternative 2019-04-09 16:24:31 i just spent about 20 mins setting up a part of the xen guest utils script to recognize alpine as the guest os. i was following this guide: https://www.linkedin.com/pulse/setting-up-alpine-linux-370-domu-vm-xenserver-72-ali-poursamadi/ but it seems that an alpine identifier wasn't in the code anymore. maybe i'll have a commit to make to the xen utils project. yay 2019-04-09 16:28:00 cool 2019-04-09 17:18:42 I have a strange problem with alpine in docker. I can telnet from a vanilla ubuntu container to an external service, but I cannot telnet from a vanilla alpine container to the external service. The alpine container just times out. Are there any differences with the networking that would cause this to happen? 2019-04-09 17:25:04 looks like there are already two PRs to add an identifier for alpine to xen. someone beat me to it 2019-04-09 17:25:36 jeffu: can you ping out from the alpine container? 2019-04-09 17:28:53 hey. how do I get a list of the command line options for /usr/bin/iconv? 2019-04-09 17:29:05 it doesn't have --help nor -? nor a man page, it seems 2019-04-09 17:29:22 and the options don't correspond to the ones i'm used to from other operating systems 2019-04-09 17:29:35 desnudopenguino: nope, can't ping 2019-04-09 17:32:31 bolt: it's different, it's from musl and not glibc 2019-04-09 17:32:35 let me see what i can find 2019-04-09 17:33:14 I can't ping from the ubuntu container either though. I can telnet from the ubuntu container 2019-04-09 17:33:51 i'm trying to translate from IBM850, which seems to be Windows' default encoding for telnet sessions 2019-04-09 17:34:25 I found I can use -f instead of --from-code. IBM850 seems to be missing, however. 2019-04-09 17:34:51 ... windows? it's just ascii though? 2019-04-09 17:36:09 the input originates from a Windows box, yes. IBM850 charset (telnet). it has norwegian characters like ÆØÅ, which I'm trying to get right. doing some processing with Alpine in a docker container 2019-04-09 17:36:36 hmm, såpass 2019-04-09 17:36:49 :D 2019-04-09 17:37:03 I don't really understand where you got IBM850 from, to be honest 2019-04-09 17:37:41 either way, doesn't look like musl's iconv supports that anyway 2019-04-09 17:39:37 oh right, you probably looked at `chcp`'s output? 2019-04-09 17:41:39 I don't actually remember where I got that from. I just know it works for converting Windows telnet stuff into UTF-8. It seems to be the default charset used by the Windows telnet client, at least in Western Europe. It says it's the default encoding for Western Europe DOS here: https://en.wikipedia.org/wiki/Code_page_850 2019-04-09 17:41:59 CP850, a.k.a. Latin-1, yeah 2019-04-09 17:43:04 ah. it does recognize CP850 2019-04-09 17:43:05 neat 2019-04-09 17:43:19 I'll give that a shot and see if my data turns out to be readable :) 2019-04-09 17:43:29 considered running `chcp 65001` for utf-8? 2019-04-09 17:43:54 messing with locales and encoding on musl is *very* painful if you're doing more than scratching the surface 2019-04-09 17:44:34 So is changing anything about the source of this data. I'd rather not poke that hornets' nest. 2019-04-09 17:44:47 fair enough, point taken 2019-04-09 17:56:26 sweet. it works! thanks 2019-04-09 17:56:36 on to the next problem :D 2019-04-09 21:06:00 hi i have after update problem to start my containers 2019-04-09 21:07:01 https://pastebin.com/k8CRq3Vn 2019-04-09 21:08:24 https://wiki.alpinelinux.org/wiki/LXC#Upgrading_from_2.x 2019-04-09 21:22:43 @clandmeter not helping 2019-04-09 21:23:06 Tuxist: what is your /proc/cmdline? 2019-04-09 21:23:07 i upgrading from 3.9.0 to 3.9.3 2019-04-09 21:23:19 ah 2019-04-09 21:23:29 unpriv containers? 2019-04-09 21:23:41 https://pastebin.com/VtQCbLu1 2019-04-09 21:23:43 yes 2019-04-09 21:24:00 cgroup_enable=memory 2019-04-09 21:24:04 i think you need to remove that 2019-04-09 21:24:46 its on the wiki: Make sure you have removed cgroup_enable from your cmdline as this will fail to mount cgroups and fail LXC service. 2019-04-09 21:25:41 i will try it 2019-04-09 21:26:32 nope not helping 2019-04-09 21:27:34 my mounting table https://pastebin.com/gz0awRSd 2019-04-09 21:27:55 did you start cgroups service? 2019-04-09 21:31:00 yes 2019-04-09 21:31:20 what is the error now? 2019-04-09 21:32:03 https://pastebin.com/fc12BnXB 2019-04-09 21:32:04 same 2019-04-09 21:33:42 possible error with systemd ? 2019-04-09 21:34:10 alpine host with debian container 2019-04-09 21:34:26 not sure 2019-04-09 21:34:41 so your cmdline is updated now and cgroup is start correctly? 2019-04-09 21:35:13 yes and cgroup is startet double checked 2019-04-09 21:36:23 https://pastebin.com/jdvWtxAg 2019-04-09 21:37:28 what does mount say 2019-04-09 21:37:54 https://pastebin.com/atHXtWjD 2019-04-09 21:39:03 that does look like correct cgroup mounts 2019-04-09 21:39:14 did you modify something cgroup related? 2019-04-09 21:39:21 the initd or rc.conf? 2019-04-09 21:39:39 did you update your /etc directory? 2019-04-09 21:39:47 update-conf 2019-04-09 21:39:59 nu 2019-04-09 21:40:09 nu ja 2019-04-09 21:41:08 no 2019-04-09 21:41:13 sorry 2019-04-09 21:41:25 i have run update conf and reboot 2019-04-09 21:42:31 no change same problem as before 2019-04-09 21:43:10 yes i'am german 2019-04-09 21:44:44 yes me not 2019-04-09 21:45:02 so there are no more configs to update 2019-04-09 21:45:09 and you still have same mount output? 2019-04-09 21:45:28 yes possible there some changes kernel config ? 2019-04-09 21:45:58 https://tpaste.us/lZPw 2019-04-09 21:46:09 thats my mount output grepped for cgroup 2019-04-09 21:46:44 what is securityfs ? 2019-04-09 21:50:23 what you see is the default mounts from cgroup service from openrc. 2019-04-09 21:52:00 ok 2019-04-09 21:52:39 i have read there was problem lxd 3.10 and musl is the fix already included in alpine ? 2019-04-09 21:57:47 thats gives me lxc-checkconfig https://pastebin.com/ahyimw0g 2019-04-09 21:58:16 did you modprobe config? 2019-04-09 21:58:23 configs 2019-04-09 21:59:08 ok now it runnig 2019-04-09 21:59:29 but the problem is probably not lxc related 2019-04-09 21:59:45 https://pastebin.com/S3t8JKDn 2019-04-09 21:59:49 are we talking about lxd or lxc here? 2019-04-09 22:00:00 lxc 2019-04-09 22:05:35 Tuxist: sorry i dont know whats wrong. you will have to fiddle a bit yourself. 2019-04-09 22:05:44 maybe SpaceToast has an idea 2019-04-09 22:06:07 oh, hello, I see I have been summoned 2019-04-09 22:07:13 can I get a recap? :) 2019-04-09 22:07:44 recap ? 2019-04-09 22:07:48 of the situation 2019-04-09 22:08:00 I wasn't *actually* here this whole time, I was detached (znc feature) 2019-04-09 22:08:01 ok 2019-04-09 22:08:40 i have update my server from 3.9.0 to 3.9.3 after that my container won't start 2019-04-09 22:09:05 ok, are you using privileged or unprivileged containers? 2019-04-09 22:09:06 https://pastebin.com/k8CRq3Vn 2019-04-09 22:09:14 unprivileged 2019-04-09 22:09:34 what user are you spawning your containers as? 2019-04-09 22:09:43 root 2019-04-09 22:12:03 how are you setting up your cgroups? 2019-04-09 22:12:39 i use thte default config + systemd 2019-04-09 22:13:17 the default config being /etc/init.d/cgroups, that you can't easily modify to get systemd? :) 2019-04-09 22:13:22 I'm asking for the details, not a summary 2019-04-09 22:14:23 i have modified th fstab 2019-04-09 22:14:31 uh 2019-04-09 22:14:54 https://pastebin.com/JKyLS7j3 2019-04-09 22:15:36 yeah that's not really how you do cgroups 2019-04-09 22:15:54 i'am not using with alpine systemd but the containers use it with debian etc ... 2019-04-09 22:16:08 ok 2019-04-09 22:16:15 https://brpaste.xyz/VUqnAw try using this service file instead 2019-04-09 22:16:17 (cgroups-systemd) 2019-04-09 22:16:22 and get the thing out of your fstab 2019-04-09 22:16:30 add it to the same runlevel as cgroups 2019-04-09 22:17:20 notice how your mount paste doesn't have "systemd" in it anywhere as well 2019-04-09 22:17:39 the way you should be doing cgroups is defaults in rc.conf, and cgroups (the service) in the boot runlevel 2019-04-09 22:17:43 (alongside the above if you need systemd) 2019-04-09 22:19:45 ok 2019-04-09 22:20:05 i have do that i will reboot now 2019-04-09 22:20:50 its not windows :) 2019-04-09 22:22:14 mhm now says Cgroup v1 systemd controller: missing 2019-04-09 22:22:52 are you sure the service is started? 2019-04-09 22:22:56 the one you got 2019-04-09 22:23:14 i have do that to check by hand 2019-04-09 22:23:25 ?? 2019-04-09 22:24:12 run /etc/init.d/cgroup-systemd start 2019-04-09 22:24:21 SpaceToast: doesnt openrc take care of those cgroups? 2019-04-09 22:24:30 clandmeter: no, this is one of the areas where openrc is stupid 2019-04-09 22:24:36 I keep wanting to send in patches 2019-04-09 22:24:43 but williamh is usually a pain to deal with 2019-04-09 22:24:46 so I still haven't ^^;; 2019-04-09 22:25:09 right 2019-04-09 22:26:03 but ok i have do that manuelly with the mount command 2019-04-09 22:26:20 but no change conatianer start 2019-04-09 22:26:42 ??? 2019-04-09 22:27:25 typingt mkdir -p /sys/fs/cgroup/systemd and mount -t cgroup -o none,name=systemd systemd /sys/fs/cgroup/systemd 2019-04-09 22:27:38 do you type that during the booting process? 2019-04-09 22:27:47 you know, when no shell is available? 2019-04-09 22:27:49 cgroups are finnicky 2019-04-09 22:27:59 please follow instructions, otherwise I can't really help 2019-04-09 22:28:01 (or even attempt to) 2019-04-09 22:28:05 ok 2019-04-09 22:30:39 i have change right now the init script do what we want 2019-04-09 22:30:55 ... 2019-04-09 22:31:15 sorry, I really can't help you if you're going to go around doing things that are different from my attempts to debug 2019-04-09 22:31:20 remote debugging is enough of a pain without that 2019-04-09 22:31:29 enjoy \o consider asking the lxc devs 2019-04-09 22:33:11 i don't understand you have say make this init scripts 2019-04-09 22:33:20 i have forgot to make runable 2019-04-09 22:36:16 i will follow your intructions now i promise it 2019-04-09 22:40:46 SpaceToast: I'am typing now oly what you say 2019-04-09 22:40:59 ? I had left 2019-04-09 22:41:26 let me see if my lxd setup dies on reboot, hold up 2019-04-09 22:41:52 ok 2019-04-09 22:42:21 i use only lxc withoud lxd 2019-04-09 22:43:57 lxd uses liblxc directly 2019-04-09 22:44:06 so if the issue is *actually* with a bump to some system package 2019-04-09 22:44:08 I would experience it 2019-04-09 22:44:17 which I did not :) 2019-04-09 22:44:47 in short, it means you definitely have a *configuration* issue 2019-04-09 22:45:00 please get me the outputs of the following commands: 2019-04-09 22:45:30 rc-update, rc-status -a, cat /etc/rc.conf 2019-04-09 22:48:24 https://pastebin.com/kKce0CWj 2019-04-09 22:48:42 please move cgroups-systemd to boot 2019-04-09 22:48:55 the place I originally asked you to put it 2019-04-09 22:49:00 https://pastebin.com/0KTxrscq 2019-04-09 22:49:12 rc-update del cgroups-systemd; rc-update add cgroups-systemd boot 2019-04-09 22:49:29 you also don't have cgroups currently up 2019-04-09 22:49:30 done 2019-04-09 22:49:33 so you'll need to reboot after that 2019-04-09 22:49:43 but first let's see rc.conf 2019-04-09 22:49:56 also, I did mean `rc-status -a`, not `rc-status`. 2019-04-09 22:51:04 in case you want to do it faster (e.g not having to copy paste a lot), consider using a curl-able pastebin 2019-04-09 22:51:35 for example, you can do 2019-04-09 22:51:47 https://paste.ee/p/XiyGV 2019-04-09 22:51:56 `rc-status -a | curl -#fF 'data=<-' brpaste.xyz` 2019-04-09 22:52:06 ok thanks 2019-04-09 22:52:12 this is after reboot, right? 2019-04-09 22:52:16 yes 2019-04-09 22:52:23 ok, what's in /etc/rc.conf? 2019-04-09 22:52:34 (you can just pastebin the whole file using the same method as above) 2019-04-09 22:55:51 https://tuxist.de/rc.conf 2019-04-09 22:56:13 I'd really rather just see the file in my browser... 2019-04-09 22:56:22 is it that hard to pipe it into curl??? 2019-04-09 22:57:43 "cat /etc/rc.conf | curl -#fF 'data=<-' brpaste.xyz" 2019-04-09 22:57:56 but i got no response 2019-04-09 22:58:02 is your prompt multiline? 2019-04-09 22:58:23 no 2019-04-09 22:58:49 https://github.com/5paceToast/brpaste/blob/master/client/brpaste.sh 2019-04-09 22:58:54 you can use this script if you're having trouble I guess 2019-04-09 23:08:06 https://brpaste.xyz/3ZrEhQ 2019-04-09 23:08:15 http no more working 2019-04-09 23:08:19 only https 2019-04-09 23:08:28 yes, it auto-redirects to https 2019-04-09 23:10:14 Tuxist: change rc.conf : rc_controller_cgroups should = YES 2019-04-09 23:12:38 done 2019-04-09 23:12:52 might need to reboot afterwards 2019-04-09 23:13:00 after reboot, try starting up your container again, and we'll re-read the error 2019-04-09 23:13:51 ok 2019-04-09 23:18:06 i need short reboot of the client vpn crashed ... :-( 2019-04-09 23:19:50 i'am back 2019-04-09 23:19:53 \o 2019-04-09 23:20:20 it's not my day 2019-04-09 23:23:06 https://brpaste.xyz/4EUa3g 2019-04-09 23:23:55 ok, now that that's more legible... 2019-04-09 23:24:29 Tuxist: in your container configuration 2019-04-09 23:24:31 lxc.cgroup.devices.allow = 2019-04-09 23:24:33 lxc.cgroup.devices.deny = 2019-04-09 23:24:36 (nothing on the right side) 2019-04-09 23:25:43 alternatively, we could also make the "devices" cgroup, but you're not planning on protecting yourself from things anyway :) 2019-04-09 23:28:02 i have do that but i got same error message 2019-04-09 23:28:22 is that so? lxc is misbehaving then 2019-04-09 23:28:26 run this, then 2019-04-09 23:28:48 https://brpaste.xyz/WwgXmg 2019-04-09 23:29:18 in the /usr/share/lxc/config/common.conf are devices 2019-04-09 23:29:25 mkdir /sys/fs/cgroup/devices; mount -t cgroup devices -o devices /sys/fs/cgroup/devices 2019-04-09 23:30:15 (and then try to start it again, and remove the devices stuff) 2019-04-09 23:31:02 https://brpaste.xyz/iUjXrw 2019-04-09 23:31:42 hmm 2019-04-09 23:31:44 what's the output of `mount`? 2019-04-09 23:33:26 https://brpaste.xyz/L8ZB8A 2019-04-09 23:33:48 um 2019-04-09 23:33:50 no 2019-04-09 23:33:52 I mean 2019-04-09 23:33:54 literally 2019-04-09 23:33:56 `mount` 2019-04-09 23:33:58 just that 2019-04-09 23:34:44 https://brpaste.xyz/tEnM0g 2019-04-09 23:35:28 ok 2019-04-09 23:35:31 this is a new lxc problem 2019-04-09 23:35:37 you have to report it to upstream :( 2019-04-09 23:35:53 basically, lxc explicitly expects the `devices` cgroup to be under /sys/fs/cgroup/devices 2019-04-09 23:35:59 but you have it under the unified /sys/fs/cgroup mount 2019-04-09 23:36:10 and we can't do much about it because openrc isn't very customizable right now 2019-04-09 23:36:31 alternatively, you can set up unprivileged containers 2019-04-09 23:36:32 after outcomment the devices the container will start 2019-04-09 23:36:42 ? 2019-04-09 23:36:50 ok, so the first workaround wasn't resetting it 2019-04-09 23:37:05 (I suspect it was order of operations on the part of bcak but that's ok) 2019-04-09 23:37:24 the short version is like so: 2019-04-09 23:37:37 if your container is privileged, it can make its own device nodes in /dev 2019-04-09 23:37:45 because root(0) == root in the container 2019-04-09 23:37:52 so lxc tries to use cgroups to stop that from happening 2019-04-09 23:37:56 (by default) 2019-04-09 23:38:08 but it hardcodes the devices cgroup location to /sys/fs/cgroup/devices (which is not what gentoo uses) 2019-04-09 23:38:17 (gentoo being the origination point of openrc) 2019-04-09 23:38:42 either way, are you satisfied with the result? 2019-04-09 23:40:30 my problem is that i need dev urandom random 2019-04-09 23:40:35 for gitlab 2019-04-09 23:40:58 um 2019-04-09 23:41:03 sure? 2019-04-09 23:42:04 ultimately, the devices cgroup handling is broken in lxc 2019-04-09 23:42:23 I suppose you could try getting a unified cgroup setup for it, or you could try using unprivileged containers 2019-04-09 23:42:46 ok 2019-04-09 23:43:13 i will no go to sleep big thanks for help 2019-04-09 23:47:33 i will write bug report 2019-04-10 00:07:46 hi all 2019-04-10 00:08:30 i picked up a raspberry pi zero today, but forgot to get a mini hdmi cable, so i am trying to figure out if i can go about this in a headless way 2019-04-10 00:08:45 has anyone run into this issue and found a resolution? 2019-04-10 00:18:55 should be ok if your boot setup enables serial console, no? 2019-04-10 00:19:02 not that i've done it 2019-04-10 00:30:21 not sure, this is a very minimal rpi 2019-04-10 00:30:33 two USBs, one mini hdmi, one microsd port 2019-04-10 00:32:09 oh wait, you mean via the headers 2019-04-10 00:32:15 i think the usb that provides power also provides serial port but i'm not sure 2019-04-10 00:32:27 there might be a header that does too 2019-04-10 00:32:42 i mean like /dev/ttyS0 or whatever :) 2019-04-10 00:32:50 not a usb device 2019-04-10 00:35:50 ah, that's somewhere to start at least, thanks! 2019-04-10 00:36:10 gotta say...i'm running alpine on my laptop and i'm loving it 2019-04-10 07:08:14 Hi there, does anyone know why $SHELL is /bin/ash even if i do apk add bash and run bash? 2019-04-10 07:21:38 natten: the bash manual page says "The full pathname to the shell is kept in this environment variable. If it is not set when the shell starts, bash assigns to it the full pathname of the current user's login shell." 2019-04-10 07:25:35 Ohh, i guess i can change the login shell then. 2019-04-10 07:25:35 Thanks :) 2019-04-10 12:30:36 `apk upgrade` returns `1 error; ...` How do I discover the error? 2019-04-10 12:31:26 <_ikke_> apk fix 2019-04-10 12:32:43 Ahhh, thanks .... something to work on now. Cheers! 2019-04-10 14:25:05 I have a vm running Alpine and nginx running. The directory /var/www/project has nginx:nginx as owner:group, but I keep getting permission errors in the nginx log. Any ideas? The user in the nginx conf is set to nginx 2019-04-10 14:31:10 Parsec300: see if `chmod -R 770 /var/www/project && chown -R nginx:nginx /var/www/project` helps 2019-04-10 14:31:43 it *could* also be something broken in your nginx config 2019-04-10 16:14:38 danieli, I will try that 2019-04-10 16:31:23 Parsec300: also make sure that the above directories are +x 2019-04-10 17:40:39 iggy, rgr 2019-04-10 19:38:44 iggy, danieli, it seems to work until I mount a shared folder to the location of that site, not matter what ownership I set it to 2019-04-10 19:42:53 So it seems like I need to find a better way to mount a shared directory 2019-04-10 19:43:16 I do my development on my mac and use Alpine only to act like my server, but then locally 2019-04-10 20:02:59 Parsec300: yeah, I've seen that, something weird with the osxfs thing... there are some issues open around the interwebs 2019-04-10 20:27:32 iggy, any other suggestions as an alternative, or just use VirtualBox? 2019-04-10 20:35:21 I stopped trying to pass things through or copied in (which worked for my use case, but doesn't sound like it would for you) 2019-04-10 20:35:34 maybe samba on the host or something? 2019-04-10 20:37:22 i do that with mount options like sshfs -o allow_other,uid=48,gid=48 2019-04-10 20:37:29 gotta go bye 2019-04-10 20:55:25 hello, I try to install Alpine on scaleway's ARM64 host using this tutorial: https://dev.alpinelinux.org/~clandmeter/other/forum.alpinelinux.org/forum/general-discussion/alpine-linux-scaleway-arm64-vps-armv8-ssd-cloud-servers.html 2019-04-10 20:55:53 But I cannot boot. if anyone has a clue... :-) 2019-04-10 20:56:21 I see some broken HTML in there, so I hope you didn't run those commands in verbatim 2019-04-10 20:56:30 I changed pathnames to 3.9 2019-04-10 20:57:10 no I didn't run verbatim, there is an edited version here also: https://community.scaleway.com/t/alpine-linux-on-arm64/6022 2019-04-10 20:57:24 and found the getty thing here: https://github.com/scaleway/image-alpine/blob/master/latest/overlay/usr/local/sbin/dynamic-getty 2019-04-10 20:58:27 I changed rc to openrc and here I am with this error: can't run '/usr/local/sbin/dynamic-getty': Exec format error 2019-04-10 21:14:38 hum... my dynamic-getty is full of HTML :-) 2019-04-10 21:19:03 iggy, fassl, thanks for the help. I used to use sshfs before. Should've thought of that ;) 2019-04-11 01:19:00 given an alpine ISO and an 8gb usb, can I create a bootable persistent usb with a sys mode install? I found the page on creating a bootable usb but I think that's a guide for diskless mode 2019-04-11 01:50:25 dieggsy, not sure about the sys mode install, but i have a persistent bootable usb 2019-04-11 02:07:05 maybe this can help you somehow, note that you would have to remove the additional fstab entries it creates when it detects installation to an usb drive for it to be able to boot. https://github.com/pfandl/alp-aut 2019-04-11 02:08:33 so make sure to do that before the make clean, also sometimes you have to manually mount the efi partition for it to succeeded, have no clue why 2019-04-11 02:14:42 fassl: like I'm wondering what happens if I use setup-alpine from a bootable usb and select a different USB as the disk to install to, then try to boot from that 2019-04-11 02:14:52 one way to find out I guess 2019-04-11 02:17:02 not sure about that, i installed it from an iso, and set every identifier to block ids, and remove the ones automatically generated, and it was fine 2019-04-11 05:29:15 dieggsy: that's how I installed mine (boot 1 usb, install to another) 2019-04-11 05:30:16 iggy: did you use sys? mine seemed to be hanging a while when generating the initramfs and I had a movie to watch so I didn't actually wait for it 2019-04-11 05:31:29 the details are a little fuzzy (like most of my life) 2019-04-11 06:51:47 hi, anyone an idea how the "fbsplash.ppm" works at the RPi ? I get it working but it ALWAYS show the splash and I never get any console or other output anymore 2019-04-11 07:15:29 iggy, danieli, I fixed my issue from yesterday. The hosting directory now is accessible by nginx by adding the '-o allow_other' option to the command 2019-04-11 07:22:52 Parsec300: how did you pass that through? 2019-04-11 08:37:38 to whoever maintains the alpine website: if you style letter colors in input fields, please also color the background. or don't color both and let browser apply system theme 2019-04-11 08:39:08 Piraty: which page, specifically? 2019-04-11 08:39:26 https://pkgs.alpinelinux.org/packages for example 2019-04-11 08:40:03 system theme shouldn't screw with css in web pages though? 2019-04-11 08:40:24 danieli: https://imgur.com/ELwReGM.png 2019-04-11 08:40:35 oh, ouch, I see what you mean 2019-04-11 08:40:40 input fields are subject theming in firefox 2019-04-11 08:41:15 not sure what's up with the arrow on those, can't remember which pseudo-elements control those 2019-04-11 08:41:22 but I could set the background of the package name input field 2019-04-11 08:41:27 i don't know in specific what else is too, but that's a common issue users have when using dark themes but webdevs assuming light themes (half-way at least) 2019-04-11 08:41:58 I guess the real question is why the hell firefox feels the need to change the background color of just input boxes. 2019-04-11 08:42:21 honestly have to agree, that seems pretty half-assed by FF, but adding a background color of the input is no big deal 2019-04-11 08:42:27 Piraty: that is issue with a lot of sites 2019-04-11 08:42:35 yes i know 2019-04-11 08:42:51 arch wiki even has a section in Firefox/ about it, that's how popular it is 2019-04-11 08:44:01 qutebrowser gets it right (by not using system theme i assume) and prints the field's bg white 2019-04-11 08:44:39 lol. luakit just does the oppotite of firefox 2019-04-11 08:45:28 root cause is in that the web devs thinks no one uses dark themes, or they simply don't care 2019-04-11 08:46:49 or unintentionally break it 2019-04-11 08:46:53 pkgs.a.o wasn't made by a web dev 2019-04-11 08:47:40 to be fair, dark themes are something that is both rare, and usually only done by power users. 2019-04-11 08:47:44 danieli: I didn't had Alpine in mind, that is a simply trend 2019-04-11 08:47:55 yeah, but I was just saying :) 2019-04-11 08:48:24 bluezinc: it's inherently prone to causing undefined behavior since it's an automatic third-party modification 2019-04-11 08:48:26 qutebrowser: https://imgur.com/hpUJVvt.png 2019-04-11 08:48:26 luakit: https://imgur.com/yolO8bw.png 2019-04-11 08:48:38 it looks even worse in luakit 2019-04-11 08:48:41 ofc, I understand you and sorry if you sounded offensive 2019-04-11 08:48:48 it inverses ff behaviour 2019-04-11 08:50:16 mps: i think you said that the wrong way around :) "sorry if you sounded offensive" 2019-04-11 08:51:10 danieli: oh, yes. looks like I'm not awaken yet, need second cup of coffee 2019-04-11 08:51:27 yeah no, me too, I just got home from a meeting 2019-04-11 08:51:38 didn't have time for coffee or breakfast, and I slept roughly two hours 2019-04-11 08:53:12 I had to work to 2:50 this night/morning and also slept a little, so better I should close my mouth now (fingers) 2019-04-11 08:54:06 don't worry, we understand what you meant :) 2019-04-11 08:55:35 btw, I solved this theme issues by using 'Dark Background and Light' add-on in firefox 2019-04-11 08:56:06 it's a shame it's necessary. I'll add a PR to change the bg color of the input anyway 2019-04-11 08:56:50 it can be set to use dark theme only for sites I'm selecting and for other their default css and 'colours' 2019-04-11 09:24:26 you can force firefox to use other themes https://wiki.archlinux.org/index.php/Firefox/Tweaks#Change_the_GTK_theme_for_content_process_only 2019-04-11 09:43:34 iggy, the command used that worked was: vmhgfs-fuse .host:/{share-name} /var/www/{webroot directory} -o uid=101 -o gid=102 -o allow_others 2019-04-11 10:17:17 @clandmeter hi, do you have a idea how the "fbsplash.ppm" works at the RPi ? I get it working but it ALWAYS show the splash and I never get any console or other output anymore 2019-04-11 11:25:41 Never used it 2019-04-11 12:24:12 tx 2019-04-11 12:49:23 hey... i am reactivating my old cubietrucks, i found a script that makes a nice 3.9.x alpine image for them right now 2019-04-11 12:49:53 i tried the repo search to find out if there's a docker package for armhf 2019-04-11 12:50:24 but generally found nothing. does one of you know if i can expect to run docker on armv7/armhf alpine? 2019-04-11 12:51:01 <_ikke_> https://pkgs.alpinelinux.org/package/edge/community/armhf/docker 2019-04-11 12:51:02 they got 2GB of ram so i _can_ also do a bit with Xen but not too much 2019-04-11 12:51:08 _ikke_: thank you 2019-04-11 12:51:18 it must have been something with the phone browser then 2019-04-11 12:51:25 because it came up empty 2019-04-11 12:51:30 :-) 2019-04-11 12:52:00 i'll see if i can also do some fun things like switching to LBU mode or boot from NAND 2019-04-11 12:52:04 if yes i'll post here 2019-04-11 12:53:14 root@svc-ubuntu16-561:~/xen-arm-builder# ls -l sdcard.img 2019-04-11 12:53:15 -rw-r--r-- 1 root root 17179869184 Apr 11 14:52 sdcard.img 2019-04-11 12:53:17 maybe 2019-04-11 12:53:22 we get lucky 2019-04-11 14:00:06 Hello. I'm having some issues with my alpine install. For every package i install i get an error saying "failed to execute: No such file or directory". 2019-04-11 14:00:10 what folder am i missing? 2019-04-11 14:00:28 <_ikke_> can you try apk fix? 2019-04-11 14:01:05 apk fix gives me several errors saying the same 2019-04-11 14:02:45 Which folder is "busybox-*.trigger" supposed to be in? is it in cache or? 2019-04-11 14:23:09 to the earlier question about installing alpine to a removable USB disk, no matter how you do the install make sure to define your mounts via UUID (not the name) or you'll have trouble booting when the drive order changes. should be a given but it's easy to forget 2019-04-11 18:21:56 Hi! Quick question: Is one supposed to install a minor release directly? Or install the major and update. I ask because the 3.9.3 armhf archive for raspberry pi contains only the kernel for raspberrypi2. So it is not bootable on a Raspberry Pi 1. On the other hand, 3.9.2 has the right kernel for RPi 1. Alpine is great, I'm a begginer (with Alpine, not with Linux), so maybe is something I overlooked. Thanks! 2019-04-11 18:27:18 DanAndresan: its a bug 2019-04-11 18:29:29 You have an rpi1 2019-04-11 18:29:54 ? 2019-04-11 18:30:43 clandmeter: yes, I do. An old one, but still working. I have also a 2 and a 3, but they are used in other projects. 2019-04-11 18:33:02 Next release should have it fixed 2019-04-11 18:36:00 Thank you for confirming my findings. Have a great day! 2019-04-11 18:37:56 good to know, will wait, too. :D 2019-04-11 23:55:31 i have installed docker on alpine linux, when i try to execute any command, e.g. "docker ps -a", i get: "Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?" 2019-04-11 23:55:38 if i run "sudo dockerd", then it blocks the shell at "API listen on /var/run/docker.sock" 2019-04-11 23:55:42 is there a way to run the dockerd in background? 2019-04-12 00:00:15 ah, nm.. had to run "sudo service docker start" :/ 2019-04-12 00:01:47 :-) 2019-04-12 00:24:23 the cubietruck thing is getting somewhere (https://github.com/mirage/xen-arm-builder/) 2019-04-12 00:24:55 i had to change initird-grsec to -vanilla in a few places, and maybe i'll do some polishing too 2019-04-12 00:25:00 but generally: wow 2019-04-12 00:26:17 booting into ram, having xen 4.12 and docker in the same little embedded box is pretty nice 2019-04-12 00:32:15 footnote, the slowest part of the process is the slightly worn sd cards 2019-04-12 06:40:03 darkfader: that is pretty cool! 2019-04-12 10:11:20 ncopa: it's definitely booting now, a few weird things like no way to have a console working 2019-04-12 10:12:10 i gotta remember how i can add my ssd as "data" disk in lbu mode now and then there'll be a /var/lib/docker and all good 2019-04-12 10:19:19 now i found i'm lacking cgroups in the kernel ;-) 2019-04-12 10:19:40 at least this a classic case where you can just fix a bit each day 2019-04-12 10:22:14 alpine domU really works 2019-04-12 10:22:19 Welcome to Alpine Linux 3.9 2019-04-12 10:22:19 Kernel 4.4.14 on an armv7l (/dev/hvc0) 2019-04-12 10:22:19 localhost login: 2019-04-12 10:22:27 enough for today ;-) 2019-04-12 10:23:38 darkfader: 4.4.14, are you using old release 2019-04-12 10:24:11 mps: i don't yet know why the kernel is stuck on that version 2019-04-12 10:24:46 i set the config to use alpine 3.9.3, but they fetch their own kernel tree for the xen patches 2019-04-12 10:25:06 didn't looked on armv7 but know that aarch64 have 4.19.x 2019-04-12 10:25:09 i don't yet dare to play around with that 2019-04-12 10:25:32 mps: 4.19.x would certainly be nicer 2019-04-12 10:26:50 I tried armv7 before official release and I think it was 4.19.x but not sure now, that was two months ago 2019-04-12 10:27:14 mps: the scripts i use to make the image are 3 years old, i think that's why :( 2019-04-12 10:27:39 but the xen version is super current. it is a bit strange. can't look at that till i got serial console working all the time 2019-04-12 10:27:46 ah, then this could be reason 2019-04-12 10:27:51 but thanks for letting me know there probably IS a current kernel 2019-04-12 10:29:08 two weeks ago I made script which download and install alpine aarch64 in qemu 2019-04-12 10:29:31 have plan to do same for armv7 2019-04-12 10:29:45 and maybe armhf 2019-04-12 10:30:18 but last week had a lot of daily work 2019-04-12 10:35:23 its nice you had time to do that :) 2019-04-12 11:40:22 Hi all, we tried to upgrade our server to 3.9. 2019-04-12 11:40:22 After installing the first few packages (libcrypto, libssl, apk-tools), we received the following error message: 2019-04-12 11:40:22 ERROR: unsatisfiable constraints: 2019-04-12 11:40:22 Huh? Error reporter did not find the broken constraints. 2019-04-12 11:40:22 the command 'apk fix apk-tools' doesn't seem to help. Any ideas how to fix this issue? 2019-04-12 11:41:29 midasi: you probably want libcrypto1.1 and libssl1.1 2019-04-12 11:41:55 yes, that's what I meant... we run 'apk fix libcrypto1.1 libssl1.1 apk-tools' to reinstall all 3. 2019-04-12 11:42:14 apk fix succeeded, but apk upgrade still throws the above-mentioned error 2019-04-12 11:42:31 what's the exact error message verbatim? 2019-04-12 11:42:43 ERROR: unsatisfiable constraints: 2019-04-12 11:42:44 Huh? Error reporter did not find the broken constraints. 2019-04-12 11:42:50 oh, that's odd, let me check 2019-04-12 11:42:55 i don't think i've seen that one before 2019-04-12 11:43:21 hm 2019-04-12 11:43:28 perhaps you have some libressl packages installed? 2019-04-12 11:43:47 good point.... let me check 2019-04-12 11:43:49 mind pasting your /etc/apk/world @ tpaste.us? 2019-04-12 11:47:09 http://tpaste.us/YMXM 2019-04-12 11:48:09 I'm not seeing anything obvious that would break it 2019-04-12 11:48:11 perhaps the openssl package causes the3 issue? 2019-04-12 11:51:31 I removed openssl which removed as well libssl1.0 and libcrypto1.0 ... but still the same issue 2019-04-12 11:53:17 aaah 2019-04-12 11:53:22 How do I configure Alpine (3.9) so that .local hostnames are resolvable through DNS? 2019-04-12 11:53:23 remove all three and install openssl + libcrypto1.1 + libssl1.1 2019-04-12 11:53:44 I get this warning from `dig` when I try to resolve a .local hostname: WARNING: .local is reserved for Multicast DNS 2019-04-12 11:54:09 midasi: instead of having the version in the package version field, it's in the package name as well - not quite sure why, but i'm sure there's a reason 2019-04-12 11:54:09 And I simply can't resolve such hostnames via DNS lookup in Go programs 2019-04-12 11:54:35 i'm guessing apk stumbles when libssl1.0 and libssl1.1 provide the same .so files 2019-04-12 11:55:08 @danieli: still the same... 2019-04-12 11:56:02 _ikke_: any clue? 2019-04-12 11:56:25 I even apk fixed libcrypto1.1 libssl1.1 openssl apk-tools... but still the same 2019-04-12 11:56:40 <_ikke_> danieli: nope, not really my expertise 2019-04-12 11:56:59 all right - I'm not sure what causes this 2019-04-12 11:59:15 thanks anyway... I'll try to verify the dependencies on each single package 2019-04-12 11:59:39 if you stick around, someone else might have a clue 2019-04-12 12:00:18 sure, thank you! 2019-04-12 12:26:42 midasi: could 'apk fix -d' (depends) help 2019-04-12 12:29:54 savantgarde: what is your resolver and how it is configured 2019-04-12 12:30:19 dnsmasq works for .local domain in my experience 2019-04-12 12:30:27 mps: No idea tbqh, it's a Docker image based on alpine:3.9 2019-04-12 12:30:47 How would I use dnsmasq? 2019-04-12 12:31:01 When it comes to the Go programs, they use the built-in Go runtime resolver 2019-04-12 12:31:50 don't know for docker images, but on normal machines install it and confiugre to use /etc/hosts and put there whatever you need 2019-04-12 12:32:34 I can manipulate /etc/hosts in the Docker image 2019-04-12 12:32:37 How would I do it? 2019-04-12 12:32:40 maybe go resolver is the source of problem, don't know how it works 2019-04-12 12:32:58 I actually tried to force Go to use the libc resolver, didn't help 2019-04-12 12:33:14 How would you solve it via dnsmasq though? 2019-04-12 12:33:29 simlpy add required record, 10.1.1.1 xyx.local 2019-04-12 12:34:51 if you want to use this resolver from local machine/docker it is enough to put record in /etc/hosts 2019-04-12 12:35:12 Ah, the IP addresses aren't known though so I can't add them to /etc/hosts 2019-04-12 12:35:20 local machine only, I mean 2019-04-12 12:35:27 They're dynamic 2019-04-12 12:36:06 then you need to integrate resolver with dhcp server 2019-04-12 12:36:10 Right 2019-04-12 12:38:45 mps: i just found this line: cd linux-stable && git checkout -f v4.4.14 && cd .. 2019-04-12 12:38:58 (so this explains why i had the old kernel ;-) 2019-04-12 12:38:59 dnsmasq could serve as dhcp server and resolver at same time 2019-04-12 12:39:09 hey 2019-04-12 12:39:23 can I ask support questions here? 2019-04-12 12:39:28 <_ikke_> yes 2019-04-12 12:39:58 darkfader: nice that you found reason for outdated kernel 2019-04-12 12:40:50 @mps: unfortunately not, we tried that already (apk fix -d) 2019-04-12 12:42:28 midasi: I had similar issue once, but unfortunately forgot how I solved it :( 2019-04-12 12:43:24 mps: unfortunately, we have the same issue on several servers (5 out of 30), but couldn't find anything they have in common beside the standard packages 2019-04-12 12:43:30 midasi: don't put '@' before nick on IRC 2019-04-12 12:43:40 ok 2019-04-12 12:44:23 I think I removed packages which depends on libressl 2019-04-12 12:46:03 by force i mean, 'apk del pkg1 pkg2' etc 2019-04-12 12:47:33 or sometimes I do 'apk del -r pkg' where pkg is package which I want to remove and of its reverse dependant packages 2019-04-12 12:47:58 s/and of/and all of/ 2019-04-12 12:48:11 I'm running alpine linux in LXC container (version 3.9), the purpose is to run nginx and everything works great except rc-service only works for starting nginx, can't use it to stop or reload. In /etc/nginx/nginx.conf I have pid set to "/var/run/nginx.pid" and I have found the master process to be the correct pid . Still I cannot do rc-service ngin 2019-04-12 12:48:11 x stop 2019-04-12 12:49:09 <_ikke_> Guest36: what happens when you do that? 2019-04-12 12:49:23 "nginx is already stopped" 2019-04-12 12:49:31 but its active when I do "ps" 2019-04-12 12:49:32 Guest36: nginx pid is set in /etc/init.d/nginx 2019-04-12 12:50:16 so if I swap that in /etc/nginx/nginx.conf it should work? 2019-04-12 12:50:35 it works in all my machines 2019-04-12 12:50:41 ok, I will try 2019-04-12 12:51:11 remove it from /etc/nginx/nginx.conf and look if it is already set in /etc/init.d/nginx 2019-04-12 12:51:40 ok sec 2019-04-12 12:54:07 yes Guest36, thank you so much it works 2019-04-12 12:54:14 I mean mps 2019-04-12 12:54:16 ;D 2019-04-12 12:55:24 yw, :D 2019-04-12 12:58:44 mps: we are trying exactly that... but instead of remove we reinstall them one by one 2019-04-12 13:00:15 midasi: ime, remove and then install is safer, but use apk.static insted of apk, more safe it is 2019-04-12 13:02:08 well, since most of these packages are related to our main application on the server I'm a bit hesitant to remove them... I don't want to lose all the configuration and data 2019-04-12 13:03:10 config data stay if you changed it 2019-04-12 13:04:19 but, archive is the only right way to have configs and other data safe 2019-04-12 13:05:05 yes we have a full backup...but still 2019-04-12 13:05:16 there is a '--purge' if you want to remove configs 2019-04-12 13:06:05 I didn't noticed any time that apk del removes configs, but still I'm careful to have backup of them just in case 2019-04-12 13:07:13 ok...let try that 2019-04-12 13:07:52 last week I got screwed /etc/apk/world due to filesystem error, and with 'apk add -l' fixed it using world file from archive 2019-04-12 16:08:09 good day 2019-04-12 16:09:23 I've had a longer discussion about Linux distributions with a friend of mine and I am (re-)considering to give alpine a try for a bigger installation. One question from my side is, what are the steps required / ways to include a new package (like ceph) into alpine? 2019-04-12 16:13:16 telmich: ceph isn't new, it's in edge/testing 2019-04-12 16:13:20 https://pkgs.alpinelinux.org/package/edge/testing/x86_64/ceph 2019-04-12 16:17:38 but yeah, let's pretend we're talking about a fictional package you want to get into stable alpine repositories 2019-04-12 16:17:58 first you'd have to create an APKBUILD, potentially using code from other APKBUILDs or a template 2019-04-12 16:18:31 second you'd need to make it build - your software might require patches to work, for instance, with musl, without systemd, etc. 2019-04-12 16:18:59 once you've taken maintainership and gotten it pushed, it'll get build and put in the edge/testing repository, which is only accessible on edge 2019-04-12 16:19:17 once it's considered 'usable' and stable (it's been tested), the maintainer can request moving it to community 2019-04-12 16:42:56 hey, on the LVM-on-LUKS wiki page, I see a note that en-us is used during boot password entry; is that still true, or is it possible to load a different keymap? 2019-04-12 16:47:21 danieli: sounds rather straight forward, thanks for the explanation! 2019-04-12 16:47:37 telmich: you're free to ask in #alpine-devel any time, if it's related to development :) 2019-04-12 16:47:41 danieli: the non-systemd part is actually one of the reasons I look at alpine 2019-04-12 17:23:17 keymap in /etc/mkinitfs/features.d works great, time to update the wiki 2019-04-12 17:28:07 telmich, I've been running Alpine on my small server at home (which is actually an EeePC900) for about a month now and it is very stable 2019-04-12 17:37:41 Parsec300: sounds promising. I'll have to understand the release model / way of running potentially bigger infrastructures first, but I'll certainly give it a try for running some ceph clusters at start 2019-04-12 18:17:50 telmich, it's memoryconsumption is minimal 2019-04-12 18:19:11 But you need to install iptables if you want a firewall (which you may want to have). It is really a minimal install and you need to install many other things. Like the ability to view man-pages and the documentation on almost everything 2019-04-12 18:19:46 But I quite like it so far 2019-04-12 18:24:24 iptables is pretty much the de-facto standard for linux firewalling 2019-04-12 18:24:53 and pf for bsd 2019-04-12 18:25:16 <_ikke_> you can use awall on AL, but it would still rely on iptables 2019-04-12 18:27:06 yes, to my knowledge it's more of an interface to iptables 2019-04-12 18:27:16 kinda like ufw 2019-04-12 18:27:23 <_ikke_> indeed 2019-04-12 18:27:44 <_ikke_> iptables is kind-of low level 2019-04-12 18:27:57 sort of kind of 2019-04-12 18:28:19 netfilter is the truly low-level part 2019-04-12 18:33:05 Indeed, iptables is a user-space interface to the netfilter kernel-level part 2019-04-12 18:33:15 how far do you people go with alpine? I can easily see it in router/server - do you also use it as a desktop? 2019-04-12 18:33:35 <_ikke_> telmich: I personally don't, but ncopa and others do use it as desktop 2019-04-12 18:33:39 telmich, it is possible, but I don't do that 2019-04-12 18:33:51 it still lacks some software I use for convenience and quite a bit of offensive software I use for security work 2019-04-12 18:34:04 but my end goal is to use alpine as a daily driver 2019-04-12 18:34:08 I might give it a try in the future when I build my own desktop again. Currently I'm on a Mac 2019-04-12 18:34:44 I had a look at alpine many years ago, I think there was a much more sparse homepage at that time - as a "third party" it looks like a lot of stuff happened since then 2019-04-12 18:35:10 danieli, do you run Kali on your desktop as your primary OS? 2019-04-12 18:35:13 it's become a pretty darn popular distribution 2019-04-12 18:35:22 Parsec300: no, and i would not recommend doing that 2019-04-12 18:35:28 <_ikke_> danieli: what are you still missing? 2019-04-12 18:35:30 And I know that desktop, if you go with full blown kde/gnome/etc. can pull in quite some complexity. I currently run devuan+i3, which is rather easy going and might actually also work on alpine 2019-04-12 18:36:25 _ikke_: you can pretty much diff some parts of the blackarch and kali repos with alpine's, and that's what i'm missing 2019-04-12 18:36:47 <_ikke_> so specific analysis tools? 2019-04-12 18:36:59 i could compose a list of it, but i've been doing it impulsively for now 2019-04-12 18:37:38 and not just for analysis, that's a task usually associated more with blue teaming (defensive work) 2019-04-12 18:38:08 <_ikke_> don't red teams need analysis tools as well? 2019-04-12 18:38:26 to some degree, but not to the same extent as blue teams do 2019-04-12 18:38:41 it's more about discovery and scanning 2019-04-12 18:38:56 <_ikke_> right, i kind of meant that with analysis tools 2019-04-12 18:40:09 we don't have, for instance, metasploit 2019-04-12 18:40:16 I run Alpine 'desktop' (awesome wm) and my wife Alpine with xfce 2019-04-12 18:41:01 wow! I just wanted to know how you deal with networking, especially vlan/bond/bridge and it's all there on the first page I looked for it 2019-04-12 18:41:03 although I have Debian in lxc container in case that I need some software which is not in Alpine 2019-04-12 18:41:04 impressive, I have to say! 2019-04-12 18:41:28 ahh, it's basically interfaces from debian 2019-04-12 18:41:29 no offense, but that's fairly trivial :) 2019-04-12 18:41:32 pretty much 2019-04-12 18:41:40 same format, but not the same as debian's 2019-04-12 18:41:40 I tend to highly disagree there 2019-04-12 18:42:01 personal opinion: i hate the network scripts on rhel derivatives 2019-04-12 18:42:13 well, I start that Debian container not to often 2019-04-12 18:42:19 nice... vlan tagged bond in bridge, exactly what I was looking for 2019-04-12 18:43:28 actually, I need Debian only for kicad (electronic cad) 2019-04-12 18:43:45 complex vlan and vrf configuration as well as bonding can be a bit difficult, but the basics are fairly simple 2019-04-12 18:43:59 i guess that's my muscle memory speaking for me 2019-04-12 21:18:34 can alpine linux be installed onto xfs? 2019-04-12 21:18:48 sure 2019-04-12 21:19:23 but there is missing support in installer or something? 2019-04-12 21:19:46 as I'm looking at it on distrowatch and it lists only ext4 2019-04-12 21:20:10 ent91: you can do install manually 2019-04-12 21:20:21 I did on f2fs 2019-04-12 23:00:46 wow, can one of you by chance translate this crosscompile error to human: 2019-04-12 23:00:47 "/usr/bin/ld: scripts/kconfig/conf.o: relocation R_X86_64_32 against `.rodata.str1.8' can not be used when making a shared object; recompile with -fPIC 2019-04-12 23:00:50 " 2019-04-12 23:00:58 i know what all the terms mean 2019-04-12 23:01:05 but not what the sentence means 2019-04-12 23:05:57 it means you're trying to link a non-pic .o file (possibly from a .a library) into a shared library or PIE executable 2019-04-12 23:12:50 dalias: thanks. i'll go through the objects it tries to link and ask them why :) 2019-04-12 23:13:36 it must have forgotten that it's supposed to crosscompile. that helps! 2019-04-13 00:54:45 how can I define in APKBUILD which config files should not be overwritten during an upgrade? 2019-04-13 00:55:33 these files should be part of update-conf instead... 2019-04-13 07:43:15 midasi: files with .conf extension 2019-04-13 14:40:55 can I retarget /usr/bin/python to python3, or should I expect widespread breakage? 2019-04-13 14:41:28 (alpine doesn't seem to be like RHEL, with half the system utilities written in python 2.7, but I imagine a lot of packaging might have to change) 2019-04-13 14:41:39 does that target /usr/bin/python2 or /usr/bin/python3 2019-04-13 14:42:00 currently 2 2019-04-13 14:42:25 that's a change i'd almost say is worth incrementing the major version of alpine over 2019-04-13 14:43:35 however 2019-04-13 14:43:47 i would be interested to see what fails if you just make that chageover today 2019-04-13 14:43:49 change* 2019-04-13 14:44:28 <_ikke_> We do intend to phase out python2 soon 2019-04-13 15:33:08 I will try it later on a less important host 2019-04-13 15:34:53 <_ikke_> things can break. I think a lot of things assume python == python2 2019-04-13 16:05:17 very strange, travis seems not to be able to download the emacs sources, and drone seems to have some aslr/exec-shield mitigation that cannot be disabled, and thus all CI tests fail on the new emacs, while in a private minimalistic alpine vm the build succeeds: https://github.com/alpinelinux/aports/pull/7074 2019-04-13 16:22:05 hi 2019-04-13 16:23:40 any tips to get mouse working? I installed i3, keyboard works but the mouse does not mouse. I tried installing xf86-input-mouse and xf86-input-synaptics and reboot. also tried installed evdev and rc-update add udev 2019-04-13 16:24:13 its a usb mouse 2019-04-13 16:25:46 rain1: xf86-input-libinput 2019-04-13 16:26:01 thanks, ill try it 2019-04-13 16:33:03 it didn't seem to have an effect 2019-04-13 16:34:32 what /var/log/Xorg.0.log says 2019-04-13 16:35:37 btw, did you installed udev and run it 2019-04-13 16:36:04 and dbus, althoug dbus should be installed by default 2019-04-13 16:36:26 I didn't add dbus 2019-04-13 16:36:46 I'll add it and i'll check the log too, thanks for the suggestions 2019-04-13 16:37:06 if you new to Alpine there is 'setup-xorg-base' script which do most things for you 2019-04-13 16:42:52 pretty sure you don't need udev or dbus to run X… 2019-04-13 16:44:41 ahills: if you want to X autoconfig input device? 2019-04-13 16:45:25 only for hotplugging 2019-04-13 16:46:21 you tried? 2019-04-13 16:47:10 Xorg.0.log: The server relies on udev to provide the list of input devices. 2019-04-13 16:49:26 with the libinput driver, right? 2019-04-13 16:51:06 ahills: I'm not talking about how it is possible to run X, just trying to help someone who is new to Alpine X, probably 2019-04-13 16:51:07 adding dbus didn't seem to help, i have run setup-xorg-base. i tried the X -configure to generate a conf and that conf did not seem to have any effect 2019-04-13 16:51:39 rain1: sorry if I'm wrong about your experience with X on Alpine 2019-04-13 16:51:48 I just installed it today 2019-04-13 16:51:56 i appreciate all the ideas for getting mouse to work 2019-04-13 16:52:40 you do not need X -configure, X can detect automatically what it need 2019-04-13 16:53:13 what says: rc-service udev status 2019-04-13 16:53:17 rain1: just to make sure, you ran setup-xorg-base, and it succeeded, and then you started X after that? 2019-04-13 16:53:22 https://bpaste.net/show/c12395212800 theres some of the xorg log 2019-04-13 16:54:41 you have some ' Cannot open device /dev/input/eventX' 2019-04-13 16:54:45 rc-service udev status ill try that and see what it says next time. I did notuce that udevd was running by ps | grep 2019-04-13 16:55:26 are you sure that the X is not running in background 2019-04-13 16:55:55 yeah, i launch it with startx and before that there is no X running 2019-04-13 16:56:32 log says that the devices are busy 2019-04-13 16:56:59 something is using them 2019-04-13 16:57:12 sorry, now have to go AFK 2019-04-13 16:59:00 bye! thanks again 2019-04-13 17:00:26 I'm back, had to reboot machine because messed br0 with wlan0 :) 2019-04-13 17:01:49 rain1: you could try rebooting, if there's nothing important happening on the machine 2019-04-13 17:02:02 yea i rebooted it 2019-04-13 17:02:09 each time i tried something 2019-04-13 17:02:38 look if the dbus and udev are running 2019-04-13 17:03:30 rain1: does "rc-update show sysinit" show "udev | sysinit" in the list? 2019-04-13 17:04:18 do you have any other services running? 2019-04-13 17:09:16 show sysinit lists devfs, dmesg, udev, udev-trigger 2019-04-13 17:09:54 ps shows me dbus-demon and udevd are running 2019-04-13 17:12:04 how are you launching X? 2019-04-13 17:12:25 startx 2019-04-13 17:12:33 and my .xinitrc file just has: exec i3 2019-04-13 17:12:43 as a non root user 2019-04-13 17:12:59 ah, I wonder if your user needs to be in a special group to access the devices 2019-04-13 17:13:38 yeah, looks like it's only readable by group "input" by default 2019-04-13 17:13:50 try adding your user to that group, then log out/in again, and try again 2019-04-13 17:14:10 ill give it a shot, ty! 2019-04-13 17:15:50 id shows my user is in gruop 'input' but still cannot move the mouse 2019-04-13 17:15:55 :/ 2019-04-13 17:16:24 I still have the same 'no input driver specified' messages in the xorg log 2019-04-13 17:16:33 and resource busy 2019-04-13 17:16:38 if you "ls -l /dev/input", do you see a dozen or so files with mode 0660 and ownership root:input? 2019-04-13 17:17:22 yeah there's event 0 to 19 and mice and mouse0 2019-04-13 17:17:47 i cant cat /dev/inptu/mouse0 as my user or as root 2019-04-13 17:17:50 resource busy 2019-04-13 17:18:37 what does this command show: 2019-04-13 17:18:48 mount | awk '$3=="/dev"{print $0}' 2019-04-13 17:19:21 do you see something like 2019-04-13 17:19:33 devtmpfs on /dev type devtmps (rw,nosuid,…) 2019-04-13 17:20:18 yeah i do 2019-04-13 17:22:57 i just installed xfce4 and actually even with startxfce i can't move the mouse 2019-04-13 17:24:08 i tried a different mouse and it worked 2019-04-13 17:24:26 so i guess this particular mouse is less supported - it works on arch linux but not here 2019-04-13 17:26:42 oh interesting 2019-04-13 17:26:45 which mouse is it? 2019-04-13 17:27:01 its "a4 tech bloody" 2019-04-13 17:27:06 v7m 2019-04-13 17:27:06 and do you still get the same errors in Xorg.0.log? 2019-04-13 17:27:08 its a really bad mouse 2019-04-13 17:27:18 but the apple mouse works 2019-04-13 17:28:28 it's weird that the mouse would cause problems with libinput and eudev… 2019-04-13 17:28:45 watching the log with tail, i get a bunch of errors when i plug in the bad mouse 2019-04-13 17:28:53 but i dont get errors when i plug in the apple mouse 2019-04-13 17:29:12 yeah it's pretty weird especially with that mouse normally working ok on my other linux 2019-04-13 17:32:16 can you check the Xorg logs when you plug it in on Arch? 2019-04-13 17:34:24 rain1: try installing "libinput"; it includes some "quirks" files that might include your mouse 2019-04-13 17:34:44 it looks like the xorg driver only depends on libinput-libs 2019-04-13 17:34:54 oh i thought that would have been installed along with xf86-input-libinput 2019-04-13 17:34:57 but ill try installing it 2019-04-13 17:35:05 no, only the shared libraries, not the other files 2019-04-13 17:35:14 im just looking for where the xorg logs are on arch right now 2019-04-13 17:35:17 apparently they are in a different place 2019-04-13 17:36:43 https://bpaste.net/show/48be958c8ff4 this is when i plug in the bad mouse 2019-04-13 19:58:05 I've noticed that some openrc init scripts have stuff like: command_args="$CRON_OPTS" 2019-04-13 19:58:15 ...which implies that you can pass arguments to them somehow; how does this work? 2019-04-13 19:59:48 trfl, the options come from files in /etc iirc 2019-04-13 19:59:55 not from arguments 2019-04-13 20:00:38 ah yes you're right, /etc/conf.d/crond in this case. Cheers! 2019-04-13 21:37:28 Is there a way to define in APKBUILD which (config) files should not be overwritten during an upgrade? 2019-04-13 21:37:28 Those files should be part of update-conf instead... 2019-04-13 22:03:29 i tried installed libinput but it didn't help make that mouse work 2019-04-13 22:09:34 midasi: I already answered this morning. files with .conf extension 2019-04-13 22:11:09 well, I had several conf files being overwritten (not in /etc). Moreover, I have other files which should not be overwritten during the upgrade. Is there a way to specify the behaviour? 2019-04-13 22:12:29 ah, that, this is specific, I don't know for such option in APKBUILD 2019-04-13 22:13:24 ok thanks... 2019-04-13 22:14:04 usually conf files for packaged goes in /etc 2019-04-13 22:15:13 yes I know, I'll try to configure the application differently to store those files in /etc 2019-04-13 22:16:41 it is currently running in a chroot under /var.... 2019-04-13 22:19:09 could you install you package under chroot, and it put conf there in /etc 2019-04-13 22:21:05 hmm yes, I could try that... 2019-04-13 22:22:00 you create 'normal' package and under chroot you do 'apk add ....' 2019-04-13 22:24:29 let me try if that works... thanks a lot! 2019-04-14 09:36:15 hi, i've installed in a virtualbox vm alpine-virt, with xorg, xfce4 and the faenza icon theme, following the instructions on the wiki 2019-04-14 09:36:53 but although i have icons on the desktop and in some apps like the thunar file manager, the icons are missing from the panels 2019-04-14 09:37:01 any idea what's going on? 2019-04-14 09:39:01 this happens also on bare metal install 2019-04-14 09:39:22 some icons missing, probably 2019-04-14 09:41:27 there is a package, faenza-icon-theme-xfce4-panel which should handle the icons on the panels 2019-04-14 09:41:33 it is installed 2019-04-14 09:41:45 maybe something wrong with this package? 2019-04-14 09:42:35 I didn't looked but noticed that on my wife's machine, same as your VM 2019-04-14 09:43:52 anyway, 'apk info -L faenza-icon-theme-xfce4-panel' could give you a hint what could be wrong 2019-04-14 09:45:09 gnome-icon-theme seems to work allright 2019-04-14 09:47:31 good to know 2019-04-14 10:52:39 hey guys, i dont really use irc that often so idk if im even connected properly tbh 2019-04-14 10:56:43 <_ikke_> you are 2019-04-14 10:58:25 ik that alpine is not meant for home use on laptops but im trying to use it on my X230 and cant figure out how to use wpa supplicant even tho i followed the wiki, i used to use wicd-curses on debian but it seems to not be working on alpine (looked up the error and seems to only be fixed on debian or the arch aur) i was wondering if there is any curses wifi manager that work properly on alpine becuase thats literally the o 2019-04-14 10:59:49 <_ikke_> I have no experience with it on Alpine, but I know someone is working on iwd 2019-04-14 11:00:31 well i guess ill have to wait 2019-04-14 11:01:11 i tried so many wifi managers on alpine, connman, networkmanager, wicd and all didnt really work all that well 2019-04-14 11:01:22 most threw error upon launch and didnt start 2019-04-14 11:01:33 which isnt surprising since this distro is aimed at servers and containers 2019-04-14 11:01:45 but i really love the simplicity of alpine 2019-04-14 11:02:08 plus no systemd is a huge plus 2019-04-14 11:06:54 _ikke_: speaking on iwd, do you have push rights to testing 2019-04-14 11:07:03 <_ikke_> yes 2019-04-14 11:08:14 I prepared and post iwd 0.16 about ten days ago, https://patchwork.alpinelinux.org/patch/4772/ 2019-04-14 11:08:36 it is not pushed to aports 2019-04-14 11:08:56 but last night I prepared new version 0.17 2019-04-14 11:09:27 <_ikke_> patchwork honestly has a lot less visibility as github 2019-04-14 11:09:41 so don't know should I wait for 0.16 to be pushed or just send new version 2019-04-14 11:10:36 especially because new version have some important security fixes 2019-04-14 11:11:08 btw, do we have security channel or ML 2019-04-14 11:11:31 <_ikke_> mps: yes to both 2019-04-14 11:11:50 and these are open or invite only? 2019-04-14 11:11:56 <_ikke_> But the use of them is not really defined yet 2019-04-14 11:12:06 <_ikke_> open atm 2019-04-14 11:12:50 thanks, do you know name of irc channel, I will look for ML 2019-04-14 11:13:17 <_ikke_> mps: the name is guessible :) 2019-04-14 11:13:23 <_ikke_> https://lists.alpinelinux.org/ 2019-04-14 11:15:04 eh :), thanks 2019-04-14 11:15:28 they are not on community page so didn't know 2019-04-14 11:15:40 <_ikke_> mps: I guess just submit the 0.17 uprade, I will look at it 2019-04-14 11:15:58 <_ikke_> It's no use to first push 0.16 and immediately 0.1 2019-04-14 11:16:03 <_ikke_> 0.17* 2019-04-14 11:17:04 ok, will post new version, but first have to ask iwd developers if it have CVE assigned because I can't find it 2019-04-14 11:17:25 _ikke_: thank you for help 2019-04-14 11:17:54 johrmungand: you can use iwd as wireless client and AP 2019-04-14 11:18:27 it has iwctl CLI which can be used interactively or from scripts 2019-04-14 11:20:09 but, it is testing and not in main and commynity, but will be moved to community, I think because BFDL told so few days ago i 2019-04-14 11:20:37 BDFL* 2019-04-14 11:21:24 <_ikke_> mps: if you verified it works properly, you can submit the move to community yourself 2019-04-14 11:21:38 <_ikke_> No need to get ncopa's approval for that 2019-04-14 11:21:55 aha, nice to know 2019-04-14 11:22:41 will do, just have to check if this version have CVE 2019-04-14 11:22:47 <_ikke_> testing is just a staging area to be able to submit packages and verify in a broader context that it works (ie, builds on all arches, no hidden issues). 2019-04-14 11:23:48 <_ikke_> mps: ping me when you have 0.17 posted 2019-04-14 11:24:08 will do, thanks again 2019-04-14 12:35:49 hey everyone! coldn't find a definitive response to this anywhere (or atleast not one that i could understand) so i figured id try the irc channel :) im trying to set up my raspberry pi 3b+ with alpine and i cant figure out whether to choose the aarch64 or armhf version. which one do i choose? 2019-04-14 12:37:44 RPi3 is 64bit, afaik, so aarch64 is for it 2019-04-14 12:39:43 ok thanks! 2019-04-14 13:03:00 kroa, it depends on which you want to run 2019-04-14 13:03:09 do you want a 64-bit os or a 32-bit one? 2019-04-14 13:03:22 3b+ has a 64-bit cpu so you have both choices 2019-04-14 13:04:49 dalias: true, but some people told that armv7 AL have a problem to boot on RPi3, I don't have board so not sure if this is true 2019-04-14 13:05:57 is there a reason to choose 32 over 64 bit? 2019-04-14 13:06:44 kroa: that's on user to know, i.e. if there are reasons 2019-04-14 13:07:01 you effectively double your ram :-p 2019-04-14 13:07:16 kinda a joke, but for most software that's pointer-heavy it's roughly true 2019-04-14 13:08:46 i would be cautious of what mps says about it having a problem with the hardware 2019-04-14 13:09:17 but for a resource-constrained device like a pi i'd personally choose 32-bit unless i had a real reason i needed 64-bit address space 2019-04-14 13:09:55 mps ok cause the thing that had me a bit hesitant when choosing version whas that the wiki tutorial for setting up "classic install" said to use armhf. but since i cant think of a reason not to use the 64 bit one i guess ill try it out 2019-04-14 13:10:18 I'm not sure about boot issues, really. kroa can try and see what will happen. 2019-04-14 13:11:31 wiki should be updated, obviously. 2019-04-14 13:12:45 *nod* 2019-04-14 13:14:24 just out of curiosity, whats the difference between the armv7 and armhf version, if theyre both 32 bit 2019-04-14 13:16:36 armhf is built with SoftFP 'in mind', i.e. use hardware floating if present in hardware and fall back to software floating point else 2019-04-14 13:16:52 huh? 2019-04-14 13:16:58 that does not sound like armhf 2019-04-14 13:17:46 no? I could be wrong, of course 2019-04-14 13:18:25 unless completely misused, armhf always means hardfloat abi (EABI hardfloat variant) 2019-04-14 13:18:51 right, that is in ABI 2019-04-14 13:19:39 what if CPU doesn't have hardware floating point 2019-04-14 13:20:14 then it can't run unless you have trap-and-emulate code (5000x slower) in kernelspace 2019-04-14 13:20:57 (ok maybe only like 1000-2000x) 2019-04-14 13:21:24 again right, but in my (possibly false) knowledge AL armhf can run on CPU without hardware floating point unit 2019-04-14 13:21:32 it definitely can't 2019-04-14 13:22:18 i think originally there were arm and armhf targets (arm was v5 baseline, armhf was v6 baseline and hardfloat) 2019-04-14 13:22:25 aha, never tried so don't know, and my presumptions were false then 2019-04-14 13:22:34 arm seems to have been dropped 2019-04-14 13:23:03 and armv7 looks like it's just the same as armhf but optimized for v7 cpus (won't run on v6) 2019-04-14 13:23:25 looking at https://git.alpinelinux.org/aports/tree/main/gcc/APKBUILD#n220 2019-04-14 13:23:50 i'm not sure where the logic that converts the alpine arch names into gcc tuples is tho 2019-04-14 13:24:01 ok thanks a lot for the help guys! gonna go try and set up my rpi now :) 2019-04-14 13:24:38 agree, this naming scheme is somewhat confusing to me 2019-04-14 13:25:10 the download page should really clarify 2019-04-14 13:25:17 since the names are not at all self-descriptive 2019-04-14 13:27:17 heh, I have a book about these arm differences but couldn't made clear mental picture in my head :) of that mess 2019-04-14 13:28:41 although must admit I didn't tried hard 2019-04-14 13:41:16 hi. dalias is right 2019-04-14 13:41:27 armhv is armv6 with hardfloat 2019-04-14 13:41:36 armv7 is optimized for armv7 2019-04-14 13:42:37 re aarch64 on rpi3, we had a nasty issue with aarch64 python3 on rpi3 2019-04-14 13:42:47 it simply segfaulted on startup 2019-04-14 13:42:57 rebuilding it with gcc 8.3.0 solved it 2019-04-14 13:44:50 i think the boot issues with armv7 should be fixed with 3.9.3 release 2019-04-14 13:52:37 :) 2019-04-14 18:05:35 hello 2019-04-14 18:09:10 hello 2019-04-14 18:09:15 I am trying to PXE boot Alpine on a 3-NICs box. The cable is on eth0. When the kernel has booted up and openrc starts up, udhcpc sends discovers on eth2 indefinitely. If i switch the cable from eth0 to eth2, it gets an IP address. 2019-04-14 18:10:13 But I would like to avoid switching the cable from eth2 to eth0. Any idea how to achieve it ? 2019-04-14 18:11:02 Oh, my bad, at this moment, openrc has not started yet 2019-04-14 18:12:57 at first I thought there was some mix-ups with the NICS but once the system is booted up, 'ip a' shows an ip address on eth2 2019-04-14 22:45:10 how would one autostart lxdm? is there an openrc service for it or something? 2019-04-14 22:46:57 Invader_Bork: as far as I can see, `rc-update add lxdm` or something along those lines 2019-04-14 22:47:10 it installs /etc/init.d/lxdm 2019-04-14 22:49:05 danieli, yep that did it 2019-04-14 22:49:08 thank you! 2019-04-14 22:53:44 oh man a fully working install (vm) with xfce4 uses 680mb of hdd space and 190mb of ram while running 2019-04-14 22:53:55 now those are numbers i haven't seen in a long time 2019-04-15 06:38:48 https://twitter.com/MashiRafael/status/1116334415096025089 2019-04-15 12:04:45 hi everyone. Is there a way to have alpine-linux loading new locales ? 2019-04-15 12:05:26 my software, uses system locales for language translation like January with fr_FR.utf8 is transformed to 'Janvier' using system locales. 2019-04-15 12:05:59 I've not found any locales package for alpine, is there a way to have it working for alpine somehow ? 2019-04-15 12:06:24 srounet: Alpine is based on musl libc, which doesn't support locales 2019-04-15 12:07:23 mps: aight, then my only chance is having the software use another technique to get the translation? 2019-04-15 12:08:52 depend where you need it, I rebuilt coreutils with enabled locales and put it manually under /usr/share and it worked, but not quite fine for datetime 2019-04-15 12:09:24 I only need it for month translation to another language 2019-04-15 12:09:45 our software is build on top of python which rely on system locales 2019-04-15 12:10:00 I'll have a look to coreutils 2019-04-15 12:11:41 srounet: I tried this only with tcsh (shell) and no other use cases 2019-04-15 12:12:47 I'll do dict-mapping then, ugly asf but well I guess it will be enough for now 2019-04-15 12:17:39 thank you 2019-04-15 12:23:10 srounet: yw 2019-04-15 13:01:00 Since the release 3.9.2 we have a problem with 'wget' and https pages. Error message: 'Disabling SSL due to encountered errors'. It seems that wget has a problem with the linked openssl 1.1.1. 2019-04-15 13:01:07 is this issue already known? 2019-04-15 13:05:18 midasi: which arch? and is the datetime correct on machine 2019-04-15 13:12:29 midasi: x86_64 - date is synchronized with ntp and is up to date 2019-04-15 13:13:21 mps: let me check if I have the same behaviour on another machine 2019-04-15 13:15:22 I had similar problem when the clock skew is high on some boxes without rtc. 2019-04-15 13:21:05 Which wget? 2019-04-15 13:21:46 Bb it regular 2019-04-15 13:21:49 Or 2019-04-15 13:24:09 mps: interesting... I forced the sync and now it's working 2019-04-15 13:24:38 sync datetime? 2019-04-15 13:24:43 clandmeter: the regular wget 1.20.3 2019-04-15 13:24:46 mps: yes exactly 2019-04-15 13:25:13 that is because ssl/tls relies on correct time 2019-04-15 13:25:55 i.e. CERT's 'not before' and 'not after' attributes 2019-04-15 13:26:14 I know, but the offset was minimal...anyway it is working now 2019-04-15 13:27:51 so, then not only on these attr's but also time of connection setup couldn't have high skew 2019-04-15 13:53:21 Invader_Bork: if you compile a self-configured (minimal) kernel, you could make it even smaller 2019-04-15 13:55:03 danieli, i've heard about alpine installs going as low as few megs 2019-04-15 13:55:33 i wonder how low could a desktop get, with a custom kernel like you mentioned and the suckless software 2019-04-15 13:56:08 5mb is about the minimum without kennel 2019-04-15 13:56:15 Kernel 2019-04-15 13:56:56 clandmeter, is there a use case without the kernel like a container or something? 2019-04-15 13:57:08 Docker 2019-04-15 13:57:17 clandmeter, figures 2019-04-15 13:57:44 It's mentions the 5mb :) 2019-04-15 13:57:56 how low could a kernel go and still have decent functionality? ... i guess it depends heavily on the use case though 2019-04-15 13:59:53 I haven't experimented with that in at least a few years, and that was on SliTaz - I can't really give you an answer to that 2019-04-15 14:15:36 good day 2019-04-15 14:15:47 I am currently adjust cdist configuration management to support alpine 2019-04-15 14:16:43 What is the notion of enabling / disabling a service for booting in openrc? 2019-04-15 14:16:56 rc-update [add|del] servicename 2019-04-15 14:16:59 ohhhhhh 2019-04-15 14:17:01 so nice 2019-04-15 14:17:06 guys 2019-04-15 14:17:08 https://wiki.gentoo.org/wiki/OpenRC_to_systemd_Cheatsheet 2019-04-15 14:17:11 https://wiki.gentoo.org/wiki/OpenRC 2019-04-15 14:17:11 you make it really hard not to be loved 2019-04-15 14:17:47 also, alpine-specific (but probably somewhat outdated) - https://wiki.alpinelinux.org/wiki/Alpine_Linux_Init_System 2019-04-15 14:20:58 https://github.com/ungleich/cdist/commit/3a2041019b5e878a0c81ef78dd719d10db48a6f6 66% done 2019-04-15 14:23:39 telmich: I see some typos in "cdist/conf/type/__package_apk/man.rst" 2019-04-15 14:23:53 at the very top - "akp" 2019-04-15 14:25:25 danieli: thanks, will fix! 2019-04-15 14:25:54 So the "best way" of finding out if a service is enabled I found `rc-update show | sed 's/ *//' | grep "^${name} |" ` 2019-04-15 14:26:02 Is there a different way that you'd recommend? 2019-04-15 14:27:15 no man page for apk? 2019-04-15 14:27:54 not yet, unfortunately 2019-04-15 14:28:01 --help on various things is helpful though 2019-04-15 14:28:28 telmich: `service X status` 2019-04-15 14:28:55 danieli, it will do ty 2019-04-15 14:28:59 are the apk packages signed? 2019-04-15 14:29:07 yes 2019-04-15 14:30:59 danieli: that seems to be doing other things as well 2019-04-15 14:31:09 checking with knotd for example: 2019-04-15 14:31:13 * knotd: `/etc/knot/knot.conf' is not readable 2019-04-15 14:31:24 even though it is enabled 2019-04-15 14:32:31 nvm - the grep is good enough for the moment 2019-04-15 14:33:05 iirc alpine had a user contributed ports like system 2019-04-15 14:33:44 is there an official repository for that? 2019-04-15 14:33:49 like arch's aur? 2019-04-15 14:34:52 ah right, finding out if it's enabled and not started, hm 2019-04-15 14:35:12 Invader_Bork: well, there's aports, but that's the packages in our repositories 2019-04-15 14:35:18 we don't have anything like aur 2019-04-15 14:36:49 so when users add aports to the tree, they just fork the git repo? 2019-04-15 14:37:52 well no, they contribute to our aports repository, one way or another 2019-04-15 14:38:07 so it's still centralized then 2019-04-15 14:39:14 well yeah, changes to aports are reviewed and tested before landing in our repositories 2019-04-15 14:39:33 of course, you could just pull an APKBUILD from somewhere and build it yourself with abuild 2019-04-15 14:39:44 and setting up your own mirror with abuild + some httpd is trivial 2019-04-15 14:40:18 i like alpine more and more haha 2019-04-15 14:40:36 i wonder how much of a hustle porting software to musl would be 2019-04-15 14:40:49 depends on how 'broken' the software is 2019-04-15 14:40:51 i guess the work is shared with the void people too 2019-04-15 14:41:23 broken meaning how much it relies on legacy features dropped by musl? 2019-04-15 14:41:23 oh... I did not even realised it was musl (I read, but forgot). Nice 2019-04-15 14:41:30 (take "broken" with a grain of salt - a lot of software depends on glibc-specific extensions) 2019-04-15 14:42:50 I assume that "everything" that runs on BSD and Linux probably works on musl, or am I totally wrong with that claim? 2019-04-15 14:43:41 somebody was saying in some other irc channel about some slowness in musl compared to glibc? danieli, did you encounter this in your experience? 2019-04-15 14:44:08 s/glibc?/glibc. 2019-04-15 14:45:14 i don't have an objective viewpoint on that for you 2019-04-15 14:45:37 additionally, i don't know what they were saying was slow, nor what they were saying 2019-04-15 14:47:48 it was more than a general pov, seemed more anecdotal rather than empiric. that's why i felt i should ask somebody more experienced 2019-04-15 14:48:06 s/more than/more like 2019-04-15 14:48:44 i haven't done too much low-level work musl 2019-04-15 14:49:09 if you have musl-specific quesions, you're free to ask them in #musl 2019-04-15 14:49:44 i have added the channel, th 2019-04-15 14:49:49 s/th/ty 2019-04-15 14:51:49 fwiw dalias is here 2019-04-15 14:52:35 dalias? 2019-04-15 14:53:09 musl maintainer/developer 2019-04-15 14:53:20 oh nice 2019-04-15 14:53:41 ohh... I can go via /etc/runlevels/* 2019-04-15 14:53:43 much easier 2019-04-15 15:03:48 ? 2019-04-15 15:05:52 Invader_Bork: in my experience musl is faster and not to small degree then glibc systems 2019-04-15 15:06:59 mps, logically if assumed the same. a newer implementation that trimmed down the fat should be faster by default 2019-04-15 15:08:05 it is easy to see difference on low-end machines 2019-04-15 15:08:37 s/if/i 2019-04-15 15:09:01 dalias, we were talking about musl vs glibc speed 2019-04-15 15:09:24 mps, that settles it 2019-04-15 15:11:18 I had Debian on different ARM boxes with low memory (512MB or 1G) and relativelly slow CPU and filesystem on SD cards for long time 2019-04-15 15:12:54 when switched to Alpine about year and a half all of them (some two or three, don't remember exactly) it was like I bought new generation hardware :) 2019-04-15 15:13:56 mps, i bet that felt great haha 2019-04-15 15:14:14 i've been thinking about getting into single board computers myself 2019-04-15 15:14:14 but it could be maybe there are speed impact on some specialized software, but I'm not in this fields of endeavor so don't have experience 2019-04-15 15:14:35 with rpi4 i might finally jump the gun 2019-04-15 15:16:18 I have Alpine on two chromebooks (arm32 and arm64) and it works really good and fast for all my works, i.e. software development 2019-04-15 15:17:19 how are the graphics on the chromebooks 2019-04-15 15:17:20 the problem are some binary only drivers for closed hardware 2019-04-15 15:17:31 do they support modern apis like vulkan? 2019-04-15 15:17:51 fortunately the lima and panfrost drivers are slowly coming along 2019-04-15 15:18:02 didn't tried because I do not need it 2019-04-15 15:19:01 jn__: yes, tried to apply patches panfrost to vanilla kernel but give up, waiting to appear in mainline 2019-04-15 15:19:20 ACTION nods 2019-04-15 15:19:41 it is not worth effort for me 2019-04-15 15:19:58 hmm with those drivers are we closer to being able to run linux distros on consumer smartphones? 2019-04-15 15:20:10 X with fbdev works quite fine 2019-04-15 15:20:50 Invader_Bork: look at postmarketOS, which is alpine based 2019-04-15 15:23:19 note also that some differences will be hardening profiles 2019-04-15 15:23:33 alpine uses fortify and stack-protector-strong for everything iirc 2019-04-15 15:23:39 ACTION is stunned 2019-04-15 15:25:03 generally musl systems use a lot less ram and thus 'feel faster' on memory-constrained systems 2019-04-15 15:25:24 mps, you made my day with postmarketOS 2019-04-15 15:25:38 part of this is musl; part of it is alpine and some other musl-based distros not using all the bloated always-running-at-startup stuff that most glibc distros use 2019-04-15 15:29:47 i see 2019-04-15 15:30:22 i've been testing a few distros lately in vms, but i think it's about time i'd install alpine on metal 2019-04-15 15:30:25 postmarket os looks very interesting 2019-04-15 15:31:08 'call/mobile-data don't work' can be made a non-issue and even a feature as long as you have wifi working 2019-04-15 15:31:40 just carry a tiny cellular modem + wifi ap device 2019-04-15 15:31:56 and now the baseband is on a completely separate and worthless device where it can do no harm 2019-04-15 15:32:37 it seems that calling at least partially works for the nexus 5 https://wiki.postmarketos.org/wiki/Devices 2019-04-15 15:32:55 dalias^ 2019-04-15 15:33:19 can probably get a used unit on the cheap 2019-04-15 15:33:42 get some parts for it in advance 2019-04-15 16:22:10 is there a 3.9 testing repo, or just edge? 2019-04-15 16:33:17 ahills: binary packages for testing in numbered releases are usually not available 2019-04-15 16:34:34 there are binary packages for edge/testing and sources for 3.9/testing 2019-04-15 16:49:24 ahills: on the mirrors (binary packages) testing is only in edge 2019-04-15 16:50:14 does alpine have support for locales? 2019-04-15 16:50:21 jn__: danieli: thank you 2019-04-15 16:51:25 if no, I'll just whitelist it in cdist so that __locale works nonetheless 2019-04-15 16:53:36 telmich: no, musl libc doesn't supports locale 2019-04-15 16:53:43 nice 2019-04-15 16:53:48 one things less to worry about 2019-04-15 16:54:51 I think I answered same question this morning to someone else ;) 2019-04-15 16:55:15 it's a bit hard to find - there are some github issues, but did not see an answer in the wiki 2019-04-15 16:55:26 maybe I should keep musl link ready 2019-04-15 16:57:04 it is here: http://www.etalabs.net/compare_libcs.html 2019-04-15 16:57:44 I think locale in libc is one of the most wrong features 2019-04-15 16:57:58 or at least for anything that is used in the shell 2019-04-15 17:01:45 I am not sure if I found a bug in stat of busybox 2019-04-15 17:02:15 yep...seems like 2019-04-15 17:03:18 So %F should be supported and is according to the help, but using it I get an error message 2019-04-15 17:04:13 should this https://www.nico.schottelius.org/temp/alpine-busybox-stat be reported to alpine or busybox team? 2019-04-15 17:09:46 i'd say busybox, unless we've patched it out 2019-04-15 17:09:48 sorry for the noise, I cannot read error messages 2019-04-15 17:09:58 usually stat uses -f or --printf 2019-04-15 17:10:05 busybox seems to use -c 2019-04-15 17:11:35 *test* 2019-04-15 17:14:33 quite funny: https://github.com/ungleich/cdist/blob/master/cdist/conf/type/__file/explorer/stat 2019-04-15 18:52:13 So, a short feedback: it took me 1 day to migrate our production dns servers managed by cdist from devuan/bind9 to alpine/knot and we gained synthetic ipv6 record support. Quite a nice experience 2019-04-15 19:41:52 is anybody using an ocsp resonder based on Alpine Linux? I didn't find any in the package browser. OpenCA would probably be a good choice as it's lightweight and can be used based on CRLs 2019-04-16 00:20:53 how do I use apk index to create an index with a number of packages larger than the linux command line maximum length? 2019-04-16 00:23:12 ddevault: you could create the index file manually 2019-04-16 00:23:44 how? 2019-04-16 00:23:58 how are the official repos maintained? 2019-04-16 00:24:16 not sure, that's done upstream on the master mirror iirc (or on the builders?) 2019-04-16 00:26:13 iirc the APKINDEX.tar.gz is nothing more than the APKINDEX, DESCRIPTION, and the signature of the unsigned APKINDEX.tar.gz 2019-04-16 00:26:27 aye, how is the APKINDEX generated 2019-04-16 00:26:40 iirc abuild does that for you 2019-04-16 00:26:54 only if you have all of the packages you want included built and ready to add locally 2019-04-16 00:27:36 there's also abuild-sign you can use to sign the apkindex.tar.gz 2019-04-16 00:27:49 aye, but that's no help if I can't make an APKINDEX -_- 2019-04-16 00:28:40 and reading through abuild, it looks like it just calls abuild index *.apk, which is going to hit the same problem 2019-04-16 00:29:03 hmm, I'm not sure how you'd want to create an apkindex of existing apks 2019-04-16 00:29:35 I just want to build a single package and add it to a repository without having an entire copy of that repository on localhost 2019-04-16 00:29:55 right now I do this by shooting the package up to the mirror and running apk index _there_ and passing it every apk file to include 2019-04-16 00:30:06 but this is breaking now because I have more packages than the maximum length of a linux command line 2019-04-16 00:30:19 I have far fewer packages than, say, community, however. So I'm curious how it's dealt with upstream 2019-04-16 00:30:31 should probably ask in #alpine-devel about how they do it, I'm not entirely sure 2019-04-16 10:10:47 morning 2019-04-16 10:11:09 i have a question about alpine-linux and raspberry pi 2019-04-16 10:11:34 i have installed alpine 3.9.3 on a raspberry pi 3 B+. Everything works fine 2019-04-16 10:12:07 but at each reboot, i no longer have any packages installed (diskless mode) 2019-04-16 10:12:47 but if i do a : apk update ; apk upgrade it will reinstall all my packages 2019-04-16 10:13:23 do i miss something ? packages list seems to be saved but not reinstalled at boot 2019-04-16 10:13:39 (and of course, i use lbu_commit) 2019-04-16 10:25:04 ok i think i found the "problem". I need to create/enable the apk cache on my sdcard 2019-04-16 10:26:20 pea, check wiki for lbu 2019-04-16 10:26:58 and yes you will probably need apk cache 2019-04-16 11:09:50 other than the raspberry pi what other single board computers have people been successfully using with alpine? 2019-04-16 11:19:30 Invader__Bork: I'm using it on Allwinner A10 and A20, exynos chromebook and rk3399 chromebook 2019-04-16 11:29:43 exynos? wasn't that a samsung cpu? 2019-04-16 11:30:40 right 2019-04-16 11:31:19 oh it's an actual samsung chromebook. cool did not know they made stuff like this 2019-04-16 11:31:46 good price too 2019-04-16 11:32:23 is anything tegra based out there? for some more graphical umph 2019-04-16 11:32:30 exynos is old one, 32bit ARM 2019-04-16 11:33:36 tegra and musl doesn't fit together, afaik 2019-04-16 11:33:58 is exynos the one you were keeping on fbdev? 2019-04-16 11:34:38 what a shame regarding tegra and musl. i dont think there are any other tegra equivalents out there are they? 2019-04-16 11:34:48 yes, simply don't have reason to try to build mali drivers 2019-04-16 11:36:50 are you familiar with the jetson nano devboard 2019-04-16 11:36:51 and on rk3399 (also samsung chromebook) 64bit ARM I don't have strong reasons to build GPU drivers although there are some but not mainlined yet 2019-04-16 11:36:51 ? 2019-04-16 11:37:23 I heard for it but don't know anything about it 2019-04-16 11:37:32 Invader__Bork: sadly those various ARM chromebooks have quite a few drawbacks on anything that isn't chromeos (limited acceleration, for starters) 2019-04-16 11:38:05 I looked at one because of the rather nice battery life 2019-04-16 11:39:34 TomJepp: I don't watch movies or play games so for me it is not problem, even is good because GPU don't drain battery :) 2019-04-16 11:40:07 mps: what about browsing? 2019-04-16 11:40:20 and do the GPUs get stuck in a high power state without drivers? some desktop x86 GPUs do :( 2019-04-16 11:41:02 Browsing is quite fine using firefox built with rust 2019-04-16 11:41:24 It was slow in previous versions, before rust 2019-04-16 11:42:05 I always want to like small ARM devices but I always get fed p with the lack of CPU power. hoping the newer generations resolve that 2019-04-16 11:42:10 *fed up 2019-04-16 11:42:14 and, I don't keep big number of tabs active, about ten usually 2019-04-16 11:43:19 main problem for me is they don't have sata interfaces for ssd disks 2019-04-16 11:44:05 SD cards are not fast although some emmc are a lot faster 2019-04-16 11:48:03 the jetson nano is $99... i need to look into vulkan support and the musl problems you mentioned 2019-04-16 11:49:09 mps, apparently has an m.2 connector https://www.youtube.com/watch?v=2JkQwu_LF3k @ 2:55 2019-04-16 11:49:25 it is not musl problem, to be clear 2019-04-16 11:50:00 Invader__Bork: pity it's a bit pricey. if you don't need the GPU it's too much I think. If you *do* want the GPU then it's probably a decent place to be 2019-04-16 11:50:06 it is the problem that upstream thinks that only glibc works on linux 2019-04-16 11:51:10 or they don't care 2019-04-16 11:52:22 Invader__Bork: about sata I'm talking about chromebooks, yes some have sata but they are rare 2019-04-16 11:53:07 mps where's the incompatibility with musl exactly? the drivers? doesn't it compile a shim like normal nvidia drivers? 2019-04-16 11:53:42 nvidia is closed source and their binary is glibc only 2019-04-16 11:54:22 don't they compile a shim between the kernel and the bin blob to comply with the gpl? 2019-04-16 11:55:11 bin blobs are glibc linked 2019-04-16 11:55:25 TomJepp, i am specifically looking for something with a decent gpu. was hoping for more than 4 gb of ram. maybe there is a way to extend it 2019-04-16 11:55:45 Invader__Bork: to be honest, I think you'll end up on x86 2019-04-16 11:55:54 and on x86 your hardware will be better supported 2019-04-16 11:56:00 something like an AMD Ryzen APU? 2019-04-16 11:56:22 mps, how does the proprietary nvidia driver work on desktop alpine? 2019-04-16 11:56:37 doesn't work at all 2019-04-16 11:57:17 not only ARM but also not work on x86_64 2019-04-16 11:57:23 TomJepp, i was looking into arm because i wanted something that could be portable 2019-04-16 11:57:56 with something like 8k-12k mAh battery i think i could get decent mileage out of it 2019-04-16 11:58:39 Invader__Bork: yes, that is one of reasons I switched to ARM and chromebooks 2019-04-16 11:59:16 it is not fast as x86_64 but is fast enough for my use cases 2019-04-16 11:59:36 and battery lasts long 2019-04-16 12:00:13 my current 'moving computer' is less than 1kg 2019-04-16 12:00:53 and works nearly 10 hours without power supply 2019-04-16 12:01:09 mps, regarding speed: i wanted to reimplement some of my stuff with compute shaders and i think that i could offset some of the lost computing power into that 2019-04-16 12:01:13 in normal use cases, of course 2019-04-16 12:01:28 that's why i am specifically looking for something that could run vulkan 2019-04-16 12:02:09 you can look at colabora site, they have a lot if info about this 2019-04-16 12:02:41 ie^W 2019-04-16 12:02:53 meh, if^W 2019-04-16 12:03:04 if^W? 2019-04-16 12:03:29 ^W means delete previous word 2019-04-16 12:03:47 so delete if from previous sentence 2019-04-16 12:03:55 got it 2019-04-16 12:04:20 i just do a s/abc/xyz 2019-04-16 12:04:53 i see that we have a linux-firmware-nvidia package in the repos. isn't that the proprietary driver? 2019-04-16 12:05:14 yes, also I do, but looks like I'm becoming lazy 2019-04-16 12:05:49 well, this is just firmware for kernel driver, not for userspace 2019-04-16 12:06:30 and I thinks is closed source as most firmware are 2019-04-16 12:09:09 i am not oblivious of kernel drivers and userspace, but i don't understand how this affects this situation 2019-04-16 12:10:20 what does the nvidia kernel driver offer? 2019-04-16 12:10:31 and what would the userspace components offer? 2019-04-16 12:10:44 firmware is loaded to adapter during initialization of the hardware 2019-04-16 12:11:30 firmware is program which runs on hardware adapter CPU 2019-04-16 12:12:13 it is independent of the OS which runs on machine, usually 2019-04-16 12:12:59 kernel drivers runs in kernel space 2019-04-16 12:13:45 and user space drivers runs in user space programs, gpu driver for X window for example 2019-04-16 12:14:10 this is simplified explanation, of course 2019-04-16 12:16:08 so basically with linux-firmware-nvidia the kernel knows about the hardware and is able to initialize it but i am not able to get gpu acceleration to work in my software because those need the userspace components 2019-04-16 12:21:21 well, yes 2019-04-16 12:22:16 actually kernel 'knows' for hardware by using appropriate kernel driver 2019-04-16 12:22:32 i suppose that with nouveau i can't hope to get any performance levels close to the proprietary ones 2019-04-16 12:23:02 and then it can upload firmware for adapter if there is firmware and adapter need firmware 2019-04-16 12:24:22 if you want to crypto valute minning better use some other glibc based distro :) 2019-04-16 12:24:37 it's not for crypto 2019-04-16 12:25:09 did you notice ':)' at the end of my message 2019-04-16 12:26:05 yes i did, but it can be interpreted many ways 2019-04-16 12:26:22 i didn't get offended don't worry haha 2019-04-16 12:26:22 it is meant to say if you need GPU to work at it full power 2019-04-16 12:27:20 yea, but there are many reasons i ended up with alpine 2019-04-16 12:27:24 and afaik, crypto minning requires full GPU power, but I'm not interested in such things 2019-04-16 12:27:49 after going trough artix, obarun, gentoo and freebsd 2019-04-16 12:27:57 then you can try with some AMD GPU's 2019-04-16 12:28:41 I think they have open source GPU driver, but don't know for sure 2019-04-16 12:29:05 yes amd has opensource drivers in the kernel 2019-04-16 12:29:44 some say that it performs even better than the proprietary ones in some use cases 2019-04-16 12:30:15 I don't need fast GPU's for my work and usage so didn't give much time to this field so my knowledge and experience with this is low 2019-04-16 12:31:01 but the reason i am discussing nvidia is because the spare laptop i can install on has nvidia gpus and the only beefy mobile arm based gpu i know of is the tegra one 2019-04-16 12:32:25 that laptop isn't the end of the world because i only use it for testing, but when it comes to the mobile gpu i don't think there is a good one with decent support 2019-04-16 12:32:33 besides tegra 2019-04-16 12:32:54 I have such notebook with nvidia and intel gpu's. I built kernel drivers for nvidia but didn't manage to do anything with userspace 2019-04-16 12:33:45 so, I simply disabled nvidia and use intel 2019-04-16 12:33:57 probably more battery life too 2019-04-16 12:34:22 yes, and not to much warm 2019-04-16 12:34:59 s/warm/hot/ 2019-04-16 12:35:27 bah what do i do, i am back to distro hopping again 2019-04-16 12:35:42 i don't really want to move on from alpine 2019-04-16 12:35:47 four or five years ago I decided to not bye any x86 computer 2019-04-16 12:36:01 it ticks all my boxes but one 2019-04-16 12:36:13 oh wow, that's a long time 2019-04-16 12:36:16 anymore, to add at the end 2019-04-16 12:37:30 I'm quite satisfied (even happy) with arm's, but waiting for risc-v 2019-04-16 12:37:56 arm is at foundations also risc, at the end 2019-04-16 12:39:16 risc-v is holly grail, i am waiting for it impatiently 2019-04-16 12:40:02 i've seen a relatively low performance but kind of expensive risc-v sbc 2019-04-16 12:40:16 aha, I'm right behind you in the row for it :) 2019-04-16 12:40:18 sooo it's getting there, almost 2019-04-16 12:40:57 ask ddevault here or on #alpine-devel, he ported alpine to it 2019-04-16 12:42:18 the sifive one? 2019-04-16 12:42:29 I think so 2019-04-16 12:42:44 he have one 2019-04-16 12:43:06 or, has, never will learn English ;) 2019-04-16 12:44:25 i don't think it has a gpu though 2019-04-16 12:44:53 how's risc-v compared to arm in power consumption though? 2019-04-16 12:45:12 don't know, really 2019-04-16 12:50:13 are you aware by any chance what's the state of the opensource mali drivers? 2019-04-16 12:51:46 huh are the proprietary mali divers actually open source?! 2019-04-16 12:51:48 do you know if the rock64 works well with alpine ? I don't find interesting informations about this arm board 2019-04-16 12:51:50 https://developer.arm.com/tools-and-software/graphics-and-gaming/mali-drivers/utgard-kernel 2019-04-16 12:51:53 I think for 400 it is opened and there are works for 860 and 760 (or it is 740) work in progress 2019-04-16 12:52:40 search for panfrost and mali860 2019-04-16 12:52:51 oh it's not, they say lower on the page that it is not the complete driver stack 2019-04-16 12:52:54 I tried it on my chromebook 2019-04-16 12:53:35 but didn't built X driver for it 2019-04-16 12:54:32 there is on gitlab both kernel and userspace panfrost but it is not yet ready for distributions 2019-04-16 12:56:02 have you heard anything about the performance of the driver compared to the stock one? 2019-04-16 12:56:34 no 2019-04-16 12:57:11 i wonder if it's a bad idea to try to develop my own pcb with a bunch of mali gpus stuck to it 2019-04-16 12:57:58 if you have time you can try for fun 2019-04-16 12:59:11 one man's quest not to distro-hop haha 2019-04-16 12:59:35 mmm i can just picture it 4 mali gpus 32 gb of ram mmmm 2019-04-16 12:59:47 imo, arm ecosystem is actually swamp, but for now can't find anything better and usable 2019-04-16 13:01:16 paradoxically, mps, I feel x86 is more open than ARM 2019-04-16 13:01:38 in terms of practical usable hardware 2019-04-16 13:01:51 i wonder if arm gives access to the full driver stack just for prototyping, or you have to order like 50k units at a minimum 2019-04-16 13:02:20 TomJepp: well, we are comparing two swamps, imo :) 2019-04-16 13:02:43 becaue i'll be like: arm rep: how many would you like to order?; me: 4; arm rep: 4k? me: no, just 4 2019-04-16 13:02:45 haha 2019-04-16 13:08:38 is there an alpine conference? 2019-04-16 13:08:45 <_ikke_> Not yet 2019-04-16 13:09:13 <_ikke_> THere was an informal meeting at fossdem 2 years ago 2019-04-16 13:09:56 fossdem is in europe, right? 2019-04-16 13:10:27 <_ikke_> yes, Belgium 2019-04-16 13:11:13 is there a discussion about doing meetings regularly? 2019-04-16 13:12:21 <_ikke_> Not that I'm aware of 2019-04-16 15:06:33 Invader__Bork: https://www.collabora.com/news-and-blog/blog/2019/03/13/an-overview-of-the-panfrost-driver/ 2019-04-16 15:13:17 mps, ty 2019-04-16 18:42:08 mps, think i might get the nanopi m4 4gb 2019-04-16 18:42:42 did anyone work with it? 2019-04-16 18:48:31 what is it? a version of RPi? 2019-04-16 18:50:46 nope 2019-04-16 18:51:02 it's one of the others that have pi in their name like rock pi and banana pi 2019-04-16 18:51:03 https://www.friendlyarm.com/index.php?route=product/product&product_id=234 2019-04-16 18:51:22 Mali-T864 GPU, Dual-Core Cortex-A72(up to 2.0GHz) + Quad-Core Cortex-A53(up to 1.5GHz) 2019-04-16 18:51:29 and 2 or 4 gb ram 2019-04-16 18:52:40 https://www.youtube.com/watch?v=knS854Taz-E 2019-04-16 18:53:16 I see, it is somewhat like my chromebook, i.e. same CPU same SOC and same amount of RAM, and similar peripherals 2019-04-16 18:54:00 haha really? 2019-04-16 18:54:02 with a lot lower price 2019-04-16 18:54:09 which one of the two chromebooks 2019-04-16 18:54:10 ? 2019-04-16 18:54:17 arm64 2019-04-16 18:54:27 rk3399 2019-04-16 18:55:07 well with the chromebook you get some other peripherals 2019-04-16 18:55:48 https://www.samsung.com/us/computing/chromebooks/12-14/xe513c24-k01us-xe513c24-k01us/ 2019-04-16 18:55:50 Invader__Bork: for like $10 more you can get the T4 which has emmc 2019-04-16 18:56:12 iggy, T4 is that the rockpi? 2019-04-16 18:56:38 it's nanopi as well 2019-04-16 18:56:58 err, nanopc, but same difference 2019-04-16 18:57:06 same company anyway 2019-04-16 18:59:49 <_ikke_> iggy: I merged your PR btw 2019-04-16 19:00:26 iggy, the m4 has emmc as well 2019-04-16 19:00:35 https://youtu.be/knS854Taz-E?t=253 2019-04-16 19:04:01 _ikke_: oh, thank you! Didn't mean to drop off mid-conversation about it, but the GF had plans for me, and I figured I'd get back to it eventually 2019-04-16 19:04:24 <_ikke_> sure, np 2019-04-16 19:05:30 I'm going to try to do a better job of keeping future work broken out better from the start 2019-04-16 19:06:52 <_ikke_> I hope the feedback helps to make it easier for future PRs 2019-04-16 19:07:52 yeah, I just got over-excited 2019-04-16 19:08:19 was thinking about some sort of linter to check the easy stuff before submitting a PR 2019-04-16 19:08:59 or a pre-commit hook or something 2019-04-16 19:09:21 <_ikke_> iggy: I'm working on the side on some linting in our CI system 2019-04-16 19:09:52 great minds... 2019-04-16 19:31:33 does the alpine rpi version include the graphics stack? 2019-04-16 19:31:43 or does it suffer from the same lack of drivers 2019-04-16 20:19:05 My chroumium cant open any page... even not settings.. after upgrade 2019-04-16 20:19:24 Ive try to fix it... reinstall it.. but nothing works 2019-04-16 20:19:31 Why is that happend? 2019-04-16 20:20:53 I just got "something went wrong while displaying this webpage" error 2019-04-16 20:21:21 No matter im try to open new page or go to chromium downloads. Settings..history whatever 2019-04-16 20:23:37 Any help? 2019-04-16 20:24:01 <_ikke_> Probably issues with the latest chromium version on Alpine / musl 2019-04-16 20:26:56 does nobody test the new version before pushing updated package? :( 2019-04-16 20:27:19 <_ikke_> I expect so 2019-04-16 20:27:31 Can i downgrade it ? 2019-04-16 20:28:52 over7head, should be possible unless you were depending on a version in edge that got replaced 2019-04-16 20:29:25 dalias how i can try ? 2019-04-16 20:30:12 Apk --help doesnt notice antyhing about downgrading package 2019-04-16 20:34:04 apk add 'package or you can manually download the apk file you want 2019-04-16 20:34:46 <_ikke_> that assumes that specific version is available in a repository somewhere 2019-04-16 20:35:21 Thanks... but i wasnt able go to wiki easy without browser 2019-04-16 20:35:28 I got firefox... 2019-04-16 20:35:57 Btw does anybodie have problems with kmod? 2019-04-16 20:36:37 I built my kernel and it rewies a couple of reboots until modules were loaded... 2019-04-16 20:36:56 But now i want update kernel and cant get modules loaded anymore 2019-04-16 20:37:51 Old modules loads but new doesnt... they are both in /lib/modules.... i tryied reboot 5 times but nothing 2019-04-16 20:37:54 _ikke_, like i said, it's only likely to be possible if the desired version is in a release not edge 2019-04-16 20:38:06 <_ikke_> dalias: nod 2019-04-16 20:38:26 i wish edge would keep like last 3 versions or something on the servers 2019-04-16 20:38:37 for cases like this where a package is broken and you want to manually downgrade 2019-04-16 20:39:03 <_ikke_> dalias: we are working on some solution to store older versions of packages 2019-04-16 20:39:18 ideally there'd be a slow archival server with all past versions for forensic purposes if a package is ever compromised 2019-04-16 20:39:39 <_ikke_> dalias: linode sponsored some resources we might use for that 2019-04-16 20:39:46 :) 2019-04-16 20:54:58 Invader__Bork: https://www.youtube.com/watch?v=6BwR6KUDRkU 2019-04-16 20:55:40 they demonstrate panfrost on same chromebook as I have :) 2019-04-16 20:56:51 maybe I will try to build these drivers for Alpine, now I see how it works 2019-04-16 20:58:43 ~/.bashrc changes doesnt work for me 2019-04-16 20:59:02 even ~/.ashrc not working 2019-04-16 20:59:19 for both user and root 2019-04-16 21:00:23 <_ikke_> you want to use ~/.profile of ash 2019-04-16 21:02:10 <_ikke_> for ash* 2019-04-16 21:03:59 LS_COLORS=$LS_COLORS:'di=1;5;33:' ; export LS_COLORS 2019-04-16 21:04:10 either not works in ~/.profile 2019-04-16 21:06:53 doesnt matter ive set color in ~/.Xdefaults, but no blinking text...not important 2019-04-16 21:07:46 does anyboide else use alpine for desktop ? 2019-04-16 21:07:58 ive read its for servers and firewalls 2019-04-16 21:08:08 over7head: yes 2019-04-16 21:08:18 i use it for desktop... its very fast with dwm 2019-04-16 21:08:25 like it 2019-04-16 21:08:36 one of my best binary distro 2019-04-16 21:08:45 <_ikke_> over7head: .profile works for me but only for a login shell 2019-04-16 21:08:46 and it is fast with awesome wm, and with xfce also 2019-04-16 21:09:51 awesome is nice 2019-04-16 21:10:11 <_ikke_> I'm using it for years now, very content with it 2019-04-16 21:10:17 i hate dwm but im keeping use it 2019-04-16 21:11:15 have awesome wm on gentoo... it runs so fluid there 2019-04-16 21:12:36 gentoo, void and alpine are my best distros 2019-04-16 21:14:47 over7head: awesome is packaged in Alpine, if you didn't noticed it 2019-04-16 21:15:17 and i cant foreget slackware when talking about best distros :) 2019-04-16 21:15:58 yes, slackware was my first serious linux distro 2019-04-16 21:16:08 mps i didnt notice it...i alredy have it when installing alpine, so i wanted to try something new 2019-04-16 21:16:28 ah, sorry then 2019-04-16 21:16:38 mps, ty for the video, i'll watch it in a few mins 2019-04-16 21:17:07 Invader__Bork: I'm downloading panfrost kernel right now 2019-04-16 21:17:16 i3 is also good... 2019-04-16 21:17:19 let me know how building the drivers went 2019-04-16 21:17:39 oh nice, so you are doing it like right now haha 2019-04-16 21:17:46 but will not try to build it now, maybe tomorrow if I find some time 2019-04-16 21:17:56 oh so not right now 2019-04-16 21:18:17 here is nearing midnight 2019-04-16 21:19:14 over7head: I prefer awesome because I can make 'applets' I need in lua easily, not to complicated language 2019-04-16 21:19:46 only text, no graphics or icons, though 2019-04-17 08:11:08 i'm trying to fix PHP Fatal error: Uncaught Error: Call to undefined function simplexml_load_string() 2019-04-17 08:11:23 <_ikke_> hendry: You are probably missing a module 2019-04-17 08:11:36 I think I have what I need installed with `apk add --update php-xml`, but how do I list the contents of this package? 2019-04-17 08:11:48 <_ikke_> php7-simplexml 2019-04-17 08:12:01 <_ikke_> https://pkgs.alpinelinux.org/contents?branch=edge&name=php7-simplexml&arch=x86_64&repo=community 2019-04-17 08:12:17 <_ikke_> the packages database is easiest when you haven't installed it yet 2019-04-17 08:12:47 <_ikke_> for installed packages, you can use apk info -L 2019-04-17 08:14:18 _ikke_: bingo, thank you php-simplexml is what I needed 2019-04-17 08:20:39 Hi, is someone successfully running apache2 on alpine 3.9.3? We get a segmentation fault on all web servers when starting apache2 with the ssl modules enabled 2019-04-17 09:12:41 ncopa: ^ 2019-04-17 09:18:11 I just installed Alpine 3.9.3 and setup an mariadb server but when i run service mariadb setup it doesn't create an database 2019-04-17 09:18:20 did something change or did i fuck up 2019-04-17 09:19:41 https://i.imgur.com/DoIAhfr.png 2019-04-17 09:40:22 nekocentral: I'm pretty sure it just creates users and stuff, not databases 2019-04-17 09:44:14 if i check the init script it runs mysql_install_db --user=mysql --rpm 2019-04-17 09:44:26 if i remove the --rpm it create the data folder 2019-04-17 09:44:34 but mariadb still doesnt start 2019-04-17 09:44:48 and the pre-install script in the package creates the user 2019-04-17 10:09:10 lets say i find a problem in the init script of mariadb, can i do an pull request for that or do i need to poke someone 2019-04-17 10:25:26 good day 2019-04-17 10:26:07 We are soon having the hack4glarus hackathon and Glarus, Switzerland. I am considering to add some alpine challenges to it. Would anyone of you be interested in joining? The website can be found on https://hack4glarus.ch 2019-04-17 10:33:57 nekocentral: I don't see what you couldn't, contributions are always welcome I guess 2019-04-17 11:15:23 <_j> Hi there 2019-04-17 11:17:14 <_j> I was just wondering, is it possible to boot alpine as an iso from GRUB? I've seen some information around but nothing conclusive on how to do it? I would be using it with an encrypted /boot (holds keys/kernels/initramfs' for my void/gentoo installations) and an unencrypted ext4 or fat partition for the apkcache/run-from-ram stuff 2019-04-17 11:50:39 ? 2019-04-17 11:52:35 Anybody? 2019-04-17 11:52:46 x|2: there are people here, yes 2019-04-17 11:52:51 x|2: welcome to irc 2019-04-17 11:52:56 feel free to ask a question 2019-04-17 11:54:06 i like Alpine. i want setup a wireless access point 2019-04-17 11:54:20 https://wiki.alpinelinux.org/wiki/How_to_setup_a_wireless_access_point#Setup_Encryption 2019-04-17 11:54:35 https://wiki.alpinelinux.org/wiki/Raspberry_Pi_3_-_Configuring_it_as_wireless_access_point_-AP_Mode 2019-04-17 11:55:31 i can not success 2019-04-17 11:56:03 i think something is wrong 2019-04-17 11:57:08 who can help me 2019-04-17 12:00:13 ? 2019-04-17 12:00:48 jn__: can you help me? 2019-04-17 12:01:50 no. it's been a long while since i last setup hostapd, and i'm going to be AFK now 2019-04-17 12:02:48 ok 2019-04-17 12:05:23 x|2 left to fast 2019-04-17 12:27:33 yay i been usefull for the first time in my life, and even made an patch 2019-04-17 12:35:18 You patched your life with how many additions? 2019-04-17 12:37:05 hehe 2019-04-17 12:38:23 was a small problem with the mariadb initd script that used a flag that is for internal use only, and for some reason it does not create the database in alpine 3.9.3, was able to find a work around which i put in the patch instead of the old flag 2019-04-17 12:44:13 Show it to the world 2019-04-17 12:50:36 already send patch in, its in patchworks 2019-04-17 12:51:20 rip patch 2019-04-17 12:51:25 :p 2019-04-17 12:51:29 :P 2019-04-17 12:52:06 hey atleast i followed the wiki haha 2019-04-17 12:53:59 Yes and you messed it up :p 2019-04-17 12:54:06 dammit 2019-04-17 12:54:22 ACTION goes back in cave 2019-04-17 12:55:40 o master cladmeter how do i unfuck it 2019-04-17 12:55:46 A cave won't help you 2019-04-17 12:55:54 haha 2019-04-17 12:55:55 you've awakened the beast 2019-04-17 12:56:03 It's a catastrophe 2019-04-17 12:56:26 the sarcasme is jumping of the screen 2019-04-17 12:56:33 Like notre dame 2019-04-17 12:56:46 and that confirmed the sarcasm 2019-04-17 12:57:21 I'll give you a hint 2019-04-17 12:58:03 o wait i forgot to update apport with a rel 2019-04-17 12:58:13 <_ikke_> \o/ 2019-04-17 12:58:26 <_ikke_> nekocentral: And while you're at it, also look at the commit subject 2019-04-17 12:58:52 let me send a V2 2019-04-17 12:59:04 It's normal to make mistakes, asks ikke :p 2019-04-17 12:59:14 damn shots fired 2019-04-17 12:59:22 Sorry... Couldn't resist 2019-04-17 12:59:23 <_ikke_> d'oh 2019-04-17 12:59:35 <_ikke_> >_< 2019-04-17 12:59:49 It's the Taiwan 🍻 2019-04-17 12:59:57 <_ikke_> Look at https://github.com/alpinelinux/aports/blob/master/.github/CONTRIBUTING.md#creating-a-pull-request-pr for details 2019-04-17 13:00:06 <_ikke_> ACTION grabs phone 2019-04-17 13:00:36 ikke, i blame the wiki 2019-04-17 13:01:07 i followed this 2019-04-17 13:01:07 https://wiki.alpinelinux.org/wiki/Creating_patches 2019-04-17 13:02:18 Fix the wiki 2019-04-17 13:03:03 i can replace the content of the wiki with the blob 2019-04-17 13:03:38 <_ikke_> That wiki describes the general patch workflow. Comitting happens before that 2019-04-17 13:03:43 Wiki doesn't mention the correct format 2019-04-17 13:04:06 A copy and paste should do 2019-04-17 13:07:20 I'll give a hug for anyone who updates it 2019-04-17 13:07:29 Even _ikke_ 2019-04-17 13:08:07 ♥️ 2019-04-17 13:09:58 i can add the format to it if i can edit it 2019-04-17 13:10:49 Thanks 2019-04-17 13:11:34 If it's a new account it could take few hours before you can post links 2019-04-17 13:12:34 GOD FUCK YOU capatcha 2019-04-17 13:12:39 i guess im not human xD 2019-04-17 13:15:26 Please mind your language 2019-04-17 13:15:33 It's not appreciated 2019-04-17 13:17:24 sorry clandmeter 2019-04-17 13:17:48 victory, only took 6 tries 2019-04-17 13:20:29 so instead of the pkgver i should change the pkgrel right because its not an update to mariadb 2019-04-17 13:20:54 <_ikke_> correct 2019-04-17 13:26:58 o masters 2019-04-17 13:27:12 https://i.imgur.com/RhUaOd3.png 2019-04-17 13:27:15 how is this? 2019-04-17 13:29:17 nekocentral: pasting to raw text paste service would be better, tpaste.us for example 2019-04-17 13:29:33 <_ikke_> commit messages are in present tense, not past tense 2019-04-17 13:31:38 http://tpaste.us/pWRE 2019-04-17 13:31:41 rewritten 2019-04-17 13:32:20 looks better 2019-04-17 13:32:32 good thank you 2019-04-17 13:32:43 but why not posted complete patch 2019-04-17 13:33:04 because i was in the git commit part 2019-04-17 13:34:01 fix missing parameter parameter in init script, instead too generic issue 2019-04-17 13:34:22 at the end, by sending patch you fix issue 2019-04-17 13:35:41 okay thank you 2019-04-17 13:36:59 http://tpaste.us/gMR5 2019-04-17 13:37:06 in this case i would write: fix missing --data parameter in init 2019-04-17 13:38:00 you have parameter word two times 2019-04-17 13:38:04 well i replaced an parameter that is marked for internal use with no documentation with the datadir parameter 2019-04-17 13:38:09 lol 2019-04-17 13:38:30 and this is why i have it checked out before i send it out 2019-04-17 13:38:43 i dont want to get laughed at again by clandmeter 2019-04-17 13:38:47 that's ok 2019-04-17 13:39:05 How would this be? 2019-04-17 13:39:20 whatever you do you risk to be laughed by someone, don't worry 2019-04-17 13:39:31 <_ikke_> mps: HAHAHAA 2019-04-17 13:39:35 <_ikke_> :P 2019-04-17 13:39:41 Fix replace undocumented parameter with datadir parameter 2019-04-17 13:39:56 or internal only parameter 2019-04-17 13:39:57 <_ikke_> why 2 verbs? 2019-04-17 13:40:01 <_ikke_> fix replace 2019-04-17 13:40:17 not sure, my brain is not working anymore 2019-04-17 13:40:17 keep in mind that first commit line will be subject of the mail 2019-04-17 13:40:23 yea 2019-04-17 13:40:43 let me look again 2019-04-17 13:41:07 "Replace internal use parameter with datadir parameter" 2019-04-17 13:41:15 add missing datadir parameter in init script 2019-04-17 13:42:07 oh, you replaced -rpm, then maybe replace sound better 2019-04-17 13:42:45 It looks like apache2 does not work with ssl modules since openssl 1.1.1. 2019-04-17 13:42:45 I always get a segmentation fault on starting apache2 services when ssl modules are enabled.Is this issue already known? 2019-04-17 13:42:48 but, don't worry, also for me is not easy to write good commit messages 2019-04-17 13:44:15 okay changed the title 2019-04-17 13:45:45 should be send now 2019-04-17 13:47:12 did i fuck this one up _ikke_ 2019-04-17 13:47:33 <_ikke_> A tiny bit ;) 2019-04-17 13:47:43 dammit what now 2019-04-17 13:48:04 <_ikke_> marking it as v2 (which is an option for send-email) 2019-04-17 13:48:11 <_ikke_> BUt no need to resend 2019-04-17 13:48:31 well it was a complete rewrite of title and an description 2019-04-17 13:48:36 and yea 2019-04-17 13:50:01 just do with them what you want i will keep an eye on my email and join IRC if my mobile connection is stable 2019-04-17 13:57:46 And I'm back from mobile 2019-04-17 13:58:47 So what is the best way to un f my patch mess except jumping off an cliff 2019-04-17 14:40:47 Is it possible to convert from edge back to release without getting into deep trouble? 2019-04-17 16:57:29 hello 2019-04-17 16:58:01 i didn't find any informations about supporting kernel 5.0 in alpine edge 2019-04-17 16:58:12 is alpine linux 'limited' to LTS kernel ? 2019-04-17 16:59:51 pea: yes, it is not limited but decided to use lts, at least for now 2019-04-17 16:59:57 pea: you can build your own kernel and use it directly. this is not alpine-specific 2019-04-17 17:00:42 alternatively, you can take the linux-vanilla APKBUILD and bump the version. this is probably not documented because very few people ever need to do that (they do it for everyone else) 2019-04-17 17:03:04 ok, thanks for the info :) 2019-04-17 17:03:47 i have read the changelog of 4.19.35 and there are backports from 5.0.8 for rockchip board, so maybe it will be good (when 4.19.35 will plan in edge) 2019-04-17 17:05:04 4.19.x will be update soon, I presume 2019-04-17 17:06:06 btw, do you url to see what is backported, I'm using rockchip 2019-04-17 17:06:53 nvm, it is announced on lwn.net 2019-04-17 17:11:43 mps: https://cdn.kernel.org/pub/linux/kernel/v4.x/ChangeLog-4.19.35 (and search for rockchip ) 2019-04-17 17:12:05 which board do you have ? I have a rock64 and i would like to test alpine on it 2019-04-17 17:15:37 I have rk3399, samsung chromebook one plus, I built mainline kernel for it, right now it is 5.0.7 2019-04-17 17:17:12 5.0.8 is out ;) 2019-04-17 17:17:16 does it work well ? 2019-04-17 17:18:17 didn't build it yet, will do probably this night, if I finish crystal lang for Alpine 2019-04-17 17:18:57 you mean, does 5.0.7 works well? 2019-04-17 17:19:01 yes :) 2019-04-17 17:19:37 have problem with suspend to ram, mmc didn't restored after wake-up properly 2019-04-17 17:46:56 ok 2019-04-17 18:39:27 s it possible to convert from edge back to release without getting into deep trouble? 2019-04-17 19:40:29 team alpine - need your help - trying to get 3.9.3 to run on my raspberry pi. Downloaded the ARMHF tar, unziped it, added files to my newly formatted Fat32 SD card as per the wiki. The pi gives me 7 blinks --> no Kernel.img ... the config.txt tells the pi to look - kernel=boot/vmlinuz-rpi . I've also tried moving the file to the root directory and changing the config.txt and it did nothing. any suggestions? 2019-04-17 19:40:50 its a Pi Zero W 2019-04-17 19:51:08 so im just noticing now... the vmlinuz and the rest of the files in the boot directory are named rpi2 while the config file calls for ones ending in rpi only... 2019-04-17 19:53:59 i tried changing the names... nothing. i tried changing the call out in the config.txt... nothing. 2019-04-17 19:57:40 i just looked at the armhfv7 image and it has the same nomenclature. im wondering if the regular armhf tar has the raspberry pi 2/3 files in it by accident 2019-04-17 20:00:46 i just went into the 3.9.2 armhf file and it has the proper files. 2019-04-17 20:01:07 so it looks like the ARMHF version of hte Raspberry Pi Tar is wrong. 2019-04-17 20:03:44 is this the proper forum to report this issue? or is there another place 2019-04-17 20:34:24 rpi-noob already known problem with 3.9.3 2019-04-17 20:34:41 hope it will be fixed within the next release 2019-04-17 21:19:02 awesome. thank you ! 2019-04-17 22:00:25 hi all 2019-04-17 22:01:27 i'm running an out of band tree. maintaining this has been swimmingly until now. I am running an apk upgrade and one of my custom packages (not originally from your aports) is reporting: "package mentioned in index not found (try 'apk update')" 2019-04-17 22:02:23 the package exists on the filesystem, and apperantly by that wording is in the index. I can manually apk add the two packages having this error fine, but apk upgrade won't pull it in from that error 2019-04-17 22:05:42 is there a way to keep kernel modules rather than them being removed when the kernel package gets upgraded? 2019-04-18 00:19:45 I'd like to download one of the Alpine releases to play with it as a webserver I would put in a Docker container, which one is good for me? I'm on this page: https://alpinelinux.org/downloads/ 2019-04-18 00:20:19 I downloaded the miniroot filesystem package but realized it was only a file system, not an ISO. So I couldn't use it right away. 2019-04-18 00:21:22 I guess I tried to be smart about it, I will try the standard one maybe. 2019-04-18 11:00:52 mps, are you in? wondering if you compiled the mali drivers, and if you did how it went 2019-04-18 11:07:07 Invader_Bork: you mean panfrost mali driver? 2019-04-18 11:07:26 mps, yes 2019-04-18 11:07:56 we talked about it a couple of days ago 2019-04-18 11:08:05 yes, kernel part works, but didn't tried userspace 2019-04-18 11:08:17 I remember ;) 2019-04-18 11:08:52 you did before the kernel part? 2019-04-18 11:08:55 I compiled mesa part but didn't made apk package 2019-04-18 11:09:03 or did you compile it as well after we spoke last 2019-04-18 11:09:34 you are right, just wanted to repeat this for you 2019-04-18 11:09:55 mps, got it 2019-04-18 11:10:11 I have to fix mmc card issue 2019-04-18 11:11:12 looks like I found possible source of problem, power button invoking suspend-to-ram could be issue 2019-04-18 11:12:03 if LID switch is used it works and by hand 'echo mem > /sys/power/state' works fine 2019-04-18 11:12:28 that is, tested only on kernel 5.0.8 2019-04-18 11:13:02 is kernel 5.x.x in testing? or did you build it yourself 2019-04-18 11:13:18 to be fair, I don't understand well how to use userspace mesa drivers 2019-04-18 11:13:40 I built it from mainline 2019-04-18 11:13:58 in Alpine kernel is version 4.19.x 2019-04-18 11:15:22 now I work on fixing some Alpine packages (gc, vim) and upgrading crystal lang, so not much time is left for panfrost 2019-04-18 11:15:57 and I have to make and test u-boot for chromebook, besides my daily work 2019-04-18 11:16:31 i was just curious about the progress, i am not in a hurry about panfrost, don't have any hardware to test it on anyways 2019-04-18 11:16:57 if i want to get alpine out of the vm and into metal i might just use the nvidia laptop with nouveau 2019-04-18 11:17:31 or get something amd apu based (that's why i was asking about kernel 5, iirc it was needed for the latest vega apus) 2019-04-18 11:17:47 'coming as the time, annihilator of the humans' - Krshna, Bhagavad Gita 2019-04-18 11:18:05 so no pressure about panfrost 2019-04-18 11:19:01 If I know how to package mesa upstream properly I could made apk 2019-04-18 11:19:44 but I will work on it for sure, just have to finish tasks which are 'in a row' 2019-04-18 11:23:50 awesome, looking forward to it 2019-04-18 11:24:54 also, I would like to have it on my arm boxes 2019-04-18 13:51:51 is there a way to undo the kill word command in ash? 2019-04-18 14:22:10 Hey, I never have used busybox(maybe in android, but I always asummed that the output of free -m is the same as htop and coreutils free -m. I feel stupid now, because I used free before installing a package and after, and my eyes rolled thinking there is some weird huge memmory leak. 2019-04-18 14:22:12 So 2019-04-18 14:22:19 what does count in the yes of busybox free utility 2019-04-18 14:22:22 as free ram 2019-04-18 14:22:28 *used 2019-04-18 14:38:50 ok free is actually an procps util sry 2019-04-19 01:09:31 is it true you guys are moving to discord soon 2019-04-19 01:11:30 it is not 2019-04-19 01:12:53 opal: no, there are plans to potentially have an official community-centric server (which means we'd have some degree of control as to how we're seen on the platform) 2019-04-19 01:13:06 absolutely no infrastructure or development conversations are going to be moved, whether or not the motion passes 2019-04-19 01:13:19 the short answer is "no" 2019-04-19 01:13:58 I've already spoken to kaniini about the extremely questionable way of communicating this, his response was "using discord in any capacity makes discord take ownership of the project", which I believe to be preposterous nonsense (and have not seen any evidence to the contrary from him since). 2019-04-19 01:14:55 ah ok thanks for the clarification 2019-04-19 01:15:14 yeah kaniini's way of putting it made it come across worse 2019-04-19 01:16:02 I'm certain he sees it that extremely 2019-04-19 01:16:06 that does not mean he's right, though. 2019-04-19 01:16:21 im aware to take kaniini's word with a grain of salt 2019-04-19 01:16:25 thank you for making sure it was correct (and finding out it wasn't) :) 2019-04-19 01:16:41 by no means am i supportive of discord at all but i understand the reasoning behind wanting a presence there 2019-04-19 01:16:55 my perspective is simple - there will be a presence no matter what 2019-04-19 01:17:02 making it official means that we can control how that presence is presented 2019-04-19 01:17:05 yeah 2019-04-19 01:17:34 further, as I've had to say a lot recently, "more better" - as long as a libre alternative is the primary channel (as irc is and will remain), supporting more things isn't bad 2019-04-19 01:17:51 (unless, of course, pleroma chat ends up being super awesome and we all move to it ;) ) 2019-04-19 01:25:39 i dont have high hopes in pleroma chat simply because i think its dangerous to offload so many utilities onto activitypub 2019-04-19 01:25:51 and transiently http+json 2019-04-19 01:25:59 er... transitively* 2019-04-19 01:27:43 i have to put my ideas aside for how i want chat to be, because im simply too busy with everything else 2019-04-19 01:28:01 well, I do mean any superior libre alternative 2019-04-19 01:28:12 my opinions on pleroma chat specifically are mostly ironic ones, thus the wink~ 2019-04-19 01:29:31 waiting for pleroma 1.0 to fix all my issues with that software :^) 2019-04-19 01:30:12 you mean GNU style: 1.3, the first stable one? :^) 2019-04-19 01:30:29 (note: I don't use pleroma or mastodon; if I was to launch my own fediverse instance (which I've considered), I'd probably use pleroma) 2019-04-19 01:32:18 i currently use pleroma but im looking to write my own ap implementation 2019-04-19 01:32:53 i just have software architecture nits to pick that pleroma wont be able to address... and it shouldnt make the effort to 2019-04-19 01:33:02 pleroma's good for general use 2019-04-19 01:33:28 im just one of those autistic unix extremist freaks who doesnt want js running in the browser 2019-04-19 01:33:47 and who wants everything to be a small interconnectable utility 2019-04-19 01:33:51 and who wants this and that 2019-04-19 01:34:28 people tell me "just contribute to existing software" but i thought the whole point was software diversity, get different perspectives toward the same goal 2019-04-19 01:34:32 well, you're dealing with a FOSS problem the FOSS way: make your own :) 2019-04-19 01:34:36 yeah 2019-04-19 01:34:43 and if it's better, I'll use it ^^ (assuming I ever do the thing) 2019-04-19 01:35:04 yeah, we will see how correct or not correct it is 2019-04-19 01:35:11 i particularly hate webapps, thats my thing 2019-04-19 01:35:31 im gonna write a desktop client for existing ap software first, because of that 2019-04-19 01:35:44 I hate relatively few things, since I realized that hate really sucks up a lot of my energy, and doesn't lead to anything productive; I prefer specific observations and specific solutions 2019-04-19 01:35:45 because my implementation will be nothing without a client to access it 2019-04-19 01:35:47 e.g "this sucks" 2019-04-19 01:35:55 "this is how to make it non-suck" 2019-04-19 01:36:05 (sometimes, that does mean writing my own, like I did for https://brpaste.xyz) 2019-04-19 01:36:22 yeah actually succeed in making alpine a reasonably healthy project again 2019-04-19 01:36:26 before pursuing that 2019-04-19 01:36:36 pleroma doesnt suck really (well it does but only because its in heavy development; im sure everyones doing their best to address the glaring issues i come up with but its easy to get mad) 2019-04-19 01:36:37 hey, I'm clearly trying ;) 2019-04-19 01:36:49 i've been following the mailing list, it is unfortunate that people are resistant 2019-04-19 01:36:55 I can pursue multiple things at once :) 2019-04-19 01:37:04 well, if people suggest specific improvements that make the proposal better, I'm all for it! 2019-04-19 01:37:21 for instance, forced rotation (which was my idea to begin with anyway) does have specific bad side effects that I overlooked 2019-04-19 01:37:22 my light dislike toward pleroma is exactly the same as my growing dislike toward alpine 2019-04-19 01:37:26 imagine if it did pass in that state! 2019-04-19 01:37:30 for alpine, it's the fact it's a dynamically-linked binary distro 2019-04-19 01:37:38 it serves its niche, just doesnt meet my tastes 2019-04-19 01:37:54 opal: one of the things I'm planning to do once it's easier to get access is to try and get every library package to have a -static version, with an install_if on -dev 2019-04-19 01:38:04 im still using it for desktop / vms / containers until i have the gentoo setup i want 2019-04-19 01:38:06 the other thing I would like to go over is make all of the openrc scripts modern and good 2019-04-19 01:38:19 neat 2019-04-19 01:38:23 don't hold your breath with gentoo, USE=static does not do as much as you think it does. 2019-04-19 01:38:28 best of luck on your ventures with that 2019-04-19 01:38:46 oh im not holding my breath lol 2019-04-19 01:38:47 (I moved to alpine *from* gentoo, for a variety of reasons, one of which was that the promise of USE-flags is ultimately hollow; "good enough") 2019-04-19 01:39:14 the idea of USE flags is that you can change exactly how your thing is built, to customize it to your needs exactly, instead of relying on what a maintainer/packager thought was right 2019-04-19 01:39:18 but that's not what happens in practice 2019-04-19 01:39:20 i want glibc on my system but static-linked with musl for anything i can, so that requires a nice careful balance of overlays and whatever other hacks there are to get it working 2019-04-19 01:39:33 in practice, the maintainer/packager decides what gets a USE flag and what doesn't 2019-04-19 01:39:39 the issue with use flags is probably the lack of care people put into their ebuilds 2019-04-19 01:39:45 yes 2019-04-19 01:39:58 the point of USE flags is to give people choice, rather than depend on the maintainer 2019-04-19 01:40:00 i'd be fine maintaining an overlay to make everything more consistent 2019-04-19 01:40:02 but the maintainer chooses the USE flags 2019-04-19 01:40:05 anyway, there are numerous irc channels on freenode that are dead because the proprietary alternative overtook it in practice 2019-04-19 01:40:08 . . . so the goal is RIP 2019-04-19 01:40:16 alpine will be no different, because the market of alpine is hacker news posting tech bros 2019-04-19 01:40:31 i look forward to the slack 'server' within 1 year 2019-04-19 01:40:32 there are numerous channels that go all but dead when the matrix relay gets cut, too 2019-04-19 01:40:34 kaniini: are you really going to go on another round of justifying your misleading statements that's making people think awful things? ;) 2019-04-19 01:40:37 people hate irc 2019-04-19 01:40:50 i hate irc but its "good enough" 2019-04-19 01:40:56 SpaceToast: i don't need to justify anything, it is what will happen 2019-04-19 01:41:00 I seem to be one of the like 5 people that *don't* hate irc 2019-04-19 01:41:17 ya, drew being another 2019-04-19 01:41:30 for what it's worth, there are reasons people are fleeing IRC 2019-04-19 01:41:39 mobile phone integration being one 2019-04-19 01:41:42 the need for a bouncer is another 2019-04-19 01:41:47 mobile phone works great for me ¯\_(ツ)_/¯ 2019-04-19 01:41:52 (which irccloud "fixes" but it's its own issue" 2019-04-19 01:41:53 but I do run a multi-tiered infrastructure for personal use 2019-04-19 01:41:56 s/"$/)/ 2019-04-19 01:41:56 SpaceToast: people want push notifications 2019-04-19 01:42:00 I have push notifications. 2019-04-19 01:42:03 go on? 2019-04-19 01:42:05 now, there are some ircv3 proposals to standardize that 2019-04-19 01:42:10 ios people want push notifs, SpaceToast 2019-04-19 01:42:10 but nobody cares ;) 2019-04-19 01:42:16 I don't know about ios 2019-04-19 01:42:19 ircv3 is a bandage 2019-04-19 01:42:19 I have push notifications on android 2019-04-19 01:42:29 opal: ircv3 is a mistake 2019-04-19 01:42:36 android's the only mobile os worth using at this point, i was tongue-in-cheek there 2019-04-19 01:42:50 anyway, I'm perfectly in favor of making a good libre alternative to irc 2019-04-19 01:42:55 and then moving to that 2019-04-19 01:42:57 my humour is harder to spot because im too lazy to type the ;) all the time 2019-04-19 01:42:59 the time wasted on trying to make irc competitive to the proprietary services could have been better spent creating something competitive to the proprietary services 2019-04-19 01:43:19 you can do both at the same time, you know ;) 2019-04-19 01:43:31 had i known what i know now, in 2005, i wouldn't have bothered with any of that work 2019-04-19 01:43:51 your only big mistake was moving chary to cxx 2019-04-19 01:44:12 we still run 3.x on my net 2019-04-19 01:44:23 4.x is not C++ 2019-04-19 01:44:25 but 2019-04-19 01:44:27 then what is 2019-04-19 01:44:33 i forget 2019-04-19 01:44:43 there is no official branch of charybdis that is C++ 2019-04-19 01:44:53 ah that was just an in-progress thing? 2019-04-19 01:44:58 never got seen through? 2019-04-19 01:45:00 yes 2019-04-19 01:45:03 gotcha 2019-04-19 01:45:23 but that's not my point 2019-04-19 01:45:40 had i known that irc would be threatened by slack and discord 15 years ago, i wouldn't have written any of this stuff 2019-04-19 01:45:50 hindsight's 20/20 2019-04-19 01:45:51 i would have just went straight to building something that could address that threat 2019-04-19 01:46:29 of course, part of this situation is also freenode's fault 2019-04-19 01:46:52 they haven't exactly moved to do anything to make freenode attractive verses slack or discord 2019-04-19 01:47:13 i even proposed something as trivial as having a webservice that generated badges that showed the # of users on the channel 2019-04-19 01:47:19 like slack and discord have ;) 2019-04-19 01:47:29 they were like 2019-04-19 01:47:30 naw 2019-04-19 01:47:34 we don't see any reason to have it 2019-04-19 01:47:51 so to be clear, i don't blame alpine for moving on with the times 2019-04-19 01:48:00 irc has a lot of friction and resistance to change 2019-04-19 01:48:03 for any network not just fn 2019-04-19 01:48:16 i am just frustrated with the fact that now we have to play serious catch up 2019-04-19 01:48:18 its what causes the fragmentation 2019-04-19 01:48:22 the discord is not an official part of the project by any stretch 2019-04-19 01:48:41 it will become one over time 2019-04-19 01:48:47 convenience is very compelling 2019-04-19 01:49:05 but alpine itself is obsolete anyway 2019-04-19 01:49:18 pedaling back a bit, SpaceToast do you think it'd be worth my time addressing gentoo's flaws or just starting from scratch and developing my own distribution? because i had plans to do so for a while... just wanted to give gentoo a fair shot first 2019-04-19 01:49:19 stuff like nix will be the technology of the future 2019-04-19 01:49:32 if it really isnt worth it then i'd like to know, since you have experience with gentoo 2019-04-19 01:49:42 oh right nixtemd 2019-04-19 01:49:44 convenience is indeed compelling, but i am 99% certain there are other people who also have very strong opinions on it 2019-04-19 01:49:54 opal: if your goal is "glibc default, static-musl as much as possible", you want portage, but you do not want the main portage tree 2019-04-19 01:49:55 hurd is the future 2019-04-19 01:49:56 opal: honestly, what is needed is something like nix but with stuff like musl and openrc ;) 2019-04-19 01:50:16 danieli: i doubt it, i'm generally the one with the strong opinions about proprietary services around here 2019-04-19 01:50:20 SpaceToast: ah 2019-04-19 01:50:27 you might get ddevault to go on about how discord is meme but that's about it 2019-04-19 01:50:35 have you looked into runsv kaniini 2019-04-19 01:50:48 portage has lots of flaws (like any package manager; ESPECIALLY source-based and extendible) 2019-04-19 01:50:53 it isnt perfect ofc but it seems like a better design 2019-04-19 01:50:53 but I think those can be fixed iteratively 2019-04-19 01:50:57 opal: i'd rather go with s6 than runit tbh 2019-04-19 01:51:01 I've already suggested a few to someone that might be able to implement them :) 2019-04-19 01:51:13 I haven't looked at s6 yet, but done lots of stuff with runit 2019-04-19 01:51:18 does s6 deal with forced doublefork any better? 2019-04-19 01:51:28 because that's the real problem with runit 2019-04-19 01:51:36 yes 2019-04-19 01:51:36 (and most daemontools-likes) 2019-04-19 01:51:39 ok, how? 2019-04-19 01:51:57 see, i treat that as a problem with the daemon itself rather than with the init service 2019-04-19 01:52:04 has a wrapper for such things 2019-04-19 01:52:12 and how does the wrapper work...? 2019-04-19 01:52:18 monitors the pidfile 2019-04-19 01:52:21 so... 2019-04-19 01:52:25 i believe programs should log to stdout and not disown their parents 2019-04-19 01:52:25 it's openrc, but less integrated 2019-04-19 01:52:38 ACTION shrugs 2019-04-19 01:52:42 did i say runsv i meant runit 2019-04-19 01:52:45 opal: that's the "correct way :tm:", but in reality, not everyone does things the "correct way :tm:" 2019-04-19 01:52:47 god my words today 2019-04-19 01:52:50 yeah we figured it out :) 2019-04-19 01:53:22 you can tell my experience with runit is limited; i have one system using it (voidlinux) and i dont generally touch it 2019-04-19 01:53:29 set stuff up, it works, i forget it 2019-04-19 01:53:39 watching twitch on it now 2019-04-19 01:53:43 I hate void 2019-04-19 01:53:48 or, to be more specific 2019-04-19 01:53:51 works fine enough on my laptop 2019-04-19 01:53:51 I hate their package manager :) 2019-04-19 01:53:55 oh yeah 2019-04-19 01:54:03 its like pacman but bad 2019-04-19 01:54:07 anyway, the distributions of the future are ones where you have a declarative configuration and the tools sync the system image to that configuration 2019-04-19 01:54:07 one of the big reasons I switched to alpine is because apk is *awesome* 2019-04-19 01:54:18 yeah well 2019-04-19 01:54:23 kaniini: have you used nix for an extended period of time? 2019-04-19 01:54:27 yeah apk's terrific... thats one of the things i wanted to emulate by creating a distro 2019-04-19 01:54:29 SpaceToast: yes, it sucks 2019-04-19 01:54:32 yes, it does ;) 2019-04-19 01:54:34 but the underlying idea is solid 2019-04-19 01:54:37 a package manager that can match the ease of use of apk 2019-04-19 01:54:38 I also know someone that contributes to it 2019-04-19 01:54:43 what we need is an 'apk' that works in that way 2019-04-19 01:54:45 apk isn't just easy to use 2019-04-19 01:54:48 apk is also fast 2019-04-19 01:54:51 AND apk is featureful 2019-04-19 01:54:53 yes 2019-04-19 01:54:58 WAY more featureful than MOST PACKAGE MANAGERS 2019-04-19 01:55:06 i know 2019-04-19 01:55:09 I told the gentoo folks about cmd: and so: virtuals 2019-04-19 01:55:14 recently 2019-04-19 01:55:16 it blew their minds 2019-04-19 01:55:18 you're telling me about things i designed 2019-04-19 01:55:26 kaniini: another idea of mine was deployable system configurations; perhaps i could look into declarative config some more and see how that could tie in with my ideas 2019-04-19 01:55:31 congrats on making something better, instead of trying to remove things that are not better ;) 2019-04-19 01:55:56 i like putting good ideas together 2019-04-19 01:56:02 declarative is a meme, imo 2019-04-19 01:56:08 it really isn't 2019-04-19 01:56:15 /etc/apk/world is declarative 2019-04-19 01:56:15 the ability to declare a system policy is very powerful 2019-04-19 01:56:18 precisely 2019-04-19 01:56:26 but it's application side 2019-04-19 01:56:26 god i love that about apk 2019-04-19 01:56:28 not user side 2019-04-19 01:56:37 users can edit the world file (same as in gentoo) 2019-04-19 01:56:40 SpaceToast: yes, and i am proposing doing that with the entire system state 2019-04-19 01:56:48 but you're SUPPOSED to do it otherwise 2019-04-19 01:56:51 the world file/dependency is singlehandedly the best thing that came out of apk 2019-04-19 01:57:03 opal: gentoo's had a world file for a VERY long time 2019-04-19 01:57:16 yeah world file isn't new 2019-04-19 01:57:18 not sure which came first, but I'd imagine gentoo has been; though their impl is a bit different 2019-04-19 01:57:24 no i know but its new to me 2019-04-19 01:57:27 :) 2019-04-19 01:57:29 having come from debian / arch to alpine 2019-04-19 01:57:40 the real trick is apk -t 2019-04-19 01:58:09 yeah, user-defined virtuals 2019-04-19 01:58:11 (by the way i have scripts that generate 'policy' packages, but they are quite ugly) 2019-04-19 01:58:27 anyway 2019-04-19 01:58:35 declarative is a meme because it is not the most efficient way to approach the problem 2019-04-19 01:58:42 the system backend can be declarative, but the frontend shouldn't be 2019-04-19 01:58:58 similarly, see ansible and co 2019-04-19 01:59:01 the frontend should allow importing the configuration, that's all i ask for ;) 2019-04-19 01:59:10 oh, that I agree with 2019-04-19 01:59:19 I just don't think squishing in the "muh declarable conf" meme is the right way to go at it 2019-04-19 01:59:30 importing/syncing configurations is an obvious step forward 2019-04-19 01:59:41 one thing i'd like to add to package management is the ability to have per-user packages; that way you have the consistency of the system package manager and can avoid language-specific ""package management"" bullshit without being root 2019-04-19 01:59:45 I'm just being careful to avoid clear undesirable states 2019-04-19 01:59:58 (such as nixos (at least, undesirable to me), and "windows group policy") 2019-04-19 02:00:28 yeah portable configuration / base installs are something i want to go for 2019-04-19 02:00:31 SpaceToast: danieli: anyway, don't take my criticism personally. i understand why things are as they are. all i ask is that you don't bullshit me ;) 2019-04-19 02:00:42 I don't bullshit you 2019-04-19 02:00:48 and I don't take criticisms of ideas personally 2019-04-19 02:01:01 I present ideas, and invite you to defeat them, or present better ones 2019-04-19 02:01:09 when you fail, I come to the conclusion that your ideas suck and you're wrong ;) 2019-04-19 02:01:13 when you succeed, I adopt them 2019-04-19 02:01:16 we all know that alpine irc is (a) already in decline, and (b) will decline faster because of alternatives 2019-04-19 02:01:17 ;) 2019-04-19 02:01:19 you've seen both sides of that coin now :) 2019-04-19 02:01:36 I disagree with a, and think you've failed to provide any serious evidence of b 2019-04-19 02:01:57 i hope it declines faster because of alternatives, but i want the alternatives to be good 2019-04-19 02:02:02 I've seen irc usage increase consistently within the last 4 months or so (when I started paying attention to it) 2019-04-19 02:02:08 i think thats an unanimous opinion i share 2019-04-19 02:02:21 -devel even goes too fast for me 2019-04-19 02:02:26 opal: people moving over to X from Y suggests that X is better than Y 2019-04-19 02:02:32 in the ways that the people moving care about 2019-04-19 02:02:38 yeah 2019-04-19 02:02:42 thus the obvious: make a FOSS thing that's even better :D 2019-04-19 02:03:17 the people with good architectural decisions usually happen to be the ones with bad ui/ux judgment 2019-04-19 02:03:20 which is unfortunate 2019-04-19 02:03:22 anyway, my plan is to write some sort of lightweight C implementation of the federated chat feature in pleroma (which isn't http-based but instead websocket based) 2019-04-19 02:03:38 you use phoenix in pleroma, right? 2019-04-19 02:03:42 yes, make tox 2.0 2019-04-19 02:03:44 channels are coming along nicely 2019-04-19 02:03:44 kind of 2019-04-19 02:03:45 it'll go really well, I swear ;) 2019-04-19 02:03:58 god tox is concentrated shit 2019-04-19 02:04:03 tried using it witha friend 2019-04-19 02:04:07 we decided on mumble 2019-04-19 02:04:10 all of the core devs up and disappeared 2019-04-19 02:04:13 yeah.. I did some work on tox a few years ago, it really was terrible 2019-04-19 02:04:21 . . . and so no one was left that knew how it works 2019-04-19 02:04:24 i was talking about server implementation 2019-04-19 02:04:29 for people who just want chat 2019-04-19 02:04:31 > decentralized 2019-04-19 02:04:33 > server 2019-04-19 02:04:34 and don't want a database or anything 2019-04-19 02:04:41 decentralised ≠ distributed 2019-04-19 02:04:48 yes, tox uses bootstrap servers 2019-04-19 02:04:50 SpaceToast: pleroma is distributed, not decentralized 2019-04-19 02:04:54 pleroma is 2019-04-19 02:04:57 I don't know if chat should be 2019-04-19 02:05:09 what value is there in having distributed chat accounts? 2019-04-19 02:05:15 same handle 2019-04-19 02:05:26 you get that with both centralized and decentralized 2019-04-19 02:05:34 it's a shame xmpp is so old and ugly 2019-04-19 02:05:43 there's a visible advantage for distributed social networks 2019-04-19 02:05:49 for chat I'm not seeing the same advantages 2019-04-19 02:05:50 xmpp's mostly ugly because the few people who create clients do a poor job at it 2019-04-19 02:05:53 danieli: the goal is to build a modern alternative to xmpp ;) 2019-04-19 02:05:59 because group policies shouldn't be a thing for chat one-on-ones 2019-04-19 02:06:01 conversations is the only app that supports omemo sanely 2019-04-19 02:06:05 and group policies for group chats should be per group 2019-04-19 02:06:08 not "per serrver" 2019-04-19 02:06:15 omemo isnt too sane 2019-04-19 02:06:17 beats otr though 2019-04-19 02:06:19 opal: not 'ugly' visually, i was aiming at everything from the protocol to the servers 2019-04-19 02:06:28 and +1 to that 2019-04-19 02:06:38 but you cant argue that xmpp is unappealing from a ux perspective 2019-04-19 02:06:49 you can make any ux you like for xmpp 2019-04-19 02:06:49 thats a huge reason not to want to use it 2019-04-19 02:06:52 as xmpp is a protocol 2019-04-19 02:06:53 yeah but people havent 2019-04-19 02:06:54 the main reason federated systems work is because they can be demonstrated easily 2019-04-19 02:06:59 same for email, people havent 2019-04-19 02:06:59 because xmpp as a protocol sucks 2019-04-19 02:07:03 people use gmail web interface 2019-04-19 02:07:11 people have, but people that care about UX don't care about good tech 2019-04-19 02:07:15 so you end up with shit like mailspring 2019-04-19 02:07:16 exactly 2019-04-19 02:07:19 same goes for mumble - I started working on an elixir implementation of murmur (without the wonky voice crypto) together with a modern client to go with it, but I haven't had much time lately :( 2019-04-19 02:07:20 awesome UI, can't even plaintext 2019-04-19 02:07:30 heh 2019-04-19 02:07:40 danieli: anyway, to answer your question, "sort of" 2019-04-19 02:07:41 god the fucking amp for email 2019-04-19 02:07:47 i hope that doesnt take off, i sincerely hope 2019-04-19 02:07:56 danieli: we use pleroma, but we change a lot of the behaviour 2019-04-19 02:07:57 ofc gmail is probably going to integrate some type of support for it 2019-04-19 02:07:59 opal: amp is so terrible, and if google keeps pushing, it probably will 2019-04-19 02:08:04 lol amp 2019-04-19 02:08:09 ya thats what im worried about 2019-04-19 02:08:12 "just add this tag and we'll make your thing faster" 2019-04-19 02:08:16 I call that extortion. 2019-04-19 02:08:18 kaniini: you're aiming at the question of whether you use phoenix, right? 2019-04-19 02:08:19 idk about you all 2019-04-19 02:08:27 SpaceToast: it is 2019-04-19 02:08:29 danieli: for example, pleroma has it's own cowboy2 handler instead of using phoenix's directly 2019-04-19 02:08:33 i agree 100% 2019-04-19 02:08:35 google's backdoored its way into everyone's tech for years now 2019-04-19 02:08:39 its scary 2019-04-19 02:08:47 captcha, analytics, now amp 2019-04-19 02:08:48 kaniini: i would do the same thing if i needed to, it's a bit painful to work around phoenix at times 2019-04-19 02:09:00 danieli: well 2019-04-19 02:09:11 danieli: we have an elixir core dev and phoenix dev on payroll 2019-04-19 02:09:15 opal: thankfully, people like me exist, that don't care about captcha, because submitting user content is for [bleep], don't care about analytics, just read what's there, and don't care about amp, because it's stupid \o/ 2019-04-19 02:09:18 so maybe phoenix will get better 2019-04-19 02:09:20 that's pretty neat, who are they? 2019-04-19 02:09:20 ;) 2019-04-19 02:09:25 i selfhosted my @wowana.me email for a while now; just recently did i also move my irl shit away from gmail 2019-04-19 02:09:29 danieli: some russian dude 2019-04-19 02:09:37 goes by max 2019-04-19 02:09:37 I selfhost @toastin.space emails 2019-04-19 02:09:39 thank god yo uexist SpaceToast 2019-04-19 02:09:44 need more like you 2019-04-19 02:09:46 for development purposes only 2019-04-19 02:09:54 I might have to make some major changes soon though 2019-04-19 02:09:59 the only things still tied to gmail are school-related 2019-04-19 02:10:07 and im not in school so emails should trickle away over time 2019-04-19 02:10:07 because openbsd decided to drop openssl support for opensmtpd :( 2019-04-19 02:10:15 I use gmail for all of the personal stuff 2019-04-19 02:10:19 because I don't care :) 2019-04-19 02:10:22 hm, I'm not sure I know who that is 2019-04-19 02:10:32 the reality is that for reasons out of my control, google probably already knows everything about me 2019-04-19 02:10:47 and I don't want my government taxes account to get lost because my server died 2019-04-19 02:10:48 I've been in the erlang community for a good while, but haven't delved too deep into elixir until recently - but they're converging nicely 2019-04-19 02:11:00 elixir is gr8 8/8 m8 2019-04-19 02:11:20 yes 2019-04-19 02:11:22 elixir is great 2019-04-19 02:11:27 why don't we just fuck off with alpine 2019-04-19 02:11:31 and run elixir on bare metal 2019-04-19 02:11:36 familiar with GRiSP? 2019-04-19 02:11:44 at some point is pleroma going to stop needing mix and all that to run 2019-04-19 02:11:56 well, GRiSP and nerves 2019-04-19 02:12:01 kaniini: write a kernel in elixir that sounds like a cool idea 2019-04-19 02:12:02 https://www.grisp.org/ and https://github.com/nerves-project/nerves 2019-04-19 02:12:04 yes, but that was mostly a shitpost ;) 2019-04-19 02:12:12 (actually, not being sarcastic for once, I know it's hard to tell) 2019-04-19 02:12:26 OTP is very well suited for self-healing long-running processes 2019-04-19 02:12:30 i realized that it was a shitpost but not a super bad idea 2019-04-19 02:12:32 . . . what process runs longer than the kernel? 2019-04-19 02:12:42 what process needs to self heal more than the kernel? 2019-04-19 02:13:00 SpaceToast: you can use erlang/elixir without using OTP but then it's not much more than $language 2019-04-19 02:13:07 anyway, ultimately, i quit alpine because it was time to move on 2019-04-19 02:13:09 danieli: are you sure you can? 2019-04-19 02:13:12 well, except the cool core language features 2019-04-19 02:13:17 (although there's a lot about alpine that pisses me off) 2019-04-19 02:13:23 SpaceToast: there's escript for instance 2019-04-19 02:13:25 what I like about elixir is (most) of the syntax 2019-04-19 02:13:35 that's what I dislike the most about it 2019-04-19 02:13:44 i prefer erlang itself's syntax 2019-04-19 02:13:45 but it grew on me, i don't think it's fully as terrible as i used to 2019-04-19 02:13:46 kaniini: that 4chan thread about the """coc"""; i assume that wasnt one of the reasons 2019-04-19 02:13:51 well, my full taste is F# 2019-04-19 02:13:56 but it has a similar problem - the environment 2019-04-19 02:14:00 indeed 2019-04-19 02:14:01 i think i can chalk that down to idiots blowing shit out of proportion 2019-04-19 02:14:03 opal: it was, in some ways, but not for their conclusions 2019-04-19 02:14:07 ah 2019-04-19 02:14:16 opal: i saw that one, lol 2019-04-19 02:14:20 F# singlehandedly advanced the state of programming language research by 10 years, imo 2019-04-19 02:14:20 yeah that was uh 2019-04-19 02:14:26 i'd have replied if not for the captcha 2019-04-19 02:14:48 i decided to just leave it and let the cesspool boil dry 2019-04-19 02:14:55 yeah probably the best choice 2019-04-19 02:15:06 opal: the alpine CoC is basically a bad response that came extremely late for the problem that jirutka was running people out of the project 2019-04-19 02:15:17 I don't really agree with the entirety of the alpine CoC, because it's ^ 2019-04-19 02:15:33 CoCs generally seem like micromanagement to me 2019-04-19 02:15:35 thats my main issue 2019-04-19 02:15:36 but then again, I wrote my own CoC as a joke response to linux adopting the contributor covenant 2019-04-19 02:15:38 now 2019-04-19 02:15:44 . . . and then people took it seriously and think it's the best thing ever 2019-04-19 02:15:51 thankfully, I got over that existential crisis 2019-04-19 02:15:58 i have some ~theories~ as to why that response happened 2019-04-19 02:16:00 "but first, let me take a selfie" 2019-04-19 02:16:00 on the extreme end, we have political outsiders backdooring themselves into foss projects by means of CoCs 2019-04-19 02:16:20 they mainly have to do with ncopa's desire to protect jirutka while appearing to take action because jirutka was doing a lot of the work 2019-04-19 02:16:54 was? 2019-04-19 02:16:55 because usually, you know 2019-04-19 02:17:05 I did notice rust isn't really getting the updates it needs 2019-04-19 02:17:14 the response to somebody running other senior devs out of a project 2019-04-19 02:17:29 is to eject the dev doing that 2019-04-19 02:17:33 but yeah CoCs are micromanagement; even i know how to act in developement circles, and if i dont act accordingly then i can expect repercussion 2019-04-19 02:17:33 and apologize profusely 2019-04-19 02:17:54 like, thats unspoken 2019-04-19 02:18:03 well, the reality is that a lot of people *don't* know what "appropriate" is 2019-04-19 02:18:14 imo the goal of a CoC isn't to actively use it to enforce things 2019-04-19 02:18:24 it's to let people with no social intelligence know what's expected of them 2019-04-19 02:18:25 appropriateness varies depending on the project lead and the surrounding community 2019-04-19 02:18:32 i don't really care about CoCs, but alpine's CoC is weird for that reason 2019-04-19 02:18:37 people with no social intelligence do not need to be contributing to projects 2019-04-19 02:18:49 SpaceToast: and yes, jirutka does not contribute much anymore, for reasons previously discussed 2019-04-19 02:18:55 social intelligence is as important as being able to code and use git 2019-04-19 02:19:04 (or whatever tools your project uses) 2019-04-19 02:19:17 ¯\_(ツ)_/¯ 2019-04-19 02:19:19 i have a feeling it happened both because one party considers the other a friend, in a sense, and because alpine needs all hands on deck to keep its nose above water 2019-04-19 02:19:22 if they can make the project better, why stop them? 2019-04-19 02:19:23 if people lack that, its a problem to be fixed elsewhere 2019-04-19 02:19:30 devs arent supposed to be teachers or parents 2019-04-19 02:19:33 just keep them from causing problems :) 2019-04-19 02:19:44 if they can't infer the rules, write them down, so they have no excuse 2019-04-19 02:19:45 keep them from causing problems by expelling them from the project swiftly 2019-04-19 02:19:55 project leads have that authority 2019-04-19 02:19:58 "expel after problems" is not keeping them from causing problems 2019-04-19 02:20:02 that's reaction TO a problem 2019-04-19 02:20:12 to keep them from causing problems, you have to give them the resources they need to not cause problems 2019-04-19 02:20:16 which is writing down things that they cannot infer 2019-04-19 02:20:18 CoCs also cause an equivalent amount of problems, just in a different manner 2019-04-19 02:20:20 danieli: yes, but handling it that way sent a bad message to other prospective contributors 2019-04-19 02:20:30 the main problem CoCs cause are in how they are enforced 2019-04-19 02:20:34 CoCs should not be enforced 2019-04-19 02:20:36 and of course, jirutka continued to run devs off until finally the CoC was enforced 2019-04-19 02:20:42 CoCs should be documentation, effectively 2019-04-19 02:20:47 which is how mine was written 2019-04-19 02:21:06 and an agreement that he could keep his commit access if he did not interact with others was put in place 2019-04-19 02:21:09 im never implementing social guidelines for any of my projects, even if they happen to gain traction 2019-04-19 02:21:35 opal: well i think "don't start screaming at contributors in reviews" is a reasonable guideline 2019-04-19 02:21:35 I have full intent in remaining BDFL for all of my current projects and doing as I please 2019-04-19 02:21:42 but none of my current projects are distros 2019-04-19 02:21:50 yeah but an unspoken one kaniini 2019-04-19 02:21:51 nor "huge software" 2019-04-19 02:21:59 mostly because I don't *make* "huge software" 2019-04-19 02:22:06 if someone screams at me i'll tell them to fuck off and block them from interacting with the development process 2019-04-19 02:22:09 I make nice self-contained utilities/wrappers :) 2019-04-19 02:22:25 sure 2019-04-19 02:22:33 opal: if a person with a broken leg is being slow in a corridor in front of you, you push them aside and make sure they learn not to walk in front of you slowly? ;) 2019-04-19 02:22:37 i wont give them room to harass me or any community i have jurisdiction over 2019-04-19 02:22:37 pleroma is probably the largest undertaking i've ever been involved in 2019-04-19 02:22:43 ya fuck cripples 2019-04-19 02:22:50 SpaceToast: but in seriousness 2019-04-19 02:22:53 those arent comparable 2019-04-19 02:22:55 well, I'll have to consider this aspect of your personality childish, then :) 2019-04-19 02:23:05 it is - people with no social intelligence aren't like that because of their choices 2019-04-19 02:23:10 no shit, i have no reason to be a boring adult 2019-04-19 02:23:15 it's not a normal state of things 2019-04-19 02:23:17 im childish if the opportunity arises to be 2019-04-19 02:23:31 like the birth rate dropping during christmas, and rising right before 2019-04-19 02:23:40 adulthood is overrated 2019-04-19 02:23:46 it is very much comparable to a physical disability, and should be taken as such. 2019-04-19 02:24:00 kaniini: maturing is the process of getting this whole "theory of mind" thing 2019-04-19 02:24:02 if i own a gym, and someone with a broken leg comes in to go on the treadmill 2019-04-19 02:24:03 most don't really get there 2019-04-19 02:24:10 ...thats what we're talking about right now 2019-04-19 02:24:12 it doesnt compute 2019-04-19 02:24:19 they shouldnt be at the gym going to the treadmill in the first place 2019-04-19 02:24:25 im gonna ask what the fuck theyre doing 2019-04-19 02:24:29 ah, so people with a broken leg shouldn't be allowed to exercise however they wish :) 2019-04-19 02:24:45 lol 2019-04-19 02:24:50 and no, if we follow your analogy, you would ban them from the gym permanently while screaming 2019-04-19 02:25:07 yeah thats why i should start a gym 2019-04-19 02:25:38 wasnt the perfect analogy but i cant come up with anything better right now 2019-04-19 02:26:00 point is, it isnt my job to teach people who dont know better 2019-04-19 02:26:09 you don't have to be the one to write the CoC 2019-04-19 02:26:19 someone else that does want to help people who don't know better does 2019-04-19 02:26:26 you also don't have to enforce it, since it's just documentation 2019-04-19 02:26:39 fine, but keep the social guidelines separate from my repository 2019-04-19 02:26:45 ¯\_(ツ)_/¯ 2019-04-19 02:26:50 as-is the alpine ones are just on a.o 2019-04-19 02:26:55 which is fine, imo 2019-04-19 02:26:58 you can educate people without doing it in the context of a software project 2019-04-19 02:26:59 could also be in docs.a.o, probably will be 2019-04-19 02:27:12 however, it requires that your software project have permanent storage facilities 2019-04-19 02:27:12 all CoCs agree more or less, right? 2019-04-19 02:27:14 outside of the git repo 2019-04-19 02:27:18 no, not at all 2019-04-19 02:27:35 as a reminder, I wrote my CoC as a direct response to SEVERELY disagreeing with the Contributor's Covenant 2019-04-19 02:27:50 i mean the same undertone can be applied to all: lurk, get a feel for the development environment, act accordingly 2019-04-19 02:27:53 to the degree that trying to patch it wasn't worth it :) 2019-04-19 02:27:54 im not talking about anomalies 2019-04-19 02:28:08 "get a feel for the development environment" -> not everyone can do this 2019-04-19 02:28:15 then they should not be contributing 2019-04-19 02:28:26 like you say, "keep it out of the context of the software project" - why should anything outside of their technical ability to improve the software project matter? 2019-04-19 02:28:45 why should a cripple avoid participating in a charity marathon? 2019-04-19 02:29:06 ever hear of terry fox? ;) 2019-04-19 02:29:35 ok i found a better response to your "broken leg" / "cripple" shit 2019-04-19 02:29:41 in what case were those people accomodated 2019-04-19 02:29:48 like explicitly accomodated 2019-04-19 02:30:07 youre drawing false similarities here 2019-04-19 02:30:45 anyway past that, i feel this conversation has started to go in circles 2019-04-19 02:30:52 gonna bbl 2019-04-19 02:30:57 havent eaten much 2019-04-19 02:33:50 the marathon didn't close, even though the runner is clearly in last place and far behind 2019-04-19 02:33:53 that in itself is accomodoation 2019-04-19 02:34:07 in fact, a much bigger one than having a single text document in your repo 2019-04-19 02:34:18 it actually required the people accommodating to sit there and wait. 2019-04-19 02:34:42 all so that they could participate fully 2019-04-19 02:34:47 something you are in a hurry to do the opposite of 2019-04-19 04:28:35 pls no discord 2019-04-19 04:35:43 again: 1) not an official part of the project, 2) alpine isn't moving there 2019-04-19 07:43:45 I would like to install alpine using virt-install. The output is stuck when using --cdrom to access the ISO. virsh console [alpine-vm] does not work either. Any ideas? 2019-04-19 10:28:53 Hello everybody, I hope I'm asking in the correct place. I'm wondering what's the best way to run my own custom kernel (already compiled) with alpine. The only guide I found is here https://wiki.alpinelinux.org/wiki/Custom_Kernel however I was wondering if it is possible to avoid recompiling inside of abuild and use my own linux repository to compile. Thanks! 2019-04-19 10:31:23 jakob1: yes 2019-04-19 10:38:01 mps: is there a guide on how to do that? Or can you elaborate please? 2019-04-19 10:43:51 it depends how you want to install it 2019-04-19 10:45:00 id you build it in the machine where you want to use it, then follow README in kernel source tree 2019-04-19 10:45:12 s/id/if/ 2019-04-19 10:46:55 if you build it on other box, then are different ways to do that but again follow guide in mainline kernel, or preferred you find on net 2019-04-19 10:47:32 so if I understand correctly this is independent from alpine? Then thanks for your help! 2019-04-19 10:48:52 right, Alpine will run on a lot of different kernel version but you have to install and set it 'by hand' 2019-04-19 10:50:01 I usually do 'make tarxz-pkg' and copy resulting tarball to destination box, unpack it, and set boot loader parameters 2019-04-19 10:50:24 u-boot mostly but syslinux or grub sometimes 2019-04-19 10:51:08 I understand, thanks! Does it require any specifics? I'm planning on replacing the kernel in 3.9.3 with a 4.14.0. Any reasons this might not work? 2019-04-19 10:55:32 don't know why it wouldn't work. I have 4.14.xx on one ARM box with v3.9 Alpine 2019-04-19 10:55:53 thanks! 2019-04-19 10:56:40 I don't see a reason to use abuild to use a custom kernel, but maybe it simplifies things for some ppl? To me at least it was just confusing 2019-04-19 10:57:16 also I don't use abuild for custom kernels 2019-04-19 10:58:03 especially because I cross compile kernels for different arch's and use cases 2019-04-19 11:03:03 hi, I had a lot of of alpinelinux vm's running on KVM without problems, even running edge version. Some days ago I migrated some to a ThreadRipper 2950x processor, and I'm experimenting random kernel panics after some hours running 2019-04-19 11:03:41 donoban: does it happen with any other operating systems? 2019-04-19 11:03:49 also, hi! I think I remember you. :) 2019-04-19 11:04:21 hi!, yes I used to enter here some years ago 2019-04-19 11:04:30 ha, it's been that long already? 2019-04-19 11:04:43 you were here around the time I joined 2019-04-19 11:04:58 I think that it was near alpine 3.3 2019-04-19 11:06:12 2015/2016... time is pretty fast 2019-04-19 11:06:42 I don't think it's been that long for me, but I honestly don't remember 2019-04-19 11:07:05 hehe 2019-04-19 11:07:31 well, I don't tested other OS with KVM, but tested a lot of different configurations with this VM 2019-04-19 11:07:55 it'd be nice if you could test with other distros, if it happens with them, it *could* be some hardware problem manifesting in a strange way 2019-04-19 11:08:15 all had this random crashes, but the VM process itself seems stable. it reboots and works fine for another X hours 2019-04-19 11:08:32 got a screenshot/paste of a sample kernel panic? 2019-04-19 11:08:36 I have one crash log, most of the times it crashes it doesn't write anything to the console 2019-04-19 11:08:40 https://share.riseup.net/#OZ8iwCG4IG7TvZal-hcdMA 2019-04-19 11:08:44 it just freezes? 2019-04-19 11:08:55 sometimes it writes log someties it doesn't 2019-04-19 11:09:06 the last 5 I think it only wrote this log 2019-04-19 11:09:11 the other 4 just rebooted 2019-04-19 11:09:33 and I also catched other log days ago which I didn't save but I'm pretty sure it was different than this 2019-04-19 11:10:14 clandmeter: you're into hardware, are you aware of any specific issues with the ThreadRipper 2950x? 2019-04-19 11:10:59 I found something about threadripper, but almost related to panics on the host system 2019-04-19 11:11:49 the host is running centos7, I tested with kernel 3.X, 4.4.178 (currently), and 5.x... 2019-04-19 11:12:04 I think that the host was stable with all of them 2019-04-19 11:13:15 is running zfs on the host nearly impossible to be related with this? 2019-04-19 11:13:37 i'd say it's unlikely, the guest doesn't care 2019-04-19 11:14:20 it'd be nice to know if it's hardware or not, do you have resources enough to run some other distro for some hours just to see if it crashes too? 2019-04-19 11:14:40 also, just curious, what memory setup (physical) do you have on the host? 2019-04-19 11:14:42 oh yeah, there is plenty of ram and CPU 2019-04-19 11:15:03 it has 64gb but it also running some vmware vm's 2019-04-19 11:15:19 but I can install some additional VM 2019-04-19 11:15:29 I'm gonna try with debian 2019-04-19 11:18:32 'general protection fault' is an interrupt triggered by the cpu when there's an access violation 2019-04-19 11:20:07 if the panics occur at seemingly random times and caused by seemingly random processes, that together with GPFs suggests RAM is the culprit 2019-04-19 11:21:52 it's a bit strange if it only happens after you started using another (and fairly new) CPU, and only in KVM and not in guests running on other hypervisors 2019-04-19 11:22:41 also the other crash that I saw in the console 2019-04-19 11:23:05 wasn't an GPF, I don't remeber exactly what was but it has clearly "kernel panic" message 2019-04-19 11:23:12 maybe I did some photo.. 2019-04-19 11:23:36 if it happens every few hours, collect one or two if you have the opportunity 2019-04-19 11:24:24 the problem is that many of them doesn't log anything 2019-04-19 11:24:29 let me paste some examlpe 2019-04-19 11:24:35 they just freeze completely? 2019-04-19 11:25:59 it reboots, I had configured to reboot in 1 second 2019-04-19 11:26:13 I had the console redirected to a qemu serial device 2019-04-19 11:26:39 most of the time I see something like this 2019-04-19 11:26:40 https://share.riseup.net/#T4yW4Qs1fBoqJjQpe2gf2Q 2019-04-19 11:26:55 it booted fine, and suddenly it reboots 2019-04-19 11:27:11 without any rc message or kernel log 2019-04-19 11:27:21 nothing at all? 2019-04-19 11:27:28 did the hypervisor log anything about it? 2019-04-19 11:28:01 let me check but I think it doesn't 2019-04-19 11:28:48 8.5K -rw------- 1 root root 90K Apr 17 13:23 alpinelinux.log 2019-04-19 11:28:59 unmodified since 17/04 2019-04-19 11:32:20 if there is some hardware ram problem, it should expect crashes on the host right? 2019-04-19 11:32:47 I'd say so, but I've seen bad memory cause very strange errors 2019-04-19 11:32:49 s/it/I 2019-04-19 11:33:17 it's ECC memory, tested by the motherboard manufacturer 2019-04-19 11:33:37 well, I'm finishing debian installation 2019-04-19 11:34:11 altougth this is pretty problematic, since it reboots inmediatelly and boots pretty fast, I can survive until I found the problem 2019-04-19 11:34:28 I mean, the 'symptoms' suggests memory, but if it's ECC and it started happening after you changed to a new CPU, I don't really know what to think 2019-04-19 11:34:52 well I changed everyghint, new motherboard/memory/cpu/harddisk 2019-04-19 12:55:42 wops, maybe I got disconnected but not noticed 2019-04-19 12:56:30 danieli: I think that I'm gonna switch this alpinelinux VM to vmware, altought on the other server it runs perfectly for years, maybe there is some problem with kvm/vmware together 2019-04-19 12:56:52 see you later and thanks for trying to help me 2019-04-19 12:58:46 donoban: I'm sorry I couldn't help. for what it's worth, someone else might be able to 2019-04-19 13:04:32 Any guide how get ad auth working ? 2019-04-19 13:11:20 What do you guys think about introducing short arguments for apk? So instead of sudo apk update && sudo apk ugprade we just do sudo apk -Uu 2019-04-19 13:11:52 Introducing them as a an alternative 2019-04-19 13:14:13 aren't scripts good enough ? 2019-04-19 13:16:59 why should they be enough? 2019-04-19 13:47:19 I think I can, if I plan my day a little bit, find time for the characters saved that way 2019-04-19 14:14:31 Have I join the #archlinux channel by mistake? 2019-04-19 14:14:41 s/join/joined 2019-04-19 14:16:35 "-Uu" doesn't really say anything about what it does in my opinion 2019-04-19 14:16:57 "update" is three more characters, but it's also 100x more verbose and self-describing 2019-04-19 14:17:12 plus, it's cool to be different 2019-04-19 14:17:28 ty, danieli 2019-04-19 14:17:45 i just wanted a honest opinion on this 2019-04-19 14:17:52 you got it! 2019-04-19 14:18:01 s/honest/objective 2019-04-19 14:18:06 better phrased that way 2019-04-19 14:18:11 agreed 2019-04-19 14:23:51 let it be said as someone who wrote a lot of shell scripts for the last decade or so that as far as script maintainability goes, it'll always be long options for me 2019-04-19 14:24:37 a good --help page does help either way, but I'm against intentionally making arguments obscure in favor of saving a few characters 2019-04-19 14:25:00 I've worked a ridiculous amount of pacman flags into my muscle memory from looking them up so many times 2019-04-19 14:26:42 TBB, i get why verbose would be better in a shell script, you understand what's happening at a glance, less confusion, less need to memorize flags 2019-04-19 14:27:56 but i'm coming rom managing a desktop and granted it might also be coming from getting used to pacman 2019-04-19 14:28:21 i'm not in favor in short flags as an alternative just because pacman is using it though 2019-04-19 14:28:58 just faster and more graceful desktop management 2019-04-19 14:29:45 it gets annoying after a while to type the same long commands over and over again 2019-04-19 14:30:21 Why are you typing the same commands over and over ? write a simple shell script lol 2019-04-19 14:31:04 maxice8[m], it ads another layer 2019-04-19 14:32:00 so instead of writing one shellscript you want people to add lots of opts to apk ? 2019-04-19 14:32:29 i don't want people to add lots of opts to apk, i would have done that myself 2019-04-19 14:32:49 why are you people so antagonizing 2019-04-19 14:33:12 i only care about the objective pros and cons 2019-04-19 14:33:41 if there is no way to have a proper adult conversation on this issue then so be it 2019-04-19 14:34:13 <_ikke_> These things often are subjective 2019-04-19 14:34:53 even subjective is better than antagonizing 2019-04-19 14:45:48 why don't you use the command line history ctrl-r is your friend. 2019-04-19 14:48:24 i think people are not antagonizing, they just offer you alternatives to your idea. that still can be an adult conversation. 2019-04-19 14:48:43 it's not just about what you say, but how you say it :) 2019-04-19 14:51:45 it's irc. it will be always ambigous. if you want to be offended, you can always find a reason. 2019-04-19 14:52:05 it's better to assume no ill, makes everything much more chill 2019-04-19 14:54:02 ah, nevermind I said anything, I'm not in the mood to discuss 2019-04-19 14:54:15 have a nice friday, people! 2019-04-19 14:54:30 I've come to the conclusion I'm probably getting old as I feel I no longer have the time to reformat everything I write 8 times as is my practice with work related emails... 2019-04-19 14:55:21 danieli, we shouldn't really even be on IRC right now, at least over here the weather is absolutely fantastic, great weather for meeting friends and having a pint or six with friends :) 2019-04-19 14:55:34 springtime, finally! 2019-04-19 14:55:54 <_ikke_> The weather is very nice in NLD 2019-04-19 14:55:58 TBB: it's been fantastic here too for a few days, and yeah, finally a semblance of spring 2019-04-19 14:56:28 TBB: to me, it's more about common decency and having at least a hint of respect for other people 2019-04-19 14:56:37 oops, sorry for the double highlight 2019-04-19 14:56:55 so much lollygagging to compensate for lack of minimum conversation skills sheesh 2019-04-19 14:57:14 I agree, one of my observations over the years is, save sarcasm for those who you know well enough to know they can take it :) 2019-04-19 14:57:33 it's not really a good tool online 2019-04-19 14:57:49 <_ikke_> Invader_Bork: sorry, but so far you seem to be more antagonizing then everyone else 2019-04-19 14:58:00 my observation is that many people hide around "sarcasm" when they're lacking minimum conversation 2019-04-19 14:58:05 honestly, i have to agree with _ikke_ on this one 2019-04-19 14:58:17 _ikke_, ill try to chill then 2019-04-19 14:58:19 my bad 2019-04-19 14:58:21 either way, let's drop that and be nice 2019-04-19 14:58:24 thanks :) 2019-04-19 14:58:53 yup, I'll keep trying to avoid that tone in the future 2019-04-19 14:59:27 TBB, you're making it really difficult for me to chill 2019-04-19 14:59:31 :) 2019-04-19 14:59:51 why? he just said he'll avoid that tone 2019-04-19 14:59:57 I don't understand the problem here? 2019-04-19 15:00:04 that's a recurring theme in my life to be honest :D 2019-04-19 15:00:17 TBB: but ha, good point. using too much sarcasm was a bad habit of mine when i was younger, but it makes for a more confrontational and hostile conversation if it's used snidely 2019-04-19 15:00:32 additionally, it tends to convey very poorly online 2019-04-19 15:04:45 danieli, because he's deferring responsibility. from what he said earlier is implied that he does that because people can't take his "sarcasm", not because he did something wrong 2019-04-19 15:04:53 to add little to short/long options in CLI, I write shell completion if I use repetitive options 2019-04-19 15:05:20 Invader_Bork: sarcasm does convey poorly online, it's not always meant in a rude way 2019-04-19 15:06:01 danieli, i don't have anything against sarcastic remarks 2019-04-19 15:06:38 what i have said is that some people justify lack of conversation skills by claiming it was sarcastic 2019-04-19 15:06:47 the good old "its just a joke bro" 2019-04-19 15:07:06 anyways this has gone for far too long 2019-04-19 15:07:35 danieli: right, I like to be sarcastic sometimes, but learned long ago that it is not good on IRC because people cannot see our face and intonation when we 'speak' 2019-04-19 15:07:58 also mps, some people are really bad at sarcasm 2019-04-19 15:08:04 Invader_Bork: please consider being more civil and polite towards other people 2019-04-19 15:08:20 danieli, so i'm the bad guy now? 2019-04-19 15:08:59 bad guy or not, you're being rude 2019-04-19 15:09:16 I AM BEING RUDE? 2019-04-19 15:09:22 <_ikke_> Can we just calm down everyone 2019-04-19 15:09:24 <_ikke_> just stop 2019-04-19 15:11:49 <_ikke_> Thank you 2019-04-19 15:12:55 the problem with me personally is, I've been on IRC for about 25 years and we never used to have problems like this back in the day. over the years the way conversations have been had on IRC has kind of found its own format, which is different from face to face conversations 2019-04-19 15:15:45 Freenode's a bit different as far as IRC goes, as it's a kind of a semi-professional setting especially when projects of different sorts are why a channel exists 2019-04-19 17:43:13 fair point someone brought up; i had this issue for a while but havent actually gotten around to addressing it, but when linux kernel is upgraded through apk, new modules cannot be loaded until i boot into the new kernel. has this issue been addressed? 2019-04-19 17:43:33 i understand that with a distro like alpine, disk usage is a concern, so maybe thats the reason behind doing it this way. i dont know 2019-04-19 17:53:05 opal: this is the case on basically every distro like that 2019-04-19 17:53:12 you can do some weird workarounds using symlinks etc 2019-04-19 17:53:24 but usually, a kernel upgrade implies you should reboot to use the new kernel :) 2019-04-19 17:54:55 kk just verifying 2019-04-19 17:57:41 opal: afaik, when you upgrade kernel old one and it's modules are removed 2019-04-19 17:58:19 yes i know lol 2019-04-19 17:59:11 solution could be to 'learn' apk to keep previous one 2019-04-19 17:59:33 teach, is probably better term 2019-04-19 18:01:07 which could be option good to have, i.e. keep previous working version till we see if now could boot 2019-04-19 18:01:26 s/now/new/ 2019-04-19 18:09:44 i was mainly wondering if it was managed with a post-install hook or something 2019-04-19 18:09:53 hell, i can look at the apkbuild for linux 2019-04-19 18:12:24 looked, and there is no post or pre install for kernel 2019-04-19 18:13:15 I thought to add such option 2019-04-19 19:48:20 trigger? 2019-04-19 19:48:50 I am trying to install bird using `apk add bird@testing` but get "ERROR: unsatisfiable constraints:" which points to libraries in edge; what is the right way to install such a package? 2019-04-19 19:49:25 you probably need to install those using the same syntax, telmich 2019-04-19 19:50:07 TBB: I need to list all dependencies individually? 2019-04-19 19:50:39 unfortunately yes, because they're in a pinned repo 2019-04-19 19:50:42 that shouldn't be the case 2019-04-19 19:50:51 that's an unfortunate side effect of mixing stable and edge 2019-04-19 19:51:00 that sounds a bit ... cumbersome. can't I tell apk to "install this package and whatever is needed from @edge"? 2019-04-19 19:51:01 when you're installing something explicitly in a tag, everything in that tag should be considered 2019-04-19 19:51:11 telmich: you have @testing and @edge separately? 2019-04-19 19:51:15 because that could do it 2019-04-19 19:51:20 SpaceToast: let me verify, but I think I do 2019-04-19 19:51:29 ok, can you try setting them to be the same? 2019-04-19 19:52:06 sure... need to checkout the syntax for that, just a sec 2019-04-19 19:52:44 you just use the same tag for each 2019-04-19 19:52:47 :) 2019-04-19 19:53:31 so I put @testing in front of both of them? 2019-04-19 19:53:45 yes 2019-04-19 19:53:48 or @edge in front of both 2019-04-19 19:54:07 note that another side effect you might have is incompatibilities between the stable and edge libraries 2019-04-19 19:54:22 e.g something needing a specific library version from non-edge, but being (yet) incompatible with the newer edge one 2019-04-19 19:54:33 it's usually easier to just switch to edge entirely, but switching back is quite hard 2019-04-19 19:54:52 that also doesn't work 2019-04-19 19:55:04 let me put it all in one document 2019-04-19 20:01:29 https://www.nico.schottelius.org/temp/apk-bird.md 2019-04-19 20:09:27 you know, if you're going to .md something, it'd be nice to actually display it ^^;; 2019-04-19 20:09:42 could you throw that up on https://brpaste.xyz? 2019-04-19 20:09:57 (or any pastebin really, I link that one by default since I wrote it :) ) 2019-04-20 07:49:29 good morning to all! anyone installed Alpine on Linode cloud? what are experiences 2019-04-20 08:06:31 Hey. I was just checking out the scripts in here: https://github.com/alpinelinux/alpine-conf and I have a really noob question. Why are the shell scripts in there have an extension of ".in" ? and not ".sh" ? 2019-04-20 08:13:12 m00n_urn: they are 'input' (source) files which are converted to .sh during build 2019-04-20 08:15:45 mps, Ah, I see. So, the Makefile takes all these input files together and builds it? 2019-04-20 08:27:59 that is how it works, usually 2019-04-20 08:37:17 cool, thanks :) 2019-04-20 12:00:50 How to install Alpine Linux in diskless mode while saving the configs and apk cache in a different medium(say, an SD card) than the installation medium(a usb drive)? 2019-04-20 12:22:33 How to install Alpine Linux in diskless mode while saving the configs and apk cache in a different medium(say, an SD card) than the installation medium(a usb drive)? 2019-04-20 12:37:08 mondalex: it'll ask you where you want to store them 2019-04-20 12:37:24 it should work with two different media 2019-04-20 12:38:34 darkfader: it only gives the option of usb, floppy, cdrom 2019-04-20 12:39:04 No SD card 2019-04-20 12:44:01 @darkfader 2019-04-20 12:44:22 i'm searching if you just need to load the mmcblk driver early 2019-04-20 12:44:28 but i think it's also missing from the script 2019-04-20 12:44:54 if you want to move on, get a second usb drive 2019-04-20 12:54:47 darkfader: mmcblk also missing from setup-alpine i.e. setup-disk 2019-04-20 12:56:23 mps: that's what i meant by the script 2019-04-20 12:57:04 probably a nice-non-hurtful topic to hack on if someone really feels like it 2019-04-20 12:58:27 I thought to hack it but didn't had enough urge 2019-04-20 13:00:06 and I'm not proficient in script hacking 2019-04-20 13:42:11 hello guys, i've made a Dockerfile with alpine, and added an executable, but i can't execute the file, even is marked as executable, any idea? 2019-04-20 13:45:24 https://pastecode.xyz/view/9808ff2d 2019-04-20 13:52:08 You are still at the / after changing to appuser 2019-04-20 13:52:53 So file gets downloaded to the /, not to the /home/appuser 2019-04-20 13:55:40 hello, what's the usual process of pushing package updates from edge repo to 3.9 repo? 2019-04-20 13:56:07 <_ikke_> FreeBDSM: only security updates and bugfixes are normally backported 2019-04-20 13:56:17 3.9 repo contains a vulnerable package libjpeg-turbo 2019-04-20 13:56:28 https://pkgs.alpinelinux.org/package/v3.9/main/x86_64/libjpeg-turbo 1.5.3-r4 2019-04-20 13:56:48 <_ikke_> You can create a pull request against the 3.9 branch updating that package 2019-04-20 13:56:50 https://security-tracker.debian.org/tracker/CVE-2018-14498 Description: get_8bit_row in rdbmp.c in libjpeg-turbo through 1.5.90 ... 2019-04-20 13:57:39 damn it, need to remember how to get to alpine's issue tracker 2019-04-20 13:57:54 ah, bugs 2019-04-20 13:58:36 <_ikke_> bugs.a.o 2019-04-20 13:59:31 _ikke_: why does issues tracker miss '3.9.*' in 'Affected versions' list? 2019-04-20 13:59:47 <_ikke_> Probably because no one created it yet 2019-04-20 14:00:05 wait, no, they are there, the list is just not sorted 2019-04-20 14:00:15 there's even 3.9.4 and 3.10.0 2019-04-20 14:00:17 Dekedro: i've "cd ~" first, also I did check if archive is decompressed in appuser, i've sh into the container 2019-04-20 14:00:37 but when i try to execute node, i get sh: node file not found 2019-04-20 14:03:19 noobelus: I think that USER command only sets your uid. Try adding `cd /home/appuser` before downloading file 2019-04-20 14:04:57 hmm 2019-04-20 14:05:01 this is so strange 2019-04-20 14:06:06 i have started docker with entry point shell sh, and i see everything extracted correctly, also the executable files, but i can't execture any of them 2019-04-20 14:07:48 bug #10306 2019-04-20 14:09:32 _ikke_: shouldn't your issue tracker get improved by adding a new filed 'package name', so that issue tracker could figure out the package's maintainer and send bugmail to him? 2019-04-20 14:11:34 FreeBDSM: that would be nice option, I also thought about it, and to automatically send mail to maintainer of package 2019-04-20 14:11:50 ncopa: would you be so kind as to look at bug 10306 whenever you have some spare time? I'm writing this to you, as you are the maintainer of the package in question. I hope the package update is trivial enough to see it available for alpine 3.9.3 2019-04-20 14:14:47 hi 2019-04-20 14:15:07 the adoption rate of newer versions of quite widely used packages like libjpeg-turbo is quite astounding: https://sourceforge.net/projects/libjpeg-turbo/files/2.0.2/ shows that .tar.gz file of the most recent version (released on 14th of Feb) has been downloaded just 1532 times 2019-04-20 15:00:17 FreeBDSM: will take a look at it when i get a chance. can it wait til over the weekend? 2019-04-20 15:01:04 update from 1.x to 2.x may not be that easy, if it breaks the ABI 2019-04-20 15:01:25 ncopa: thank you! Yes it can. But honestly, I hoped that there's a simplified process of downporting a version from edge branch to the v3.9 2019-04-20 15:01:37 (the newer version of the package is already available on edge) 2019-04-20 15:02:20 may be it is simple, just need to check if it breaks ABI 2019-04-20 15:04:13 how does one check ABI compatibility? why not automate that by running tests? 2019-04-20 15:50:10 Good morning. "apk upgrade" tells me I've 2 errors, but I can't see any way to actually get the text of those errors... where would I find them? 2019-04-20 15:50:37 <_ikke_> tugrik: try apk fix 2019-04-20 15:51:19 :blush: thanks _ikke_ - I looked right past that option 2019-04-20 15:56:36 is openwrt by any chance based on alpine? 2019-04-20 15:56:43 <_ikke_> nope 2019-04-20 15:57:35 are there any projects similar to openwrt but based on alpine? 2019-04-20 15:58:38 well, there's alpine with acf 2019-04-20 15:58:46 I just run plain alpine on my gateways 2019-04-20 16:00:17 acf = Alpine Configuration Framework? 2019-04-20 16:00:21 <_ikke_> yes 2019-04-20 16:05:54 how do I unpin a package from @testing? 2019-04-20 16:07:07 ah, just apk add it without the @ 2019-04-20 16:17:07 Should alpine-standard-x86_64 boot on qemu with kvm enabled? x86 works fine for me but x86_64 gives kernel panic. Host OS: alpine edge x86_64 2019-04-20 16:17:37 Without kvm, both x86_64 and x86 works fine 2019-04-20 16:58:49 a project (PhantomJS) requires GLIBCXX_3.4.9 and GLIBC_2.7. Are they available in alpine? 2019-04-20 17:10:03 I have some doubts on that 2019-04-20 17:10:07 gonna try building it to verify 2019-04-20 17:10:15 but no, the whole point is that we use musl instead of glibc 2019-04-20 17:10:20 (well, a good chunk of the point :) ) 2019-04-20 17:10:46 ah, so that's why I have to rely on stuff like https://github.com/israelroldan/docker-sencha-cmd/raw/phantomjs-2.1.1/dockerized-phantomjs-2.1.1.tar.gz 2019-04-20 17:10:53 no? 2019-04-20 17:10:56 you can just build it 2019-04-20 17:10:58 at least I assume so 2019-04-20 17:11:14 as I said, I'm testing that :) 2019-04-20 17:11:27 how? 2019-04-20 17:11:33 what do you mean how? 2019-04-20 17:12:09 you download the GLIBCXX and GLIBC resources somewhere and build them using `build-base` package? 2019-04-20 17:12:14 no? 2019-04-20 17:12:20 then how? 2019-04-20 17:12:21 I downloaded the phantomjs sources and seeing if I can build them 2019-04-20 17:12:26 oh 2019-04-20 17:12:34 GLIBCXX is a symbol 2019-04-20 17:12:36 it's not a program 2019-04-20 17:14:05 SpaceToast: but http://phantomjs.org/build.html says that Linux build requires GNU C++ compiler 2019-04-20 17:15:34 SpaceToast: I'm not really much advanced in building packages, so I'm looking forward to the results of your build attempt 2019-04-20 17:15:46 GNU C++ compiler != GNU C Library 2019-04-20 17:15:50 those are separate things 2019-04-20 17:15:57 GNU C++ compiler = g++ (extension of gcc) 2019-04-20 17:16:05 GNU C Library (GLIBC and GLIBCXX) = glibc 2019-04-20 17:19:01 this build process is pretty badly designed though 2019-04-20 17:19:10 why can't you use system qt libraries, for instance? 2019-04-20 17:19:58 for what? 2019-04-20 17:20:05 ? 2019-04-20 17:20:12 I don't really get what you suggested 2019-04-20 17:21:34 I'm saying that phantom js' build process is badly designed 2019-04-20 17:21:44 right 2019-04-20 17:22:07 this project's development also stopped indefinitely 2019-04-20 17:22:39 yes 2019-04-20 17:22:49 perhaps you might consider finding a better alternative :) 2019-04-20 17:23:09 ok, it doesn't build because they're using broken sources as well 2019-04-20 17:23:16 (not surprising, given it's abandoned now) 2019-04-20 17:23:41 B. afaik - this project has no alternatives. A. I don't really need it, Sencha Cmd does. And I need Sencha Cmd 2019-04-20 17:23:58 https://github.com/dhamaniasad/HeadlessBrowsers 2019-04-20 17:24:28 okay, that covers the B part 2019-04-20 17:25:13 I don't think alpine would support sencha cmd 2019-04-20 17:25:20 it's not open source, and seems strange 2019-04-20 17:25:35 anyway, you might talk to sencha, given that they depend on an abandoned project 2019-04-20 17:25:51 I'm not asking for alpine to support it. I'll just make it work on alpine and that's it. 2019-04-20 17:26:16 if you're going to be making things work on alpine, you may want to learn about how compiling works 2019-04-20 17:26:25 (and a few other choice subjects) 2019-04-20 17:26:32 I think eventually it will come with experience 2019-04-20 17:26:54 I've compiled some programs, but never really felt I understood what I was doing 2019-04-20 17:32:54 SpaceToast: what did you mean by broken sources? 2019-04-20 17:33:11 it means that the sources have errors in them 2019-04-20 17:33:17 that cause compilation errors 2019-04-20 17:33:36 not like warnings or ignorable errors? 2019-04-20 17:34:19 lots of warnings, but no 2019-04-20 17:34:26 broken means broken, not "maybe has problems" :P 2019-04-20 17:34:32 and ignorable errors aren't usually a thing. 2019-04-20 17:45:55 +1 for "ignorable errors aren't a thing" 2019-04-20 17:54:11 SpaceToast: thanks for the info 2019-04-20 19:03:50 Hello friends - I wish to report a bug, although I'm not sure if such a problem I have is indeed a bug or not 2019-04-20 19:04:52 lbu ci -ev fails because openssl no longer supports: $openssl list-cipher-commands (line 129) 2019-04-20 19:14:25 c705, https://bugs.alpinelinux.org/issues/10275 ? 2019-04-20 19:15:22 fabled, thank you - checked a week ago and saw nothing 2019-04-20 19:34:36 speaking of c705, how come alpine is not using libressl instead of openssl? 2019-04-20 19:35:15 <_ikke_> Invader_Bork: Becase libressl break compattibality 2019-04-20 19:35:17 <_ikke_> broke* 2019-04-20 19:35:20 <_ikke_> and other reasons 2019-04-20 19:35:39 musl breaks compatibility too 2019-04-20 19:35:43 <_ikke_> They are no longer compattible with openssl, so it's no longer just a drop-in 2019-04-20 19:36:56 libressl felt closer to how alpine does things 2019-04-20 19:37:13 but it's not the end of the world, i am not making a feature request or anything 2019-04-20 19:37:17 just was curious 2019-04-20 19:37:30 <_ikke_> The state of openssl improved a lost since then 2019-04-20 19:37:36 <_ikke_> so the distinction is a lot less clear 2019-04-20 19:37:57 <_ikke_> There are mailing lists posts with all the reasons 2019-04-20 19:38:05 so there was a discussion 2019-04-20 19:38:28 <_ikke_> Alpine did use libressl of a while even, but switched back in 3.9 2019-04-20 19:38:55 what made you go back? 2019-04-20 19:39:23 void is doing quite fine with libressl 2019-04-20 19:40:28 <_ikke_> https://lists.alpinelinux.org/alpine-devel/6308.html 2019-04-20 19:41:17 q66, they're doing musl too right? 2019-04-20 19:41:21 _ikke_, ty 2019-04-20 19:41:33 Invader_Bork: as a variant 2019-04-20 19:41:36 you get a choice 2019-04-20 19:42:13 q66, do they have busybox as well? 2019-04-20 19:42:18 nope 2019-04-20 19:42:26 well, it's a package 2019-04-20 19:42:29 but coreutils is the base 2019-04-20 19:42:41 Whilst I have some folks here, which program on boot actually sources the apkovol and rebuilds? 2019-04-20 19:43:11 alpine-base? 2019-04-20 19:43:14 <_ikke_> something in the initramfs 2019-04-20 19:43:27 <_ikke_> no, alpine-base is part of apk-ovl 2019-04-20 19:44:01 <_ikke_> ie, it's in the world file in the apkovl, and hence, gets installed 2019-04-20 19:44:04 initramfs-init 2019-04-20 19:44:19 c705: https://github.com/alpinelinux/mkinitfs this repo, initramfs-init.in is the script 2019-04-20 19:44:50 SpaceToast, Thank you kindly 2019-04-21 06:15:59 Can we use lbu for backing up /etc to a different location other than the one chosen in setup-lbu? I used `lbu package|pkg -v [|]` and it backed up to the desired location but on reboot it was not as desired. Please help 2019-04-21 11:31:09 good day 2019-04-21 12:27:35 ncopa: There's a bug in latest radicale when importing 'load_entry_point'. 2019-04-21 13:18:56 Is Alpine Linux compatible with the first raspberry pi? (raspberry pi b) 2019-04-21 13:23:25 heya 2019-04-21 13:37:08 flipflop99: i think so 2019-04-21 13:37:35 flipflop99: scroll down at https://alpinelinux.org/downloads/ 2019-04-21 13:51:55 It doesn't say specifically in the wiki, and the download for raspberry has 3 options: armhf, aarch64 and armv7? 2019-04-21 13:53:52 armhf should be the correct one 2019-04-21 14:01:17 flipflop99: oh, indeed. in that case, what wej says. 2019-04-21 14:06:44 What are the others for? They are also under the raspberry downloads 2019-04-21 14:07:38 <_ikke_> aarch64 is for raspberry pi 3 2019-04-21 14:07:44 <_ikke_> (64 bits) 2019-04-21 14:20:03 I see, thanks! 2019-04-21 15:10:41 <_ikke_> ls 2019-04-21 16:55:37 Greetings, I want to run a os on alpine linux using containters, should I use lxd or lxc? 2019-04-21 16:57:22 <_ikke_> lxc runs fine on alpine, I have no experience with lxd 2019-04-21 16:58:47 _ikke_, lxc should allow running an entire os correct? 2019-04-21 17:00:21 <_ikke_> Not the kernel 2019-04-21 17:00:26 <_ikke_> it's still a container 2019-04-21 17:02:13 than I need lxd 2019-04-21 17:02:34 <_ikke_> lxd is still containers 2019-04-21 17:02:49 <_ikke_> if you want an entire os, you need something like qemu/kvm/libvirt 2019-04-21 17:03:05 <_ikke_> or virtualbox 2019-04-21 17:05:01 _ikke_, I'm already using kvm, I did some reading and found that lxd can be used to run oses with better performence 2019-04-21 17:06:36 <_ikke_> Because you are not running an entire OS, only userland 2019-04-21 17:07:19 <_ikke_> You are running processes on the host 2019-04-21 17:08:51 _ikke_, I can run a router within lxd for example , this would that means I'm running the entire os? 2019-04-21 17:09:39 <_ikke_> no, you are not 2019-04-21 17:10:32 _ikke_, unless I'm missing something, this guys http://www.makikiweb.com/Pi/lxc_openwrt.html runs openwrt within a lxc/lxd 2019-04-21 17:10:53 <_ikke_> It's still not an entire OS 2019-04-21 17:11:16 <_ikke_> You are still using the kernel from the host 2019-04-21 17:11:28 <_ikke_> daggs1: are you familiar with what a chroot is? 2019-04-21 17:11:42 _ikke_, yes 2019-04-21 17:11:50 <_ikke_> a container is a chroot on steroids 2019-04-21 17:11:59 I see. 2019-04-21 17:13:45 daggs1, A container shares the kernal with the host. A VM loads its own kernal, so in this sense, it is it's own "OS" 2019-04-21 17:14:20 so tecnically speaking, if I run os A with two vms, one openwrt and the second is alpine, I'll get the same result but with better performence with both vms within lxc? 2019-04-21 17:16:15 You're asking if performance is better with LXC opposed to VMs? Theoretically, containers should perform better than VMs due to the sharing of the kernal, yes 2019-04-21 17:16:47 and memory 2019-04-21 17:19:26 <_ikke_> But it really depends on the workload 2019-04-21 17:20:17 router + tvheadend 2019-04-21 18:01:40 './configure --help' gives './configure: exec: line 154: bash: not found' on edge lxc. Anyone noticed this problem 2019-04-21 18:07:19 <_ikke_> mps: while trying to build it? 2019-04-21 18:07:26 nvm, found reason. some packages I build yesterday required CONFIG_SHELL set to bash 2019-04-21 18:07:39 <_ikke_> ah ok 2019-04-21 18:07:47 I forgot to remove it 2019-04-21 18:08:47 I was perplexed why would tcsh configure script require bash :D 2019-04-21 18:10:42 new, fixed and polished tcsh is ready :) 2019-04-21 18:11:35 btw, anyone know if Dan Theisen, current tcsh maintainer is active? 2019-04-21 18:12:33 I don't know if it is ok to send relatively big changes without asking maintainer 2019-04-21 18:13:19 hm, this question is better to post to #alpine-devel, I think 2019-04-21 18:30:25 <_ikke_> mps: I guess you can try to e-mail the maintainer 2019-04-21 18:31:54 will do, but asked in case he is on some of the Alpine channels 2019-04-21 18:33:04 <_ikke_> Doesn't ring a bell 2019-04-21 18:33:08 Why does getent on alpine not support shadow, even though it uses the shadow system? 2019-04-21 18:33:23 <_ikke_> alpine by default does not use shadow 2019-04-21 18:33:52 oh... I did not intentionally change it on a new system, how did I get it? 2019-04-21 18:34:01 tcsh change maintainer with every patch, and most patches I sent to bugs.a.o actually, from first one to work 2019-04-21 20:01:40 if apk upgrade says "2 errors" - how do I find out what those errors are? 2019-04-21 20:02:27 <_ikke_> apk fix 2019-04-21 20:02:36 ok, that was easy 2019-04-21 20:02:42 thank, _ikke_ 2019-04-21 20:03:16 <_ikke_> :) 2019-04-21 20:03:37 <_ikke_> Would be nice if every tool had a fix command 2019-04-21 20:03:43 I am looking for the corresponding "apk break" 2019-04-21 20:03:47 true that 2019-04-21 20:03:51 hi 2019-04-21 20:04:11 yesterday, i upgraded my rp3 to 3.9.3 and now it won't boot. 2019-04-21 20:04:22 mount error 2019-04-21 20:04:31 <_ikke_> https://bugs.alpinelinux.org/issues/10302 ? 2019-04-21 20:04:34 mount: mounting /dev/mmcblk0p2 on /sysroot failed: invalid argument 2019-04-21 20:05:22 _ikke_: is there documentation about how/who is running alpine in terms of organisation? 2019-04-21 20:05:47 <_ikke_> telmich: We are currently working on that (as well as formalizing it a bit) 2019-04-21 20:05:49 hmmm, not sure. had been running 3.8 and used apk upgrade 2019-04-21 20:07:01 Mounting root failed 2019-04-21 20:07:13 Initramfs emergency recovery shell launched. Type 'exit' to continue bood 2019-04-21 20:08:08 if i type mount, then i can see all my files 2019-04-21 20:08:32 fdisk -l looks good 2019-04-21 20:09:18 mbucc: what arch you had previously 2019-04-21 20:09:24 except mmcblk0p1 is the partition marked as bootable. 2019-04-21 20:09:46 alpine 2019-04-21 20:09:49 3.8 2019-04-21 20:10:01 no, armhf or aarch64 2019-04-21 20:10:05 doh 2019-04-21 20:10:55 armhf i believe 2019-04-21 20:11:44 I don't have RPi's but remember that some have issues with aarc64 version 3.9.3 which will be fixed in next release 2019-04-21 20:12:31 ok. 2019-04-21 20:14:24 you can ask tomorrow again when here will be more connected users, probably someone have experience with 3.9.3 on RPi 2019-04-21 20:15:24 will do, thx. 2019-04-22 10:50:24 Hello :) 2019-04-22 10:50:34 <_ikke_> Hi 2019-04-22 10:52:17 I'm currently trying to set up Alpine on my ZFS root, but I get `mointing expool/ROOT/alpine on /sysroot failed: resource busy`. I suppose that's because I set ROOT=zfs=expool/ROOT/alpine on my kernel cmdline. But when I don't set that my initramfs doesn't load the ZFS module :c 2019-04-22 10:53:26 I've used dracut before, so maybe I'm just trying to force its behaviour onto mkinitfs? 2019-04-22 13:13:53 Hey I'm experiencing 2 sets of issues. I have the ryzen radeon vega8 combo. I just realized the ryzen-stabilizator from git also comes with services and activated those and hope it'll crash a lot less now. 2019-04-22 13:18:42 furthermore X doesn't start same as lightdm. Is there a better suited display manager for OpenRc than lightdm? its a very complex program config wise and I bet I did something wrong there like not manually creating the lightdm user 2019-04-22 13:30:13 zava: slim 2019-04-22 13:31:47 mps: thank you kind sir 2019-04-22 13:33:06 it is simpler, less resource hungry and quite fine, I use it on x86_64, x86, armhf, armv7 and aarch64 2019-04-22 13:34:03 even rebuilt it for my use without dependency on consolekit and polkit 2019-04-22 13:47:01 yay it starts X 2019-04-22 13:47:30 do I need an exec before the i3 in my .xinitrc? right now once im logged in my fans start blowing maxspeed and nothing happens 2019-04-22 13:48:04 yes, last command in .xinitrc should be window manager 2019-04-22 13:48:27 if you do not use Xfce :) 2019-04-22 13:48:46 xfce can start without .xinitrc, I think 2019-04-22 13:49:52 not sure about though, maybe it somehow puts startxfce4 there 2019-04-22 13:50:15 I now get as far as the slim password prompt on X. afterwards neither "exec i3" nor "exec mate-session". slim should let me choose which desktop environment i want right? 2019-04-22 13:51:13 don't put 'exec' in .xinitrc , simply i3 or whatever WM you use 2019-04-22 13:51:21 ok thanks 2019-04-22 13:51:40 will also read docs now to make sure Im not beating around the bush 2019-04-22 13:52:04 in my .xinitrc last line is 'awesome' literally 2019-04-22 13:53:13 if you want to see it I can post .xinitrc to tpaste.us 2019-04-22 13:53:40 i'd not say no to that! 2019-04-22 13:54:22 ok 2019-04-22 14:03:09 hooray. I realized I put my comp into an endless loop of starting $HOME/.xinitrc 2019-04-22 14:03:26 by just copying all xinit xsession xserverrc files from /etc/X11 2019-04-22 14:05:14 i3 starts up. mate-session doesnt find the dbus session bus 2019-04-22 14:18:57 mps: by which I meant yes sure I'd like to see it 2019-04-22 14:42:21 hooray it all works now. 2019-04-22 15:22:41 are there any good web gui firewall systems that are based on alpine or can be installed on top of alpine? I'm looking for something that's similar to pfsense in features, but is linux based (I have an old desktop that due to a bios bug that will never be fixed can't boot bsd) 2019-04-22 15:33:00 programmerq: great question! I'm intending to use alpine for exactly that. I'd be interested to know what you find 2019-04-22 15:35:53 the firewall projects I've found aren't gui based so far, which may end up being what I need to do, which is fine 2019-04-22 15:36:08 but I've been spoiled by the consumer router user experience for so long. :) 2019-04-22 15:36:24 <_ikke_> programmerq: I've never used it, but did you take a look at ACF? 2019-04-22 15:36:39 same. I'm screwed up so many FreeBSD upgrades I'm kind of afraid to use pfsense 2019-04-22 15:37:10 I will definitely check out acf 2019-04-22 15:40:24 that is really cool 2019-04-22 20:35:48 Hi, please tell me how to create a permanent directory /user in /run with users group permissions. 2019-04-22 20:36:23 /run is in tmpfs so it won't survive reboots, you can have it be created at every boot 2019-04-22 20:37:38 <_ikke_> The best place currently is to do it in the init.d service 2019-04-22 20:41:40 @_ikke_ is there a service or configuration file where it can be set in init.d? 2019-04-22 20:42:25 <_ikke_> BlueKiwi: Is it not related to a specific service? 2019-04-22 20:42:47 my guess is he is referring to XDG_RUNTIME_DIR 2019-04-22 20:44:28 I need it for custom bash script. Which is running without root permission 2019-04-22 20:45:13 <_ikke_> BlueKiwi: You can add a script in /etc/local.d/ and then rc-update add local 2019-04-22 20:52:09 @ @_ikke_ thx for the recommendation 2019-04-22 22:22:22 hi! I'm having issues installing latest scikit learn on alpine linux 2019-04-22 22:22:39 this happens inside a docker container, does anyone else have this? 2019-04-22 22:22:48 seemingly started happening last week 2019-04-22 22:38:29 g1eb: you'll have to provide more info. Is it running in a container based on Alpine or Alpine is the host os? What errors do you see? etc 2019-04-22 22:51:15 terror, yeah, it's running an alpine:edge container, i can share Dockerfile and logs 2019-04-22 22:54:50 g1eb: please do, I don't even know what package you are talking about at this point 2019-04-22 22:55:17 Here's the build log: https://dpaste.de/DRMC and the Dockerfile http://vpaste.net/w9ccD 2019-04-22 22:56:18 this used to work fine before last week so I'm tempted to say something changed in scikit-learn to cause this error 2019-04-22 23:06:16 g1eb: this probably is not alpine related 2019-04-22 23:09:33 yeah, i think so too, im just lost at where to start looking, maybe i could try scikit-learn irc haha :) 2019-04-22 23:19:52 g1eb: that looks like a fine idea to me 2019-04-23 07:01:54 @clandmeter hi, I tried Alpine (armhf) at an RPi1, didn't worked, tried 3.9.2 and it works again :) guess your fix for 3.9.3 killed the RPi1 ;)# 2019-04-23 08:31:24 asgoo: what is the issue? 2019-04-23 08:31:56 @clandmeter it doesn't boot at all (no rainbow screen etc) 2019-04-23 08:32:18 that the bootloader 2019-04-23 08:38:12 asgoo: that would be strange if its related to my change. 2019-04-23 08:39:10 idk to be honest, but otherwise not too much changed there 2019-04-23 08:39:27 also there are no config-rpi files anymore, just config-rpi2 2019-04-23 08:51:19 ah i see, im stupid 2019-04-23 08:51:28 ok ill add it back 2019-04-23 08:57:22 asgoo: i just added rpi kernel back to armhf, stupid mistake :| 2019-04-23 08:57:27 should be in next release 2019-04-23 09:04:05 @clandmeter tx :) 2019-04-23 09:05:14 you can copy the kernel and initramfs and modloop from prev release 2019-04-23 09:05:16 forn ow... 2019-04-23 10:00:41 @clandmeter I installed 3.9.2 and updated, good enough for now, tx for quick fix 2019-04-23 11:03:47 Hi, any guide for AD auth login ? :) 2019-04-23 11:13:45 afaik there's no /etc/nsswitch.conf in musl so in Alpine 2019-04-23 11:15:02 https://github.com/pikhq/musl-nscd 2019-04-23 11:19:02 Andreetje that was for you 2019-04-23 11:19:21 I'll have a look.. trying to understand it ;) 2019-04-23 11:19:29 I'm new into Linux/Alpine... ;) 2019-04-23 11:20:00 I suggest you to use SSSD 2019-04-23 11:20:14 its much easy than winbind 2019-04-23 11:20:52 i couldn't had time to test AD login with Alpine 2019-04-23 11:21:08 apk add sssd 2019-04-23 11:21:10 I will do in time but I'm quite busy nowadays 2019-04-23 11:21:18 ant then.. the magic config ;) 2019-04-23 11:23:03 trick is, configuration of sssd.conf, nsswitch.conf and pam modules correctly 2019-04-23 11:23:45 sounds quit hard 2019-04-23 11:23:46 you can even change domain password on linux with "passwd" command 2019-04-23 11:25:04 start with sssd howto on redhat site 2019-04-23 11:26:08 I'll have a look thnx 2019-04-23 11:30:43 i think this nsswitch.conf issue is one of the big obstacles regardind musl and so Alpine 2019-04-23 11:31:24 therefore the link you've sent 2019-04-23 13:06:37 would it be possible to add a small patch to the lxc aport to fix an issue with containers on zfs backend? https://bugs.alpinelinux.org/issues/10259 2019-04-23 13:07:28 (was included upstream but next release may take a while) 2019-04-23 15:25:01 I've found myself in dependency hell, courtesy of Perl 2019-04-23 15:26:44 installing request tracker ticketing system, needs a ton of perl modules... all were in the repos except GnuPG::Interface and PerlIO::eol, and Email::Address was an unacceptable old version 2019-04-23 15:28:07 so I decided I'd create my own packages with apkbuild-cpan... that worked for Email::Address just fine, but the rest has me stumped as they require other perl modules not present in repos 2019-04-23 15:28:27 when I try to build those other dependencies, I'm finding them conflicting with existing packages 2019-04-23 15:28:50 namely perl-utils 2019-04-23 15:31:11 I thought I might be able to just use CPAN directly, but I can't seem to find it in the repo 2019-04-23 15:35:54 well, I may have just figured out how to use CPAN 2019-04-23 15:50:43 my container doesn't build 3.9 properly ... when trying to install xdebug it throws an error that says autoconf is missing (despite being installed in it's own step) 2019-04-23 15:51:25 probably some paths have changed where the php util that builds in xdebug can no longer set a variable for some reason 2019-04-23 15:51:32 or the varialbe is now wrong 2019-04-23 17:35:28 if it's just a path/version issue, an autoreconf might help 2019-04-23 17:35:33 ash_worksi: ^ 2019-04-23 17:38:24 iggy: thanks 2019-04-23 17:38:29 I'll look into it 2019-04-23 17:38:34 for now I'm pinning to 3.8 2019-04-23 18:55:34 hi everyone, I just hit bug #10209 (https://bugs.alpinelinux.org/issues/10209) and currently it doesn't seem queued in for any upcoming release. Is there a process to check what fixes or upgrades get selected for release? 2019-04-23 18:59:47 <_ikke_> karthanistyr: everything in master (edge) is going to be in the next major release 2019-04-23 18:59:52 <_ikke_> (3.10) 2019-04-23 19:03:29 ok, so if I understand correctly, this is lost to the lower releases unless there is some special treatment 2019-04-23 19:04:03 <_ikke_> Yes, it could be backported to existing releases 2019-04-23 19:05:21 for example, in my case, ideally this is backported to 3.9.x, is there an avenue to +1 the potential backport? 2019-04-23 19:06:34 <_ikke_> karthanistyr: You could provide a PR targeted to the 3.9-stable branch with the same fix if applicable 2019-04-23 19:07:42 interesting, thanks! 2019-04-23 22:17:55 Good evening (again), so I did my PR, and the contributing guildelines say "wait for a dev to look at it", so I take it that I won't have to open a bug at bugs.alpinelinux.org or some other form of request for consideration at this stage? 2019-04-23 22:20:43 No, just wait until a dev will eventually review it :) 2019-04-23 22:21:30 Cogitri: thanks, let's be hopeful then 2019-04-23 22:23:08 Yup, fast review times are important :) 2019-04-23 22:54:24 Anyone also having https://bugs.alpinelinux.org/issues/10309 ? 2019-04-24 05:18:29 mps: I just replied to your maintainer email re: tcsh 2019-04-24 05:18:40 let me know if you need any help 2019-04-24 06:22:56 sorry if this is a noob question. what's a good way to check out the contents of an .apk file? 2019-04-24 06:23:22 seems like "apk info" only works on packages installed in the system 2019-04-24 06:32:12 ah, so they're just tarballs 2019-04-24 06:32:23 so "tar tvf file.apk" works quite well 2019-04-24 07:46:29 mepholic: Hi! didn't know you by real name, although we talked here few times, iirc 2019-04-24 07:49:19 I asked two or three times here if you (your real name) are here in about last ten days but didn't got any answer, probably we are at opposite TZ's 2019-04-24 07:53:59 that, and my name doesn't highlight me 2019-04-24 07:54:01 I've fixed that 2019-04-24 07:54:14 i'm sort of a nightowl anyways so I probably shoulda seen it 2019-04-24 07:56:40 never mind, regarding my question about tcsh, I used current development branch which fixes some UTF issues and some other things to build it locally 2019-04-24 07:57:29 and added /etc/csh.login and /etc/csh.cshrc files 2019-04-24 07:58:51 and /bin/csh symlink to /bin/tcsh, who will dare to use 'csh' name for another executable :) 2019-04-24 07:59:34 if you mind I can post aports patch to you for review 2019-04-24 08:07:40 sure, I can take a look 2019-04-24 08:07:51 I'm not sure what alpine's official stance is on the symlink thing though 2019-04-24 08:07:56 that's mostly why I mentioned it 2019-04-24 08:09:05 busybox is 'full' of symlinks :) and there are symlinks in other packages 2019-04-24 08:10:33 here is the patch http://tpaste.us/xpJQ 2019-04-24 08:12:14 regarding maintainership, I don't mind about it, just asked because didn't know you are active and wanted for tcsh to have active maintainer 2019-04-24 08:12:35 I'm more active in the adelie community these days 2019-04-24 08:12:41 but I don't mind maintaining it 2019-04-24 08:12:48 i maintain tcsh over there too :P 2019-04-24 08:12:54 so I might end up pulling this patch anyways 2019-04-24 08:14:07 however you think is ok, at the end you can add me there 2019-04-24 08:14:29 add a contributor line for yourself for sure 2019-04-24 08:15:36 I use tcsh as my interactive shell for decades, few times tried to switch to some other with more 'features' but always went back 2019-04-24 08:16:51 i think zsh has a t?csh emulation mode 2019-04-24 08:17:07 zsh actually uses less resources than bash with a stock config on each 2019-04-24 08:17:13 yes, tried it but it not true tcsh 2019-04-24 08:17:15 well, zsh in bash emulation mode at least 2019-04-24 08:17:25 probably not :P 2019-04-24 08:17:25 isn't zsh without any config borderline unusable? 2019-04-24 08:18:58 not really 2019-04-24 08:19:24 we ship zsh by default on adelie with a custom prompt, bash compat mode, and basically 0 other configuration 2019-04-24 08:19:34 and if you didn't know to look, you'd probably think it was just bash 2019-04-24 08:20:22 maybe I'm just used to my pimped config too much 2019-04-24 08:21:54 mepholic: ok, will wait a day or two for your comment and if you tell me it is sane will send it to patchwork.a.o 2019-04-24 08:28:56 mps, was community/tcsh/002-test_123-posix_fix.patch fixed upstream? 2019-04-24 08:30:20 yes, because that I removed it 2019-04-24 08:31:08 i'm trying to find the upstream commit that fixes it 2019-04-24 08:31:24 https://github.com/tcsh-org/tcsh/blob/master/Fixes 2019-04-24 08:31:26 would have been nice to get attribution for that considering I spent hours debugging that 2019-04-24 08:31:55 and added a comment and everything citing the exact issue in my patch 2019-04-24 08:32:09 uhm it is not listed there 2019-04-24 08:33:01 yeah i don't see it there 2019-04-24 08:33:02 hmm 2019-04-24 08:33:03 hold on 2019-04-24 08:34:16 I blindly removed it because I abuild told me that it is already applied 2019-04-24 08:35:01 https://github.com/tcsh-org/tcsh/commit/2ad4fc1705893207598ed5cd21713ddf3f17bba0 2019-04-24 08:36:17 that commit was AFTER the last tagged release 2019-04-24 08:36:23 I think I ended up duplicating work 2019-04-24 08:36:37 it was also BEFORE the timestamp on my own patch 2019-04-24 08:37:02 it's also exactly the same patch because there's really only one way of fixing that without reworking the entire test 2019-04-24 08:37:30 yes, I see, https://github.com/tcsh-org/tcsh/pull/1#issuecomment-282035528 2019-04-24 08:38:48 mps: lets see if we can get the tcsh people to get us a real tagged release 2019-04-24 08:39:25 I don't like using git commit ID's, especially on github archives 2019-04-24 08:39:34 would be nice, thought to contact them but I don't have github account 2019-04-24 08:39:45 I can do that 2019-04-24 08:40:17 also, I'm against git commit but sometimes this is only solution 2019-04-24 08:40:33 yeah 2019-04-24 08:40:47 ok, nice, please inform me when you get anything from them 2019-04-24 08:42:06 btw, I have private patches for some software but don't send them because only option is github and I dislike it 2019-04-24 08:42:37 check out the repo, do a git log 2019-04-24 08:42:45 and send them directly to whoever committed the most :P 2019-04-24 08:42:52 git shortlog can help you determine that 2019-04-24 08:43:07 email is the way to go 2019-04-24 08:43:50 well, I have some kind of hesitation to bother people by mail if their mail address is not public 2019-04-24 08:45:32 for example, I thought to send email to you more than a month or two but always hesitated because didn't wanted to disturb you 2019-04-24 08:47:27 I have a feeling mps and mepholic are in close timezones :) 2019-04-24 08:49:02 well my name is in my highlights now, so I shouldn't have that issue again 2019-04-24 08:49:34 mps: don't hesitate to contact me though, it's no bother, and if it becomes one there's always /dev/null :P 2019-04-24 08:49:57 that's how I see email boxes anyways 2019-04-24 08:50:04 considering that many people rarely check them 2019-04-24 08:51:48 now I know, but yesterday was not sure. although remember that we 'talked' here few times 2019-04-24 09:14:28 where should I put alias foo="bar" for automatic sourcing ? 2019-04-24 09:15:32 I tried ~/.profile, but its not sourced if I try to run docker --rm -it johndoe bar 2019-04-24 09:15:33 <_ikke_> for ash, when having a login shell, in ~/.profile 2019-04-24 09:24:43 _ikke_ Does not work for me :/ Strange ...docker run --rm -it test sh -c 'az' -> sh: az: not found Dockerfile: https://pastebin.com/wjLbzQPn 2019-04-24 09:25:03 <_ikke_> muhaha: Might be that it's not giving you a login shell 2019-04-24 09:25:44 ah 2019-04-24 09:26:03 And is possible to export it for all users before login session? 2019-04-24 09:26:41 <_ikke_> /etc/profile 2019-04-24 09:26:55 <_ikke_> https://linux.die.net/man/1/ash 2019-04-24 09:29:11 I am aware of this, but I need to be able to run it without login session (docker run --rm -it test sh -c 'az' ) 2019-04-24 09:30:09 Maybe some alternative to bashrc ? 2019-04-24 09:30:18 <_ikke_> try sh -lc 'az' 2019-04-24 09:30:38 <_ikke_> Not sure if that would have consequences 2019-04-24 09:32:54 Does not work, iwill have to create executable in /usr/bin I guess 2019-04-24 09:53:13 uh, when i do `apk upgrade` i get this response: ERROR: unsatisfiable constraints:\nHuh? Error reporter did not find the broken constraints. 2019-04-24 09:53:30 same when doing `apk fix`, why could this be? 2019-04-24 10:46:44 muhaha: you should not use an alias, a shell function will work for you much better 2019-04-24 10:47:39 tcely: well, I did printf "#!/bin/sh\n_az() { . /azure-env/bin/activate && command az "'"$@"'" ; deactivate; }; _az "'"$@"'"" >> /usr/bin/az && chmod +x /usr/bin/az 2019-04-24 10:47:49 I hope its ok 2019-04-24 10:49:16 besides the odd quoting it looks ok 2019-04-24 10:50:13 Its wacky, .. and of course > redirect 2019-04-24 10:50:14 nevermind, the quoting is just the stupid font for this window 2019-04-24 11:32:18 anyone around here know mandarin? 2019-04-24 11:35:41 I'm trying to contact an SoC vendor for information and/or access to their SDK 2019-04-25 01:27:00 chromium 73 is sluggish as hell compared to 68 2019-04-25 01:27:17 and ublock origin crashes; none of my other extensions do 2019-04-25 01:27:57 im running chromium from terminal so i can catch any errors that come up, but uBO isnt segfaulting or anything that would print an error 2019-04-25 01:28:03 dunno whats up 2019-04-25 02:39:19 I've run into an issue that I can't quite put my finger on. PerlIO::eol will not install via CPAN and is not part of the Alpine repos 2019-04-25 02:39:42 :D contribute it to the repos! 2019-04-25 02:39:46 it fails in the testing phase 2019-04-25 02:40:05 send the errors via a paste service 2019-04-25 02:40:10 I was thinking of it, while installing Request Tracker (RT) this week I ran into a bunch of perl modules that don't exist 2019-04-25 02:40:46 https://bpaste.net/show/cf573c2e1386 2019-04-25 02:41:14 I attempt the same module installation in the same manner on an ubuntu machine and it installs flawlessly 2019-04-25 02:41:27 leading me to believe it's either a musl thing or a compiler flag issue 2019-04-25 02:43:02 fresh install of alpine 2019-04-25 02:56:47 it's probably bugs in perlio::eol 2019-04-25 02:57:01 but the paste you showed doesn't have the test results 2019-04-25 02:57:09 it just says some failed without showing what :/ 2019-04-25 03:06:53 I've tried building from source--basically the same process--and enabled the debugging output, but it wasn't any more helpful 2019-04-25 03:07:04 it shows it getting to test 6 and stopping 2019-04-25 03:08:28 can you figure out what test 6 is? 2019-04-25 03:08:30 i'll see what better output i can find 2019-04-25 03:08:43 I can't, the test file is so cryptic I can't tell what's what 2019-04-25 03:09:30 yeah.. 2019-04-25 03:09:33 same here 2019-04-25 03:09:37 perl is so awful 2019-04-25 03:12:03 https://bpaste.net/show/c927a436cf0b 2019-04-25 03:13:45 https://metacpan.org/source/SHLOMIF/PerlIO-eol-0.17/t/1-basic.t 2019-04-25 03:14:19 there are supposedly 23 tests here, but I don't know what's what 2019-04-25 03:20:49 cool, now i can run vmx 2019-04-25 03:21:05 but how can i doto restar rio if hangs 2019-04-25 03:21:23 i can hit ctl+alt+supr to reboot 2019-04-25 03:21:35 but there another key combo? 2019-04-25 03:22:19 e-werdd, yeah, same as what i found 2019-04-25 03:22:23 perl is awful 2019-04-25 03:23:25 ups 2019-04-25 03:23:27 wronmg window 2019-04-25 10:54:28 is it possible to set boot of one lxc container to depend on another 2019-04-25 11:43:44 Anybody can help me to install docker-compose with apk 2019-04-25 11:44:01 running docker-compose -v gets me 2019-04-25 11:44:02 pkg_resources.DistributionNotFound: The 'pyrsistent>=0.14.0' distribution was not found and is required by jsonschema 2019-04-25 11:44:11 <_ikke_> docker-compose is currently broken on edge 2019-04-25 11:44:18 Damn it! 2019-04-25 11:44:29 <_ikke_> You can fix it by running apk add py-pristent manually 2019-04-25 11:44:47 Why isn't this a dependency ? :O 2019-04-25 11:44:53 <_ikke_> That's being fixed 2019-04-25 11:45:25 <_ikke_> https://github.com/alpinelinux/aports/pull/7115 2019-04-25 11:47:14 Are you sure py-pristent is the name of the package ? 2019-04-25 11:47:47 I'm getting an error 2019-04-25 11:47:48 <_ikke_> py-pyrsistent 2019-04-25 11:48:01 <_ikke_> even py3 2019-04-25 11:48:05 <_ikke_> py3-pyrsistent 2019-04-25 11:48:57 Downloaded, but docker-compose still needs some more dependancy I think 2019-04-25 11:49:53 <_ikke_> py3-jsonschema 2019-04-25 11:50:17 <_ikke_> which needs py3-attrs as well 2019-04-25 11:50:40 Good, thanks! 2019-04-25 11:52:25 Also, more of a noobish question, when you apk del something is there anything left of that packages ? 2019-04-25 11:52:36 Config, dependancys etc ? 2019-04-25 11:53:32 <_ikke_> dependencies get automatically removed (if they are not required for some other package) 2019-04-25 11:54:16 SO the fact that I downloaded those 2 dependacys manually means they won't get removed if I decide to apk del docker-compose ? 2019-04-25 11:54:24 <_ikke_> correct 2019-04-25 11:55:08 <_ikke_> They are part of /etc/apk/world, meaning they are explicitly installed) 2019-04-25 11:55:27 Cause I accidentally downloaded py2, and deleted it, is there anything left of it anywhere &? 2019-04-25 11:56:59 And could I delete them from /apk/world, meaning that if the package gets fixed, it will auto delete the dependency if I decide to unninstall docker-compose ? 2019-04-25 11:57:55 <_ikke_> removing them from /etc/apk/world could mean they might end up being uninstalled automatically 2019-04-25 11:58:11 <_ikke_> KH405: and those kinds of packages should leave nothing behind 2019-04-25 11:58:33 Aight, thanks alot! Slowly learning more and more about linux/alpine in general :P 2019-04-25 11:58:43 Now back to docker :P 2019-04-25 12:00:15 dalias: hi. two questions if you don't mind regarding http://www.etalabs.net/sh_tricks.html : 1. why do you quote `eval "$foo=\$bar"`? is it to resist IFS values which could delimit inside variable names? 2. re echo, what do you think of this approach to combine your two approaches: https://0x0.st/zbAY.txt ? 2019-04-25 12:06:02 avih, i quote because its a pattern you should follow when writing shell. its no cost and avoids costly analysis of whether its safe not to 2019-04-25 12:06:59 echo looks right offhand 2019-04-25 12:07:04 dalias: no disagreement there. however, to the best of your knowledge, in this specific construct, and assuming $foo is a valid variable name, is there anything which can interfere other than IFS? 2019-04-25 12:08:39 well, IFS could also be '=', but still it'd be to resist IFS 2019-04-25 12:10:14 i cant think of anything else offhand 2019-04-25 12:10:31 k. thanks for your time. 2019-04-25 12:15:37 np 2019-04-25 17:40:38 :qa 2019-04-25 17:41:05 :d 2019-04-25 18:47:34 Good day people 2019-04-25 18:52:04 <_ikke_> hi 2019-04-25 18:53:11 nekocentral: o/ 2019-04-25 19:45:09 nekocentral: hi 2019-04-25 20:28:13 Hi team, who acts on packages flagged as out of date? 2019-04-25 20:28:30 <_ikke_> terror: The package maintainer usually (they get an e-mail) 2019-04-25 20:30:11 How is it packaged out of date ? someone makes an issue on bugs.alpinelinux.org ? 2019-04-25 20:30:24 s/packaged/flagged 2019-04-25 20:30:46 <_ikke_> north1: via pkgs.alpinelinux.org 2019-04-25 20:31:38 ah nice, it has an auto-check mechanism ? 2019-04-25 20:32:24 <_ikke_> I believe there is a service that can automatically flag packages out-of-date, but I'm not sure if it's still working 2019-04-25 20:34:21 I guess Repology will have to do then 2019-04-25 20:34:54 <_ikke_> THat's what I use personally 2019-04-25 20:35:59 Has some false-positives (I'm still questioning why some distros package their git versions as `latest upstream release + 1`), but it works well enough for the most part 2019-04-25 20:36:20 cogitri: the secret is check if there is at least 5 or 6 people on the newer version 2019-04-25 20:36:30 if there is only 1 or 2 you can write it off as wikimedia 2019-04-25 20:43:49 Is there a way to see contents of a package that is not installed ? i want to compare 2 different versions of the same package to see what changed between each 2019-04-25 20:44:24 <_ikke_> afaik, you can't 2019-04-25 20:44:34 <_ikke_> only through pkg.a.o 2019-04-25 20:44:42 bummer 2019-04-25 20:44:54 would be super useful to see what changed when bumping a package up 2019-04-25 20:44:58 apk-file 2019-04-25 20:45:02 <_ikke_> oh 2019-04-25 20:45:17 Neat 2019-04-25 20:45:29 <_ikke_> Didn't know that existed already 2019-04-25 20:46:16 not perfect but useful 2019-04-25 20:47:48 mps: i don't think apk-file can list whole contents of a package 2019-04-25 20:48:16 i want something equal to apk info -L but for a remote package so i can just pipe them through diff 2019-04-25 20:49:50 apk-file --arch x86_64 mupdf 2019-04-25 20:50:15 see what you get as response 2019-04-25 20:50:30 http://ix.io/1Hdt 2019-04-25 20:50:32 not perfect as I wrote 2019-04-25 20:51:37 well completely useless for the usecase i asked but yes apk-file is a nice tool 2019-04-25 20:52:25 agree, apk info --remote -L would be nice 2019-04-25 20:53:52 does pkg.a.o has an API for fetching contents of a package ? 2019-04-25 20:54:08 <_ikke_> 3not yet 2019-04-25 21:00:38 oh 2019-04-25 21:02:01 _ikke_: The auto flag doesn't seem to be working from a random sample of packages I use 2019-04-25 21:02:47 <_ikke_> https://bugs.alpinelinux.org/issues/9589 2019-04-25 21:03:20 _ikke_: thanks 2019-04-25 21:05:34 <_ikke_> So apparently it still runs 2019-04-25 21:05:37 _ikke_: is anyone looking into that bug? Sounds like a 10' fix 2019-04-25 21:05:55 <_ikke_> I'm just checking 2019-04-25 21:07:02 <_ikke_> terror: what packages did you check? 2019-04-25 21:08:07 I can't remember of the top of my head, this was 1 month ago 2019-04-25 21:08:18 <_ikke_> ok 2019-04-25 22:19:30 any thoughts on why haveged init script would depend on the network being up to run? seems to keep it from starting. commenting out "need net" fixed it. 2019-04-25 22:19:44 seems a bogus dependency 2019-04-25 22:29:40 maybe it harvests entropy from network activity, but in any case I think that it's a mistake 2019-04-25 22:35:37 nah it doesn't. does cpu benchmarks and some jitter timing stuff. 2019-04-25 22:36:01 kernel does network i think 2019-04-25 22:38:41 guess i can do a PR on github 2019-04-25 22:40:38 hello 2019-04-25 22:40:54 migil: hi 2019-04-25 22:40:54 I think the init script here is maintained by an Alpine maintainer (Jakub Jirutka in this case) 2019-04-25 22:42:02 can i @tag them or something? or does it have to go through a bug tracker then? 2019-04-25 22:42:29 i saw the maintainer and /whois'd jakub yesterday, theyre not on here by that name at least 2019-04-25 22:43:00 m4t: he is not here 2019-04-25 22:43:07 k 2019-04-25 22:43:40 alpine uses cgit, I don't know if/how you could make a PR on it 2019-04-25 22:44:03 doggone: alpine has a github mirror one can makes PRs for. 2019-04-25 22:44:05 I usually create a bug report at bugs.alpinelinux.org , but it's a drag you have to create an account 2019-04-25 22:44:09 it is not obligatory to mention or mail maintainer if you make patch or PR 2019-04-25 22:44:12 you can also send a patch with git send-mail 2019-04-25 22:44:35 ah that is nice 2019-04-25 22:44:58 i unfortunately chose a provider (protonmail) for my dev email that does not provide smtp :( 2019-04-25 22:45:05 doggone: anon bug reports are disabled because of spammers 2019-04-25 22:45:37 i'm getting my first board soon using ARMv8-A (Cortex-A72) https://libre.computer/products/boards/roc-rk3399-pc/ do i use aarch64? (sry newb) 2019-04-25 22:45:53 migil: yes 2019-04-25 22:46:02 okay thats what i thought 2019-04-25 22:46:11 thanks very much 2019-04-25 22:47:04 I'm typing right now from rk3399 box ;) 2019-04-25 22:47:30 thats awesome! 2019-04-25 22:48:42 missing is X driver for mali T860 2019-04-25 22:50:07 okay 2019-04-26 00:03:00 opened a PR and @tagged the maintainer, thanks 2019-04-26 02:54:07 _ikke_: Anyway I can tell when docker-compose will be fine again without checking everyday? 2019-04-26 02:54:22 So I can fix my /apk/world file 2019-04-26 02:57:18 KH405_TV: should be fixed by https://github.com/alpinelinux/aports/commit/3b978a80fe81e423b3e35cb20fa52e392b7f27c6 i think 2019-04-26 02:57:27 or are there more dependencies missing in the chain ? 2019-04-26 03:09:09 Tryed it this morning and there was 2 packages missing 2019-04-26 03:09:21 oh, which ones ? 2019-04-26 03:09:35 py3-pyrsistent and py3-jsonschema 2019-04-26 03:10:05 the commit i sent is about py3-pyrsistent being moved to the community repo so py3-jsonschema can use it 2019-04-26 03:10:15 But it was 2 days ago 2019-04-26 03:10:26 I'm on edge and it still isn't there 2019-04-26 03:10:34 so the py3-pyrsistent dependency should be fixed, anything else on py3-jsonschema ? 2019-04-26 03:10:51 What do you mean? 2019-04-26 03:11:28 i am on edge too, just installed docker-compose, any specific command i need to run to see if it works or j ust running the binary is enough ? 2019-04-26 03:11:41 docker-compose -v 2019-04-26 03:12:40 https://i.imgur.com/BuSdy6J.png 2019-04-26 03:12:42 seems to work here 2019-04-26 03:12:46 This morning it was returning a bunch of errors 2019-04-26 03:13:36 Alright then! Will delete them from my apk/world file and see how it does :P 2019-04-26 03:13:46 hmmm, are you using a mirror ? maybe it is not synced 2019-04-26 03:15:36 https://imgur.com/a/FEfPojG 2019-04-26 03:15:44 That's a rip from me dawg 2019-04-26 03:15:56 Yeah, i'm using a mirror tough 2019-04-26 03:18:25 https://mirrors.alpinelinux.org/ says it's synced tough 2019-04-26 03:18:48 hmm, i am lost then, it works here 2019-04-26 03:19:26 sorry, can't help any further 2019-04-26 03:20:15 Maybe I should move that to apline-devel ? 2019-04-26 03:26:24 apk -U upgrade ;-) 2019-04-26 03:26:48 Nope ... 2019-04-26 03:27:23 its a shorthand for the command you gave 2019-04-26 03:27:31 what is your problem? 2019-04-26 03:29:18 Apparently py3-attrs and py3-pyrsistent are dependancy for docker-compose 2019-04-26 03:29:35 But if I remove them from my apk/world file, they unninstall 2019-04-26 03:30:02 py3-jsonschema should depend on them 2019-04-26 03:30:27 but yeah i have no clue what went wrong in your system 2019-04-26 03:31:35 I don't think it's my system ^^' 2019-04-26 03:31:43 Pretty sure the package is still broken 2019-04-26 03:32:18 on which branch are you? 2019-04-26 03:32:22 edge 2019-04-26 03:32:27 i just tested it isn't, even deleted my local repo from which i packaged py3-pyrsistent 2019-04-26 03:33:02 oh 2019-04-26 03:33:11 no it is broken yeah. 2019-04-26 03:33:13 my bad 2019-04-26 03:33:31 lol 2019-04-26 03:33:49 forgot to remove py3-jsonschema from ~/packages/community 2019-04-26 03:35:01 Rip 2019-04-26 03:35:18 This should fix it 2019-04-26 03:35:24 https://github.com/alpinelinux/aports/pull/7334 2019-04-26 03:35:38 Anyway, back to my first question, how can I tell when docker-compose will be fine again without checking everyday? 2019-04-26 03:35:53 once that PR is merged 2019-04-26 03:36:42 lol, 3 minutes ago 2019-04-26 03:40:03 KH405_TV: what do you expect from us? 2019-04-26 03:40:52 That you share your knowledge 2019-04-26 03:41:21 and what will i get in return for it? 2019-04-26 03:41:28 or we 2019-04-26 03:41:44 I mean isn't this a support channel? 2019-04-26 03:42:01 slow channel, ask your question and wait for a response. :), litteraly in the title 2019-04-26 03:44:41 KH405_TV: the only way to get notified is if you create a ticked for issues. 2019-04-26 03:44:55 ticket even 2019-04-26 03:45:23 and you are on edge, so expect things to be broken. 2019-04-26 03:45:26 Alright, thanks! 2019-04-26 03:45:52 Sure, i'm not mad cause it is broken, i'm just wondering if there is a way to know when the package will work fine 2019-04-26 03:46:12 the ofifcial way is by creating a ticket 2019-04-26 03:46:52 but that doesnt mean it gets fixed right away 2019-04-26 03:47:02 you need to monitor it yourself 2019-04-26 03:47:41 ok 2019-04-26 03:47:44 we dont get payed to support people, so we do it when we have free time. 2019-04-26 03:47:55 https://github.com/alpinelinux/aports/pull/7334 if it passes CI should be good to go 2019-04-26 03:48:21 north1: are you maxice8? 2019-04-26 03:48:32 clandmeter: yes 2019-04-26 03:48:57 ACTION adds a new entry in his nickname to github references  2019-04-26 03:49:06 :) 2019-04-26 03:49:15 time for luch bbl 2019-04-26 03:49:25 ohh, well goodnight man 2019-04-26 03:49:27 Thanks for the help 2019-04-26 03:49:40 yes i go by Leo and maxice8 is just temporary 2019-04-26 03:53:44 Can i get a CI-malfunction for https://github.com/alpinelinux/aports/pull/7333 ? 2019-04-26 04:08:35 When moving packages from repos should pkgrel be bumped ? 2019-04-26 04:36:07 <_ikke_> north1: ncopa said that should not be necessary 2019-04-26 04:36:35 _ikke_: ok 2019-04-26 04:36:50 <_ikke_> re docker-compase, someone already opened a PR for it, but it's hanging a bit: https://github.com/alpinelinux/aports/pull/7115 2019-04-26 04:37:58 i see 2019-04-26 04:38:07 needs a lot of changes to get merged 2019-04-26 04:38:20 <_ikke_> yes, indeed 2019-04-26 08:08:15 north1 is Leo on github ? 2019-04-26 08:08:54 <_ikke_> yes 2019-04-26 08:10:34 <_ikke_> and maxice8 as well 2019-04-26 08:24:58 tmhoang: yes 2019-04-26 08:59:53 if i want to install a working DE the "alpine-desktop" is mandatory no matter what WM i use right? the wiki lists a few but can't seem to find openbox/fluxbox/lxde the wiki tut is all about xfce 2019-04-26 09:03:22 boxemall: ime, xfce is best option for DE, although I personally use awesome WM only 2019-04-26 09:03:48 i can't wrap my head around "tilers" 2019-04-26 09:03:55 xfce is most stable of them, afaik 2019-04-26 09:03:57 i'm the mouse guy... 2019-04-26 09:04:21 i'm limited on ram (old netbook) so i'd prefer something a tad lighter than xfce 2019-04-26 09:04:47 i do like xfce but it's not as light anymore. any i'm trying to stay away from gtk3 2019-04-26 09:05:07 on my wife's notebook with 2GB ram xfce works quite fine 2019-04-26 09:05:57 < 1GB.... 2019-04-26 09:06:08 firefox is memory eater 2019-04-26 09:06:10 every meg counts, mate 2019-04-26 09:06:21 i know. FF can only run solo and max 2 tabs 2019-04-26 09:07:07 then mate DE maybe 2019-04-26 09:07:39 if you want to have DE, solo WM is quite fine, imo 2019-04-26 09:08:42 isn't mate even heavier than xfce? on my vbox testing any distro i used (it was a shitload really) all mate version closked in about 200-300 meg heavier than xfce 2019-04-26 09:08:51 clocked* 2019-04-26 09:10:09 could be, but it is gtk-2 afaik 2019-04-26 09:10:19 for experimenting... i don't like slim... does alpine support lightdm? 2019-04-26 09:10:30 yes 2019-04-26 09:10:56 although I prefer slim, it is simpler and less resource hungry 2019-04-26 09:11:26 even built it for me without support for consolekit and policykit 2019-04-26 09:11:39 hmm. right. but for experimenting (atm running virtual) i want to try out different DE's and slim doens't support switching iirc 2019-04-26 09:12:01 'switching' ? 2019-04-26 09:12:22 switch de for a session 2019-04-26 09:15:49 long time passed when I looked at this, so can't remeber, but I think it allows switch 2019-04-26 09:16:29 i'll see in a few 2019-04-26 09:23:57 seems plain. nothing to click. prolly can only switch by editing a file. 2019-04-26 09:25:08 F1 2019-04-26 09:25:38 ah damn... tried installing virtual box additions rebooted no i have a red console... uess i'm screwed lol 2019-04-26 09:56:15 mepholic: Hi! any news from tcsh upstream 2019-04-26 09:56:25 oh no I forgot 2019-04-26 09:56:39 uno momento 2019-04-26 09:57:01 niente problemo :) 2019-04-26 09:57:14 LOL 2019-04-26 09:57:22 wanna see a really funny issue? 2019-04-26 09:57:24 https://github.com/tcsh-org/tcsh/issues/11 2019-04-26 09:57:26 ahahahha 2019-04-26 09:57:37 i'm pretty sure that's NOTABUG 2019-04-26 09:58:10 I tested it few days ago and didn't see it as bug 2019-04-26 10:17:36 Hello... 2019-04-26 10:19:17 Hello there 2019-04-26 10:19:40 I'm trying to use VirtualBox VM to preapre an USB drive for a 'diskless' Alpine Linux machine... Now I can get the VM to see and use the drive, but VirtualBox doesn't support booting off USB. Therefor I used a 'trick' with creating a 'raw' vmdk, that is just a 'pointer' to the actual USB drive (on my host system). 2019-04-26 10:19:56 Howerver, this is now recognized as a 'regular' hard drive by Alpine... 2019-04-26 10:20:31 Is there a way (like a kernel parameter) that would 'force' Alpine Linux to treat this drive as an USB drive instead? 2019-04-26 10:24:12 mps: https://github.com/tcsh-org/tcsh/issues/12 2019-04-26 10:27:30 mepholic: nice. now we can only wait few days, I think. thanks 2019-04-26 10:27:40 danci1973: Not sure how you'd force ot yo recognise it as usb drive, but maybe you could just use Qemu instead of VB not have to work around this? 2019-04-26 10:28:07 s/VB not/VB to not/ 2019-04-26 10:28:23 Anyway, I will prepare patch with only csh.login and csh.cshrc and post it to patchwork over this weekend 2019-04-26 10:28:24 well, considering I helped them answer another issue before submitting that, hopefully they are kind and expedient 2019-04-26 10:28:34 sounds good 2019-04-26 10:28:44 you know, I'm not familiar with alpine's patchwork 2019-04-26 10:28:48 or where it even is 2019-04-26 10:28:58 that's probably a bad thing considering I'm a package maintainer 2019-04-26 10:29:04 I didn't see they mentioned you in fix of your patch 2019-04-26 10:29:18 where at? 2019-04-26 10:30:07 i think I determined that that tcsh patch I had in there was duplicated effort because I didn't even bother looking to see if they fixed it after the latest release 2019-04-26 10:30:12 and it turns out that they did 2019-04-26 10:30:15 well, I had to remove your patch from aports because it is already applied upstream, and didn't noticed they mentioned you 2019-04-26 10:30:21 Cogitri: Thanks, I guess I'll have to try that... 2019-04-26 10:30:42 before I got around to root causing it myself and applying the same fix that they did 2019-04-26 10:31:05 aha, duplicate work, I see 2019-04-26 10:32:36 I'm struggling with mupdf link flags right now, hope will have time this night for tcsh and some other packages 2019-04-26 11:33:14 mps: my description of the issue and why the fix works was better though :P 2019-04-26 11:33:39 or rather, why the original test was wrong 2019-04-26 11:36:28 don't see mepholic in commits history 2019-04-26 11:37:21 mps: https://git.alpinelinux.org/aports/tree/community/tcsh/002-test_123-posix_fix.patch 2019-04-26 11:37:25 the comment at the top of the patch 2019-04-26 11:39:14 i worded it poorly 2019-04-26 11:39:57 oh, your patch is not accepted? 2019-04-26 11:40:51 mps: Need help with mupdf? 2019-04-26 11:41:28 Cogitri: yes, I asked on #alpine-devel 2019-04-26 11:42:33 it looks like they preparing next release, i.e. 1.15.x probably 2019-04-26 11:47:47 mps: nah, it's the same patch that's upstream, basically with a little description of the change 2019-04-26 11:48:35 considering that tcsh is a codebase that's effectively in maintenance mode, if I was maintaining the codebase, any commit would have a detailed commit message that described what changed and why 2019-04-26 11:49:04 eh, If I start to count my patches and contributions which other take and credited :) 2019-04-26 11:49:11 I figure that could be used to build a changelog, and contain addition information that might be of interested in the parties that are still using tcsh 2019-04-26 11:49:39 jeez I can't talk tonight 2019-04-26 12:08:25 dalias: actually, |IFS=" " eval 'printf %s\\n "$*"'| is incorrect. eval is a special built-in utility, therefore assignments are specified to remain after the command - not as intended for this use case. 2019-04-26 12:09:39 it becomes trick-ish to do without a subshell (wither iterate the items in a loop, or save IFS state+value and restore it on exit) 2019-04-26 14:49:03 hello, got a problem here with a server locking up from time to time, one week nothing but today two times in a row. it does not seem to be a panic (i guess) because it is set to reboot after 120 seconds on a panic, but it does not. also tried to work with magic sysrq keys, but it doesn't respond to that. does somebody have some hints how i should proceed from here? 2019-04-26 14:50:24 could it be a swap death? 2019-04-26 14:53:08 what is that, running out of swap? 2019-04-26 14:54:02 fassl: Did you try to use sysrq with a physical keyboard or iLO/iDRAC remote console? 2019-04-26 14:54:28 misterunknown, physical keyboard 2019-04-26 15:03:59 fassl: Ok. Do the processes still run, or is there not any hint of the system doing something? Maybe you could activate a per-minute logging of several things (ps aufx, free -m, netstat -tunep) to see at which point the server hangs. We did that recently to debug a weird error with a debian-based HPGen10 server. 2019-04-26 15:04:11 (Turned out to be a mainboard error) 2019-04-26 15:05:42 oops there it went again. misterunknown, physical keyboard but it seems dead as well, i guess i should try via serial? 2019-04-26 15:07:29 it seemed it locked up soonish after i started the samba services, they run on top of a 3-way zfs mirror on top of luks devices 2019-04-26 15:17:24 fassl: Ok. Do the processes still run, or is there not any hint of the system doing something? Maybe you could activate a per-minute logging of several things (ps aufx, free -m, netstat -tunep) to see at which point the server hangs. We did that recently to debug a weird error with a debian-based HPGen10 server. 2019-04-26 15:17:49 (I wrote that earlier) 2019-04-26 15:19:15 If it's somehow samba-related I'm not an expert, didn't use that lately. 2019-04-26 15:22:10 i will try to add logging 2019-04-26 15:23:52 would it make sense to add a beep every second to see whether this still works when it locks up? 2019-04-26 15:53:35 fassl: Well, I don't know if that brings any useful information, apart from the exact point in time. 2019-04-26 15:55:10 Even if you know the exact point in time, it would be hard to guess what caused this. 2019-04-26 15:56:11 For me, this sounds like a hardware issue anyway. Did you run any other Distros on this particular hardware without any issues? 2019-04-26 16:14:29 misterunknown, i ran centos for a while, but never this long, it did not seem to have issues but then also i did not run this infrastructure there 2019-04-26 16:19:16 i'll try a memtest 2019-04-26 16:20:13 dalias: in case you're interested, that's currently my preferred implementation for echo which ignores IFS, doesn't use subshell or alter variables, and is reasonably fast: https://0x0.st/zb2b.txt 2019-04-27 03:16:28 This is completely unrelated to alpine as a distro, but i'm trying to create a equivalent of QUBES OS based on KVM and alpine linux, you guys have any tips/packagges that I should install? 2019-04-27 03:17:06 Or any documentation on someone who ever did this? 2019-04-27 03:30:22 what's qubes? 2019-04-27 03:44:04 A security based distro using XEN 2019-04-27 04:38:17 mps: 2019-04-27 04:38:24 https://github.com/tcsh-org/tcsh/issues/12 2019-04-27 09:02:14 Any way of stripping down alpine further without compiling a custom kernel? 2019-04-27 09:02:38 I'm trying to get it on a 128mb disk 2019-04-27 09:07:20 <_ikke_> gardar: you could win some by specifically selecing what firmware packages you need, but that would require building your own linux-firmware virtual package 2019-04-27 09:07:47 <_ikke_> https://pkgs.alpinelinux.org/package/edge/main/x86_64/linux-firmware 2019-04-27 09:10:52 _ikke_: Thanks, I'll look into it.. Is there some quick way I could see from running the livecd what firmware I would need? 2019-04-27 09:14:14 <_ikke_> gardar: someone came up with this: modinfo `lsmod | grep -v Module | awk -e '{ print $1}'` | grep firmware 2019-04-27 09:16:47 _ikke_: Neat, thanks :) 2019-04-27 09:30:03 gardar: install linux-firmware-none and pick which firmware you need 2019-04-27 09:30:14 by default the installer installs all kind of firmware 2019-04-27 09:30:37 <_ikke_> tmhoang: Ah, nice, did not know that existed 2019-04-27 09:31:30 :D 2019-04-27 09:32:13 tmhoang: sweet, that saves me compiling a custom one, thanks :) 2019-04-27 09:33:49 gardar: I think 128mb should be ok for you. My existed system saying OK: 117 MiB in 58 packages after # apk add linux-firmware-none (and remove all linux-firmware-*). # df -h say 2019-04-27 09:34:00 /dev/vda2 9.7G 148.6M 9.0G 2% / 2019-04-27 09:34:03 /dev/vda1 92.8M 14.5M 71.4M 17% /boot 2019-04-27 09:34:36 I think this qemu vm is fresh. s390x though so should be minus/plus some MB on x86. 2019-04-27 09:37:58 x86_64 is way bigger ? OK: 235 MiB in 58 packages 2019-04-27 09:38:17 /dev/vg0/lv_root 7.8G 264.9M 7.1G 4% / 2019-04-27 09:38:19 /dev/sda1 92.8M 20.9M 65.0M 24% /boot 2019-04-27 09:39:59 Hmm do you think x86 might be smaller than x86_64? 2019-04-27 09:40:31 I can test for you. 1 sec. 2019-04-27 09:40:42 need to download kernel + initrd and install one ... 2019-04-27 09:40:46 few mins top 2019-04-27 09:56:25 anyone here tried installing Fedora latest 29/30 in text mode ? It sucks lol you can't do anything custom (lvm, swap, etc.) in text mode, you've got to do it in gui mode 2019-04-27 09:56:58 anaconda is the name for the installer 2019-04-27 10:00:32 I've always found the centos/fedora text based installer exceptionally bad, compared to the debian one 2019-04-27 10:02:37 gardar : x86 : OK: 175 MiB in 52 packages. /dev/sda1 92.8M 15.7M 70.1M 18% /boot . /dev/sda2 9.7G 207.9M 9.0G 2% / 2019-04-27 10:03:07 crossing checking if any difference in number of packages ... 2019-04-27 10:05:35 gardar: https://paste.ubuntu.com/p/KkQ9CfDCDc/ 2019-04-27 10:06:26 Thanks! 2019-04-27 10:07:00 oh you have vim installed, that's probably taking around 50mb 2019-04-27 10:09:00 around 30mb 2019-04-27 10:09:04 vim and dep : lua5.3-libs 2019-04-27 10:09:41 removing : libcom_err e2fsprogs e2fsprogs-libs 2019-04-27 10:10:29 mepholic: hi! I see, so we can expect new release soon 2019-04-27 10:11:26 gardar: I think you can remove syslinux once in a while ... 2019-04-27 10:11:52 as long as the mbr is installed, you are fine. upgrading kernel might require reinstalling syslinux 2019-04-27 10:12:15 you can do that from an external usb device so no need to install syslinux on your 128mb disk 2019-04-27 10:17:49 gardar: I think install dropbear instead of openssh* is one way. Another/final is to recompiling linux-vanilla. 2019-04-27 10:18:03 # for i in $(apk info ); do apk info $i | tail -n 3; done 2019-04-27 10:19:22 i'm surprised to see that there is no wireguard package in alpine 2019-04-27 10:19:41 Piraty: in testing/ 2019-04-27 10:20:06 https://pkgs.alpinelinux.org/packages?name=wireguard&branch=edge&repo=testing&arch=x86_64 2019-04-27 10:20:14 and should be moved to community, imo 2019-04-27 10:20:35 <_ikke_> if it's working good enough, feel free the submit a PR doing so 2019-04-27 10:20:52 Piraty: adding a wildcard would show wireguard-tools 2019-04-27 10:21:10 that's counter-intuitive 2019-04-27 10:21:17 mps: lets make it to 3.10 :) 2019-04-27 10:21:29 <_ikke_> wireguard is implemented in the kernel 2019-04-27 10:21:43 I talked about this with ncopa but he had split mind about this so I didn't insist 2019-04-27 10:22:07 _ikke_: it's a module 2019-04-27 10:22:16 <_ikke_> yes, so in the kernel 2019-04-27 10:22:20 so _ikke_ was right 2019-04-27 10:23:22 last time we talked about moving wireguard ncopa was in favor to move it for 3.10 2019-04-27 10:23:47 lets remind him on Monday I guess 2019-04-27 10:24:26 does alpine crossbuild the packages? in specific: wireguard-rpi ? 2019-04-27 10:24:33 tmhoang: will note your proposal and tell him that you agree with me ;) 2019-04-27 10:24:58 mps: +1 from me, if that's worth anythin ;) 2019-04-27 10:25:20 mps: Gigi D'Agostino - I'll Fly With You 2019-04-27 10:25:50 Piraty: wireguard is very good. I use it from time when the first versions are released and never had any issue 2019-04-27 10:26:51 Piraty: packages are native build. wireguard-rpi is a package built on 3 arm arch : https://git.alpinelinux.org/aports/tree/testing/wireguard-rpi/APKBUILD#n24 2019-04-27 10:27:17 yeah i'm looking at it 2019-04-27 10:36:59 <_ikke_> Piraty: alpine builds everything native 2019-04-27 10:37:17 <_ikke_> tmhoang: oh, what tmhoang said :-) 2019-04-27 10:37:19 saves some hazzle i guess 2019-04-27 10:37:42 <_ikke_> cross-compiling can give incorrect results 2019-04-27 10:50:29 I'm having trouble setting up persistent interfaces (via `/etc/network/interfaces`) in an alpine container, anyone got experience with that? 2019-04-27 10:50:49 loopback works, but e.g. a dummy interface doesn't. 2019-04-27 11:05:42 Found the solution, I needed to add pre-up iproute2 commands to create the interface first. Not the case for "lo". 2019-04-27 11:18:58 <_ikke_> did you add something like auto dummy before/ 2019-04-27 12:52:21 hello, I've been trying to reset password with by adding single and init=/bin/ash without success any other way to do it? I've searched without success 2019-04-27 12:53:04 I suppose the simplest way would be booting a live usb and changing the password from there 2019-04-27 12:54:29 really? why is it so difficult?!!! 2019-04-27 13:27:38 _ikke_: yes, added e.g. `auto dummy0` before. That doesn't seem to create an interface 2019-04-27 13:28:48 adrian_1908: did you added dummy in /etc/modules 2019-04-27 13:29:03 i.e. is module loaded 2019-04-27 13:30:44 oh, you are trying inside container 2019-04-27 16:56:39 how do I query "which packages does file /etc/xxx belong to?" 2019-04-27 16:57:36 <_ikke_> apk info -W /etc/xxx 2019-04-27 16:57:43 ahh, so easy 2019-04-27 17:00:13 interesting, the busybox cron is quite powerful 2019-04-27 17:05:02 _ikke_: in terms of organisation of the team, you mentioned recently some discussions. Is there anything public about it? 2019-04-27 17:06:26 <_ikke_> mailing lists, irc chatlogs mostly 2019-04-27 18:20:06 ikke i'm back haha 2019-04-27 18:21:54 I'm thinking of taking over a package that hasnt been updated in 4 months that is up to date on edge but 2 full releases behind on stable 2019-04-27 18:22:44 any recommendations on how to do it 2019-04-27 18:23:53 <_ikke_> nekocentral: stable releases usually only get bug / security fixes 2019-04-27 18:24:09 <_ikke_> If you have a little patience, 3.10 is just around the corner 2019-04-27 18:24:21 no i mean for a package 2019-04-27 18:24:22 nekocentral: write mail to current maintainer and alpine-devel first 2019-04-27 18:24:54 maybe fill issue on bugs.a.o 2019-04-27 18:25:02 <_ikke_> nekocentral: yes, I'm talking about packages as well 2019-04-27 18:25:15 yea but wierd it mps is that he keeps edge up to date but 3.7-3.9 is all 2 releases behind 2019-04-27 18:25:35 btw the package i'm taking about is Gitea 2019-04-27 18:25:55 <_ikke_> 3.10 will contain what is in edge at the point that it's branched 2019-04-27 18:26:03 a nice 2019-04-27 18:26:30 <_ikke_> so stable versions won't get major updates to packages 2019-04-27 18:27:04 <_ikke_> 3.9 is now on 1.6.4, so it will probably remain in 1.6.x 2019-04-27 18:28:22 yea was wondering as 1.7.0 -> 1.8.0 was a big update but 1.6.4 -> 1.7.* not to mush 2019-04-27 18:28:32 atleast from experience 2019-04-27 18:29:16 <_ikke_> right, but it's the policy to only backport patch releases 2019-04-27 18:29:31 a okay good to know 2019-04-27 18:29:59 <_ikke_> I must say, in general, there might be some exceptions to that rule, but they have to have a good reason 2019-04-27 18:30:17 like security reasons? 2019-04-27 18:30:31 <_ikke_> Well, that usually counts as a patch release 2019-04-27 18:30:42 nekocentral: you can backport it from edge if you need new features on stable 2019-04-27 18:31:27 yea sorry _ikke_ my mind doesn't work anymore made to many burgers today xD 2019-04-27 18:31:59 good to know mps 2019-04-27 18:34:03 going back to try to learn a bit ansible 2019-04-27 20:40:14 the wiki page ( https://wiki.alpinelinux.org/wiki/Alpine_Linux:Overview ) states grsec patches are applied to alpine's kernel builds. is this true? 2019-04-27 20:40:45 <_ikke_> Piraty: Not anymore 2019-04-27 20:40:54 Piraty: Yo 2019-04-27 20:40:57 that's what i assumed 2019-04-27 20:41:00 thanks 2019-04-27 20:41:06 hi north1 2019-04-27 20:41:15 Piraty: hi 2019-04-27 20:41:55 i remember you from some other channel, guess #voidlinux 2019-04-27 20:42:09 Piraty: Yes 2019-04-27 20:42:12 #xbps too 2019-04-27 20:42:16 probably 2019-04-27 21:16:02 are alpine's releases just static tags on the master branch or is work done on backporting stuff to version branches? 2019-04-27 21:21:34 Piraty: the versions (3.6, 3.7, 3.8, etc.) are git branches in the aports repo 2019-04-27 21:21:46 packages in them can be updated 2019-04-27 21:22:31 although the most common type of update is probably bug fixes 2019-04-27 21:22:41 rather than feature updates 2019-04-27 21:23:22 the master branch corresponds to 'edge', AFAIUI 2019-04-27 21:33:41 ok, skimming through wiki as first phaes of research, hoping they are kind of in sync with the real world 2019-04-27 21:38:48 The wiki is somewhat inaccurate in many cases. Reader beware! 2019-04-27 21:40:52 c705: yes, does Alpine have something like docs.voidlinux.org ? 2019-04-27 21:47:42 north1, Besides wiki.alpinelinux.org, I haven't found anything yet 2019-04-27 21:48:18 READMEs in https://github.com/alpinelinux might be useful as well 2019-04-27 21:49:16 I'm kind of hacking around Alpine myself right now - eventually once I really get my head wrapped around exactly what's going on, I may start contributing better documentation 2019-04-27 21:49:51 If anyone would like to share with me whatever documentation they have, that would be appreciated 2019-04-27 22:22:56 c705: try asking in #alpine-docs 2019-04-27 22:29:03 tcely: Thank you sir 2019-04-28 00:10:31 any alpine docker publisher here 2019-04-28 01:15:46 Anyone also hitting https://bugs.alpinelinux.org/issues/10309 ? 2019-04-28 03:20:19 I've got one more dumb question when following https://wiki.alpinelinux.org/wiki/KVM 2019-04-28 03:20:28 it says to do rc-update add libvirtd 2019-04-28 03:20:34 But it dosen't exist 2019-04-28 03:21:59 Try installing libvirt-daemon 2019-04-28 03:22:24 Do I need both? 2019-04-28 03:22:48 both what ? libvirt-daemon is a subpackage of libvirt 2019-04-28 03:24:09 Isn't the daemon suppose to be included in libvirt? 2019-04-28 03:24:18 It has been split 2019-04-28 03:24:29 meaning can I unnistall libvirt and only add libvirt-daemon ? 2019-04-28 03:24:38 you can't 2019-04-28 03:24:41 libvirt-daemon depends on libvirt 2019-04-28 03:24:43 Or I guess Libvirt-daemon is a dependacy on libvirt ? 2019-04-28 03:24:48 just install libvirt-daemon 2019-04-28 03:24:50 Aight ^^ 2019-04-28 03:26:03 Also, i have no clue why libvirt-daemon has /usr/sbin/libvirtd but keeps out other daemons 2019-04-28 03:26:17 I'm still getting this error 2019-04-28 03:26:18 Failed to connect socket to '/var/run/libvirt/libvirt-sock': No such file or directory 2019-04-28 03:27:18 I tought by starting libvirtd it would create it 2019-04-28 03:27:30 i guess it should 2019-04-28 03:27:37 try rc-status to see if it crashed 2019-04-28 03:28:56 It did? 2019-04-28 03:29:02 Well it says stopped 2019-04-28 03:31:30 Nvm, I got it 2019-04-28 03:31:34 But it still not working lol 2019-04-28 03:33:07 So here's another dumb question not related to alpine, i've got a container running which has brideg network, does that Failed to connect socket to '/var/run/libvirt/libvirt-sock': No such file or directory represents the container or the host ? 2019-04-28 03:36:32 you are running libvirt in a container? (lxc?) 2019-04-28 03:37:10 https://hub.docker.com/r/forumi0721alpinex64/alpine-x64-webvirtcloud 2019-04-28 03:37:26 This is what i'm trying to make it work 2019-04-28 03:37:41 It's just a webUI, for KVM/QEMU 2019-04-28 03:37:56 So this runs in a container, but the hypervisor runs bare-metal 2019-04-28 03:44:33 i have no experience with that product. if the error happens in the container that its an issue of the container. 2019-04-28 03:45:03 socket file has nothing to do with bridged networking 2019-04-28 03:45:42 Well I mean where is it "looking" for /var/run/libvirt/libvirt-sock 2019-04-28 03:45:48 I'm pretty sure it is looking in the container 2019-04-28 03:45:54 And needs to be looking on my host machine 2019-04-28 03:46:10 like i said, i dont know the product,do i dont know what its doing. 2019-04-28 03:46:18 so* 2019-04-28 03:46:33 Aight, thanks for your help ^^' 2019-04-28 03:46:44 Now I will need to pass it trought SSH i guess ? 2019-04-28 03:47:19 where does libvirt run? 2019-04-28 03:47:25 in the container or on the host? 2019-04-28 03:47:30 host 2019-04-28 03:47:50 Well I guess their is a libvirt ''client'' of some sort on the container? 2019-04-28 03:47:59 So it can mess with it 2019-04-28 03:50:04 and the container tries to connect to libvirt 2019-04-28 03:50:19 Yes, the daemon 2019-04-28 03:50:27 so if it uses a socket it need to be mounted between host and container 2019-04-28 03:50:56 You think that just mounting the socket would work? 2019-04-28 03:51:10 that should work yes 2019-04-28 03:51:52 I could try that, there is also a way to connect trough my host by SSH 2019-04-28 03:51:59 I think that would be easier 2019-04-28 03:52:55 cant it listen on a port? 2019-04-28 03:53:11 what is ''it'' ? 2019-04-28 03:53:18 libvirt 2019-04-28 03:53:29 good question 2019-04-28 03:53:59 most network based applications you can choose between socket or port 2019-04-28 03:54:22 both can be made work with docker 2019-04-28 03:54:31 you need to choose you weapon 2019-04-28 03:56:03 I understand, i'll try by SSH, the documentation is jsut very sparse 2019-04-28 03:56:17 https://github.com/retspen/webvirtmgr/wiki/Setup-SSH-Authorization 2019-04-28 03:56:27 I'm trying to understand what he's doing there 2019-04-28 04:47:29 Umm ... $ touch ~/.ssh/config && echo -e "StrictHostKeyChecking=no\nUserKnownHostsFile=/dev/null" >> ~/.ssh/config 2019-04-28 05:08:23 is there an rss feed for the packages site? 2019-04-28 05:18:04 nm, found it 2019-04-28 05:18:59 KH405_TV: creating an ssh config so the client won't check host keys. I guess he doesn't want to deal with making a confirmation dialog like ssh-askpass 2019-04-28 05:21:11 reason why you don't want to use regular virt-manager? 2019-04-28 06:22:49 doggone: Sorry for the late reply, it's a server, no UI 2019-04-28 06:23:17 I'm getting this error now: Error: internal error: Unable to get DBus system bus connection: Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory , DBUS is installed and started 2019-04-28 06:26:29 you can use virt-manager on your own computer 2019-04-28 06:26:39 what is throwing the error? the webvirtmgr? 2019-04-28 06:28:13 Yes 2019-04-28 06:28:53 You,re right, but I might have wanted to go and check my server with my cellphone or something 2019-04-28 06:29:25 check the location of the dbus socket, it might be wrong 2019-04-28 06:29:32 No, it's right 2019-04-28 06:30:05 maybe a permissions problem then? 2019-04-28 06:31:04 Umm, maybe? 2019-04-28 06:31:39 But yeah, I will install virt-manager on my laptop and probably use it there 2019-04-28 06:32:00 But my main PC is Winblows, and I want to check my server from my phone 2019-04-28 06:34:44 ah yea, and virt-manager doesn't really work on Windows, I think -- though you can install it via Cygwin 2019-04-28 06:35:33 Personally I use ConnectBot on my phone to SSH and then use virsh 2019-04-28 06:36:34 Yeah, but you can't check a windows machine :S 2019-04-28 06:36:54 to use the console then? 2019-04-28 06:37:08 You could use a SPICE client 2019-04-28 06:37:20 That's exactly what webvirtmanager is 2019-04-28 06:39:24 maybe you're right, maybe I should just install virt-maanger on my laptop 2019-04-28 06:39:28 Sorry, I'm not trying to persuade you :) I just dislike webinterfaces. 2019-04-28 06:44:35 I think you convinced me :P 2019-04-28 07:02:09 Now to install virt-manager 2019-04-28 07:02:24 I stil need libvirt and dbus on the host right doggone? 2019-04-28 07:36:11 Anyone using pulseaudio ? 2019-04-28 07:52:20 I'm installing alpine on my laptop, which will probably gonna use it ^^' 2019-04-28 07:52:38 good luck, pulseaudio on alpine is kinda broken :D 2019-04-28 07:52:53 Any other way to get audio :O ? 2019-04-28 07:53:06 Can use alsa really 2019-04-28 07:53:48 Never heard of that 2019-04-28 07:58:16 Do you know about virt-manager /qemu/kvm ? 2019-04-28 08:00:19 KH405_TV: I suppose I know virt-manager 2019-04-28 08:00:19 Used it for sometime for a VFIO VM 2019-04-28 08:01:18 I wanted to know what are the prerequisite for the host? (virt-manager) won't be installed on it, only QUEMU/KVM 2019-04-28 08:01:26 Do I need libvirt and dbus on my host ? 2019-04-28 08:05:50 Anybody running LXC on Alpine? I'm struggling with sharing a /dev entry from host to guest and wonder if this is LXC or Alpine speficic issue 2019-04-28 08:10:57 Also when I start virt-manager I get ''module not found libvirt'' 2019-04-28 08:11:05 Even tough it's installed and running 2019-04-28 08:11:37 KH405_TV: Uh? 2019-04-28 08:11:41 Virt-manager is a VM manager as the name suggests 2019-04-28 08:11:55 If you want to use bare Qemu you don't need it 2019-04-28 08:11:55 Yeah, you launch it using virt-manager 2019-04-28 08:12:31 Whenever I type virt-manager I get ''Module not found error: No module named libvirt'' 2019-04-28 08:19:45 Does anyone know how i can set the nfs threads persistently? i read something about putting a file in /etc/sysconfig but this directory does not exist. i know i can set it like this echo 48 >> /proc/fs/nfsd/threads, but it is not persisten after resetting the nfsd service 2019-04-28 08:34:50 KH405_TV: Maybe you're mixing up py2 virt-manager and py3 libvirt? 2019-04-28 08:39:53 Cogitri: can you explain? 2019-04-28 08:39:58 That's probably the issue 2019-04-28 08:40:11 Ohh, I understand what you mean 2019-04-28 08:41:09 How do i know whci one is installed tough ? 2019-04-28 08:41:42 I did apk add libvirt and virt-manager 2019-04-28 08:41:53 Did it install a different version of python of both ? 2019-04-28 08:59:09 KH405_TV: Well, it's possible that it's broken, virt-manager has been switched over to py3 very recently 2019-04-28 08:59:31 Any way to unbroke it? 2019-04-28 09:06:24 I can look at it once I'm home, okie? :) 2019-04-28 09:06:49 Sure, I might go to sleep soon, just whisper me if you find anything please? 2019-04-28 09:09:39 I can't find any PR about it on github 2019-04-28 09:19:49 Alright, what is your handle on GH? Can ping you then 2019-04-28 09:23:47 Just ping me here, Ill read it when i wake up! 2019-04-28 09:29:04 Also, should dbus be installed on my host or it's useless? 2019-04-28 09:30:04 Well, if it's required libvirt should pull it on 2019-04-28 09:32:32 In case anyone is interested in hacking on some alpine/open source related stuff with some other cool people: on 31st of May the hack4glarus.ch hackathon happens again and Alpine is a bit a bigger topic there 2019-04-28 09:32:49 Damn, where is it ? 2019-04-28 09:33:25 Ohh Switzerland 2019-04-28 09:33:29 That's a bit far from me :P 2019-04-28 09:33:39 KH405_TV: where are you at? 2019-04-28 09:33:48 Canada 2019-04-28 09:34:42 Ahh, very nice 2019-04-28 09:35:04 there is a saying here that canada is like big Switzerland - I've only been to Vancouver so far, so I cannot judge much 2019-04-28 09:35:22 Never been to swtizerland :P So can't really tell you 2019-04-28 09:35:37 There's a shit load of flooding happening right now tough 2019-04-28 11:05:26 does 'reboot' on alpine/busybox not unmount and sync filesystems? 2019-04-28 11:14:28 steinex: it would be surprising if it did not - why do you think so? 2019-04-28 11:21:56 steinex: Shouldn't OpenRC take care of that? 2019-04-28 11:57:52 Anybody is able to use virt-manager? 2019-04-28 15:31:12 hey I want to setup a headless torrent server with my old laptop and alpine. I was wondering what the busybox adduser option -S did. As I'm typing this I realize what I need to do to searx it 2019-04-28 15:31:27 what is a Systemuser in comparison to a non systemuser 2019-04-28 15:36:37 systemuser is for dbus I assume for example 2019-04-28 15:40:41 System users are users that aren't used by actual users but by daemons/services, yes 2019-04-28 15:40:59 ty 2019-04-28 15:46:53 dat feeling when you ssh into the headless and it works 2019-04-28 17:46:23 Is there a way to query for all packages that depend on a specific so: ? 2019-04-28 17:47:29 like list all packages that depend on so:libz.so.1 2019-04-28 17:49:09 apk info -r so:libz.so.1 2019-04-28 17:49:22 see apk info --help 2019-04-28 17:49:38 Doesn't include remote packages 2019-04-28 17:49:50 Sorry i wasn't clear 2019-04-28 17:52:11 remote? 2019-04-28 17:52:19 do you mean indirect dependencies? 2019-04-28 17:56:10 dalias: I believe remote mans not yet installed 2019-04-28 17:56:34 https://pkgs.alpinelinux.org/packages will also be of some help 2019-04-28 18:18:30 oh i see 2019-04-28 21:07:15 is there a way to just boot the netboot images from the downloads on alpine (initramfs, kernel, modloop) using grub? to get an installer from a working grub/fs 2019-04-28 21:07:42 i've tried specifying alpine_dev=sda1:ext2 modloop=modloop-vanilla (with sda1 being where the images are), but init craps out on boot 2019-04-28 21:11:12 it almost loops like the kernel/initramfs doesn't come with enough modules to actually get the modloop off disk 2019-04-28 22:18:14 How y'all doin... I'm trying to get alpine to boot up on little thin client I have axquired 3 of. 2019-04-28 22:18:14 I can't seem to get the wifi to work to update it. 2019-04-28 22:19:28 I'm pretty sure that it has a Ralink wifi chip, but I can't get it it show up with lsusb 2019-04-28 22:20:07 setup-interfaces doesn't show it eirther 2019-04-29 00:29:11 Quick support question: I’m seeing really long boot times with a fresh sys install of 3.9.3. busybox crond hangs for over 30 minutes. Any idea what’s wrong? 2019-04-29 00:29:56 It does eventually finish booting after 45 minutes or so, but clearly something is not right. 2019-04-29 00:41:36 asimpson: there were some changes for running out of entropy. Maybe that explains it. 2019-04-29 00:42:30 tcely: can I do anything to speed up the boot? 2019-04-29 00:43:03 figure out which daemons are using /dev/random and switch them to /dev/urandom (if this is the cause) 2019-04-29 00:43:45 tcely: thanks! 2019-04-29 00:47:48 tcely: do you know when those changes were introduced? Might it be a better move to downgrade to 3.8 instead? 2019-04-29 00:48:23 asimpson: the changes were to provide more entropy. I'm not sure a downgrade would help. 2019-04-29 00:48:57 this is all just guesswork from your described symptoms anyway 2019-04-29 00:50:07 tcely: I appreciate it! 2019-04-29 01:28:27 OneManHateGroup: some wifi devices are sdio 2019-04-29 01:31:08 altho idk if ralink makes an sdio card ;/ 2019-04-29 01:32:02 if there's nothing on the web about the thin client model and i had three of them, i'd take one apart to see what it's composed of 2019-04-29 07:16:58 hello, is the stress-ng package broken? i get open_by_handle_at: symbol not found and name_to_handle_at: symbol not found 2019-04-29 07:54:56 fassl: on edge or v3.9 ? 2019-04-29 07:54:58 edge seems fine 2019-04-29 08:07:52 tmhoang, v3.9 i'd say, at least thats what it says in /etc/issue , i tried to enable edge/main repo and apk add --upgrade stress-ng but i am not sure if this would then upgrade this tool to edge? 2019-04-29 08:09:26 fassl: don't do that. let me see with 3.9 2019-04-29 08:15:45 tmhoang, hm, i think it is not available in 3.9 and i got it from edge/testing 2019-04-29 08:16:52 fassl: now that makes sense :) 2019-04-29 08:17:09 that symbol should be in newer musl, big guess 2019-04-29 08:17:36 yeah i think so too 2019-04-29 08:18:13 so i should not mix the repos like i did and when i need something from edge/testing i should upgrade the whole system to that? 2019-04-29 08:18:26 fassl: if you really need it, you could open a PR and request to move it to community :) 2019-04-29 08:19:39 fassl: you are right, kind. But try what I suggested. It would help you. 2019-04-29 09:49:28 tmhoang, would that mean just moving the stress-ng package to community folder and the commit? 2019-04-29 09:49:31 tmhoang: i alrady moved it 2019-04-29 09:49:35 last week 2019-04-29 09:49:40 and cleaned it up 2019-04-29 09:50:09 hmm 2019-04-29 09:50:27 i didnt move it yet 2019-04-29 09:50:30 i will do it 2019-04-29 09:50:32 later 2019-04-29 10:27:59 fassl: clandmeter would help you :D 2019-04-29 10:42:27 Anyone knows how to set the threads for nfsd permanently? 2019-04-29 10:51:22 tmhoang, i guess it is not necessary anymore, he said he will move it ;) 2019-04-29 11:18:18 What package provides /usr/share/bkeymaps? setup-keymap doesn't work for me and I can't find it on p.a.o :c 2019-04-29 11:19:43 b2ed890e7f876a42229e6181268828c2493c972c 2019-04-29 11:19:49 its main/kbd 2019-04-29 11:20:25 kbd-bkeymaps 2019-04-29 11:24:44 we need a couple of volunteers who can help us moderate the bug tracker (redmine) 2019-04-29 11:24:51 no coding skills required 2019-04-29 11:25:14 some experience with redmine (and/or gitlab) is a plus but not required 2019-04-29 11:25:52 it seems that we have spam written by humans (for example https://bugs.alpinelinux.org/users/9969) 2019-04-29 11:26:06 so i think we may need humans that can help us deal with it 2019-04-29 11:27:31 what is done with spam postings like that? 2019-04-29 11:27:39 <_ikke_> removed 2019-04-29 11:27:56 I assume they just try to post links on important sites to try and inflate the pagerank on target sites 2019-04-29 11:28:12 esainane: thats what i think too 2019-04-29 11:28:25 could also be a part of the fakenews machinery 2019-04-29 11:28:34 <_ikke_> So removing them is the best cause of action 2019-04-29 11:28:41 but i think the intention is to get page rank 2019-04-29 11:28:59 I guess that's plausible, afaik there were countermeasures against links from things that looked like comments 2019-04-29 11:29:19 <_ikke_> We do want to allow relevant links to be posted 2019-04-29 11:29:26 ncopa: I have some experience with redmine, I built it for Alpine stable and maintain it for one of my customers 2019-04-29 11:29:30 my main point is that we could have use of some volunteers helping us with it 2019-04-29 11:29:45 would it not be better to modify redmine to change the link attributes? https://en.wikipedia.org/wiki/Nofollow 2019-04-29 11:29:56 <_ikke_> tcely: We still want the spam to be removed 2019-04-29 11:30:11 mps: you also have coding skills. i think this is something that someone without coding skills could help us with 2019-04-29 11:30:46 this volunteer could also help us assign the ticket to the package maintainer 2019-04-29 11:31:12 and maybe even ask the package maintainer to register to redmine if he/she does not have account 2019-04-29 11:31:32 right, I prefer to work with code, but if the help needed in this I'm ready to help till someone else take that role 2019-04-29 11:31:44 _ikke_: getting no boost from those links hopefully means we see less spam 2019-04-29 11:32:18 ncopa: a single sign on across alpine would help. I have a bugs account but I'm loathe to sign up again for wiki, whatever else 2019-04-29 11:33:51 I'm not sure I can make an ongoing commitment, but I could look into adding in rel="nofollow" to redmine comments 2019-04-29 11:34:14 tcely: https://bugs.alpinelinux.org/issues/335 2019-04-29 11:34:54 esainane: do you mean automatically add rel="nofollow"? 2019-04-29 11:35:21 i dont think we bother to add rel="nofollow" manually. better just delete in that case 2019-04-29 11:35:30 Yes, automatically 2019-04-29 11:35:54 I think adding it whereever class="external" is added to a link would be sufficient, but possibly not necessary 2019-04-29 11:35:58 Still looking 2019-04-29 11:36:15 esainane: can you please create a ticket on redmine, and add whetever you find to that ticket 2019-04-29 11:36:43 <_ikke_> Can't seem to quickly find an existing plugin for that 2019-04-29 11:36:45 On redmine's redmine, or alpine's redmine? :) 2019-04-29 11:36:59 esainane: on alpine's redmine 2019-04-29 11:37:01 There was already some work done it looks like, http://www.redmine.org/projects/redmine/repository/revisions/2334/diff/trunk/lib/redmine/views/other_formats_builder.rb - just not for all cases 2019-04-29 11:37:18 http://www.redmine.org/issues/5249 2019-04-29 11:38:33 tcely, good find, looks like that implements what I was going to do 2019-04-29 11:38:40 i think we can probably find volunteers to help us manage some of things manually, brute force 2019-04-29 11:38:58 there are problably people who wants to contribute, but does not have coding skills 2019-04-29 11:41:03 maybe i should ask on alpine-user mailing list too 2019-04-29 11:41:17 for people in other timezones 2019-04-29 11:47:21 Worth a shot 2019-04-29 12:00:45 Anybody knows if there is a problem with libvirt / libvirtd / Dbus ? 2019-04-29 12:01:28 Everything looks like it's working fine and i'm getting failed to connect to socket: /var/run/dbus/system_bus_socket : No such file or directory 2019-04-29 12:01:54 Well, does that dir exist? 2019-04-29 12:02:05 node1:/run/dbus# ls -a 2019-04-29 12:02:05 . .. system_bus_socket 2019-04-29 12:02:05 node1:/run/dbus# 2019-04-29 12:02:42 /run/dbus vs /var/run/dbus ? 2019-04-29 12:03:11 symlinks 2019-04-29 12:03:12 I don't know why /var/ dosen't show 2019-04-29 12:03:38 /var/run is a symbolic link to /run 2019-04-29 12:04:08 unless it hasn't been created.. 2019-04-29 12:04:25 Try showing us the output from: ls -l /var/run /var/run/dbus/system_bus_socket 2019-04-29 12:05:00 lrwxrwxrwx 1 root root 4 Apr 2 20:20 /var/run -> /run 2019-04-29 12:05:00 srwxrwxrwx 1 root root 0 Apr 29 07:57 /var/run/dbus/system_bus_socket 2019-04-29 12:06:26 U okay bot? 2019-04-29 12:06:28 algitbot: what did you see? 2019-04-29 12:07:20 saw that date of the system_bus_socket and took pr from it =) 2019-04-29 12:07:42 lol 2019-04-29 12:08:04 So is one of the packages borked ? 2019-04-29 12:08:35 What's this show: ls -ld / /var /run /run/dbus 2019-04-29 12:09:36 I'm sorry mr Bot 2019-04-29 12:09:37 drwxr-xr-x 23 root root 4096 Mar 21 14:04 / 2019-04-29 12:09:37 drwxr-xr-x 9 root root 480 Apr 29 07:57 /run 2019-04-29 12:09:37 drwxr-xr-x 2 root root 60 Apr 29 07:57 /run/dbus 2019-04-29 12:09:37 drwxr-xr-x 11 root root 4096 Apr 2 20:20 /var 2019-04-29 12:10:00 might be better to use hastebin or something like that for the paste 2019-04-29 12:10:09 Yeah, you're right 2019-04-29 12:10:12 I have no ideas left on why it's claiming that socket doesn't exist 2019-04-29 12:10:15 looks like algitbot does not like 29 Apr 2019-04-29 12:10:28 A*pr* 2019-04-29 12:10:46 algitbot: please behave 2019-04-29 12:10:53 Apr 28 2019-04-29 12:11:04 ncopa: grep "\' ;-) 2019-04-29 12:11:42 im thinking: [^a-zA-z]pr 2019-04-29 12:12:09 1pr6565 2019-04-29 12:12:29 ACTION shrugs 2019-04-29 12:13:10 Anyone has this problem with dbus ? 2019-04-29 12:14:17 KH405_TV: unless you were in here before for the same issue it does sound familiar. I haven't tried it myself though. 2019-04-29 12:14:28 1pr6565 2019-04-29 12:14:34 pr6565 2019-04-29 12:14:43 The bot is gonna hate you 2019-04-29 12:14:48 I wasn't tcely 2019-04-29 12:14:49 pr#6565 2019-04-29 12:15:11 CVE 2016 ^^ 2019-04-29 12:15:16 Apr#6565 2019-04-29 12:15:22 huh 2019-04-29 12:15:32 GH-7000 2019-04-29 12:15:50 Is it triggering on # ? 2019-04-29 12:16:17 Apr6565 2019-04-29 12:16:17 Apr#29 2019-04-29 12:16:25 Apr#6565 2019-04-29 12:16:41 pattern = "[^0-9A-Za-z][Pp][Rr] ?#?(%d+)", 2019-04-29 12:16:54 its a Lua pattern 2019-04-29 12:17:11 I would have guessed python for the bot 2019-04-29 12:17:30 im relocating. bbl 2019-04-29 12:21:13 It also says verify the libvirtd is running on the host 2019-04-29 12:21:18 Which it is 2019-04-29 12:21:50 And where is this /var/run/dbus/system_bus_socket should be ? 2019-04-29 12:21:56 host or client? 2019-04-29 12:25:27 ncopa: you might try this: pattern = "%s-[Pp][Rr][%s#]-(%d+)" 2019-04-29 12:28:57 Now i'm getting : no polkit agent available to authenticate your action 2019-04-29 12:30:31 Install one of the polkit-* 2019-04-29 12:30:36 (e.g. polkit-gnome) 2019-04-29 13:01:19 https://pkgs.alpinelinux.org/package/edge/main/x86_64/polkit 2019-04-29 13:01:22 Would this work ? 2019-04-29 13:09:32 That's just Polkit itself 2019-04-29 13:09:53 You have to install a polkit agent (think of it as a frontend for the polkitd daemon) 2019-04-29 13:33:48 Should I install the same one as my background 2019-04-29 13:34:01 Windows manager * 2019-04-29 13:37:14 Or DE ? 2019-04-29 13:37:18 I'm using awesome ... 2019-04-29 14:27:24 I'm sorry this is sort of a noobish question, i'm just wondering what the differnece bewteen the multiple clients are 2019-04-29 14:33:03 Or which one would you recommend ? 2019-04-29 14:44:57 KH405_TV Are we talking about Alpine release flavours? 2019-04-29 14:50:30 Best polkit client ^^' 2019-04-29 14:51:02 A bit unreleated to alpine but I haven't figure out how to fix my issue which might be related to alpine ? 2019-04-29 14:59:51 Once I downlaoded mate-polkit, how do I start it ? 2019-04-29 15:03:54 mepholic: here is the patch to add csh.cshrc and csh.login to tcsh https://patchwork.alpinelinux.org/patch/4823/ 2019-04-29 15:16:59 Why can't I just rc-service mate-polkit start ? :O 2019-04-29 15:23:41 KH405_TV:having it installed is all you need iirc 2019-04-29 15:23:52 KH405_TV: it will be called 2019-04-29 15:26:50 It's not working :( 2019-04-29 15:28:14 Are you sure about that ? 2019-04-29 15:29:17 so i'm setting up an alpine deployment and just want to make sure i'm doing a few things right 2019-04-29 15:29:36 KH405_TV: did you installed mate-desktop and start lightdm 2019-04-29 15:30:32 i want normal users to have rw for /dev/ttyUSB0, i set up an /etc/init.d/foo script to check for /dev/ttyUSB0 and set permissions to 666, is there a better way to do this? 2019-04-29 15:31:14 geosmin: add user to dialout group 2019-04-29 15:31:37 mps: i did that initially but that gives them rw to other files as well 2019-04-29 15:32:08 that said, i can't really see any risks involved there... 2019-04-29 15:32:26 ah, you want to limit normal user to only ttyUSB 2019-04-29 15:32:51 i guess i should mention this is a minimal "workstation" tool. launches chromium kiosk pointing to company webapp 2019-04-29 15:32:58 No, I'm using i3WM 2019-04-29 15:33:05 and not allow access to ttyS0 for example 2019-04-29 15:33:20 /dev/ttyUSB0 is to read values from a piece of equipment and use xdootol to fill forms 2019-04-29 15:33:33 mps: right, but i don't see much risk in allowing other ports 2019-04-29 15:34:45 then, you can edit /etc/mdev.conf and adjust according your needs 2019-04-29 15:35:23 KH405_TV: does it start consolekit at start 2019-04-29 15:35:26 mps: exactly what i was looking for, thank you 2019-04-29 15:37:18 mps: No it dosen't do I just add consolekit? 2019-04-29 15:38:05 then, you have to use X DM which starts consolekit and polkit for you 2019-04-29 15:38:48 iirc, slim DM does that 2019-04-29 15:39:07 I'm not sure I'm following you 2019-04-29 15:41:03 I don't have consolekit and polkit installed so cannot check it now, but I remember when I had them they were started by slim DM 2019-04-29 15:46:20 How exactly do I do that? 2019-04-29 15:46:28 just write slim DM ? 2019-04-29 15:47:20 apk add slim 2019-04-29 15:49:33 The SLiM project appears to have been abandoned and the project homepage is dow :( 2019-04-29 15:54:13 Is there another one that is maintained ? 2019-04-29 15:57:00 maybe it is abandoned upstream but it works quite fine 2019-04-29 15:59:54 mps: edited mdev.conf but permissions haven't changed 2019-04-29 15:59:57 Can I just add it to .xinitrc ? 2019-04-29 16:00:50 maybe, but I didn't tried so not sure. but you can try and see if it works 2019-04-29 16:00:56 ttyUSB[0-9] root:dialout 0666 @ln -sf $MDEV modem 2019-04-29 16:01:11 crw-rw---- 1 root dialout 188, 0 Apr 29 11:51 /dev/ttyUSB0 2019-04-29 16:01:19 I just add mate-polkit? 2019-04-29 16:01:21 geosmin: reload mdev or restart machine 2019-04-29 16:01:29 mps: this is post-restart 2019-04-29 16:01:55 geosmin: not sure, try and see 2019-04-29 16:03:22 I remember that I added patch to put ttyUSBx in dialout group and in worked, but forgot what I did to activate it 2019-04-29 16:04:10 KH405_TV: could you start consolekit by invoking '/etc/init.d/consolekit start' 2019-04-29 16:04:43 mps: /etc/init.d/mdev does not reference mdev.conf (though perhaps the binary does by default) 2019-04-29 16:05:36 note /etc/mdev.conf says "this is a sample file" commented out at the top 2019-04-29 16:06:21 /etc/init.d/consolekit start 2019-04-29 16:06:30 sample from upstream, but distributors usually extend it 2019-04-29 16:07:01 Oops, file is not found it says :( 2019-04-29 16:07:01 KH405_TV: does it starts 2019-04-29 16:07:10 No :( 2019-04-29 16:07:36 aha, so you didn't installed it 2019-04-29 16:08:25 Install what ? 2019-04-29 16:09:07 consolekit2 2019-04-29 16:09:30 No ^^' I'm pretty new to alpine, sorry 2019-04-29 16:09:31 What is this ? 2019-04-29 16:09:36 and, sorry, it doesn't have /etc/init.d/ file 2019-04-29 16:09:55 I'm trying to stick to a minimal install the most I can 2019-04-29 16:10:04 So i'm paying attenntion to what I install :P 2019-04-29 16:10:11 looks like it needs dbus started to be activated 2019-04-29 16:11:07 why not install slim which depends on consolekit and polkit and dbus 2019-04-29 16:11:31 then when you start slim all these should run 2019-04-29 16:12:39 geosmin: 'mdev -h' says 'mdev -s is to be run during boot to scan /sys and populate /dev. 2019-04-29 16:13:27 to have hotplug you should add udev apk 2019-04-29 16:14:57 mps: 'echo /sbin/mdev >/proc/sys/kernel/hotplug' worked, unsure how to do that at boot though 2019-04-29 16:15:48 this can be set in /etc/sysctl.d/ file 2019-04-29 16:15:59 mps: I'll read on the subject a little bit more, but I don't feel comfortable to install something that the last time it has been updated is 20143 2019-04-29 16:16:02 2013 * 2019-04-29 16:17:00 KH405_TV: why not, there are a lot of software which are not updated every year 2019-04-29 16:17:03 mps: so 'echo $previous_echo_cmmand >> /etc/sysctl.d/foo.conf'? 2019-04-29 16:17:45 geosmin: no, look at sysctl guide somewhere how to set parameters in files 2019-04-29 16:18:40 it should be 'kernel.hotplug=/sbin/mdev', iirc 2019-04-29 16:19:14 It's been abandonned for 6 years 2019-04-29 16:19:24 Do you know how many CVE come out each year lol 2019-04-29 16:19:27 KH405_TV: you can also use lxdm, gdm, lightdm... 2019-04-29 16:20:02 KH405_TV: give me CVE for slim, I will try to fix it 2019-04-29 16:20:40 lightdm looks cool 2019-04-29 16:20:58 If the CVE isn't out yet dosen,t mean it dosen't exist 2019-04-29 16:21:04 there is a fork which has seen some development lately - https://github.com/PeteGozz/slim 2019-04-29 16:23:35 if no one can show a bug in software does it means there is bug in this software or just guessing in dark 2019-04-29 16:26:38 php is updated 'daily' and there is always some bugs, so fresh releases doesn't mean that the software are secure 2019-04-29 16:27:31 but this will be better for #alpine-offtopic 2019-04-29 16:30:09 Not saying newer software don't have bugs :P But at least they will or might het fixed 2019-04-29 16:30:15 But Yeah i understand you're point 2019-04-29 17:58:24 where can I find version release dates? I don't see it on this post: https://www.alpinelinux.org/posts/Alpine-3.9.0-released.html 2019-04-29 17:58:48 <_ikke_> https://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases 2019-04-29 17:58:53 ah, I see it in git log 2019-04-29 17:59:09 oh, that's prettier; thanks _ikke_ 2019-04-29 17:59:34 not sure it includes release dates of minor versions 2019-04-29 17:59:35 _ikke_: shouldn't you be helping me do something in #git right now? 2019-04-29 17:59:42 :P 2019-04-29 17:59:45 <_ikke_> haha :D 2019-04-29 17:59:56 <_ikke_> ncopa: it doesn't 2019-04-29 18:00:16 no, i think _ikke_ should enjoy life on his balcony right now ;) 2019-04-29 18:00:58 <_ikke_> If only I had one :D 2019-04-29 18:01:03 lol 2019-04-29 18:01:59 gah; I always want to say in my brain ampq instead of amqp 2019-04-29 18:26:55 so I was having issues getting xdebug to install; it was claiming that autoconf was missing despite me installing it on it's own step; I figured this was perhaps an issue with 3.9 being that vars are set differently from version to version sometimes 2019-04-29 18:27:06 py3-urllib3 is currently v1.25.1, but certbot requires urllib3 1-21.1 to 1.25 -- Requirement.parse('urllib3<1.25,>=1.21.1') 2019-04-29 18:27:32 so I was informed that "an autoreconf might help" ... I thought the idea was to just install autoreconf 2019-04-29 18:28:07 but autoreconf is part of the autoconf package... so should I just add `autoreconf` as a standalone command to my container build? 2019-04-29 18:28:27 can fix with "pip3 install --upgrade certbot" which installs older version of urllib3 2019-04-29 18:29:00 namely urllib3-1.24.2 2019-04-29 18:30:35 this applies to edge 2019-04-29 18:32:40 hmm... whatever issue I was having seems to have disappeared 2019-04-29 18:36:45 so wait; for docs that take the form of `-f, --foo BAR` can you use BAR with -f ? can you exempt BAR from --foo ? 2019-04-29 18:38:22 what are foo and bar in this case 2019-04-29 18:39:19 opal: foo is a flag and bar is a package 2019-04-29 18:39:21 oer 2019-04-29 18:39:22 erm 2019-04-29 18:39:31 bar is an argument 2019-04-29 18:39:39 for the flag 2019-04-29 18:39:57 so like `-t, --virtual NAME` is in the docs 2019-04-29 18:40:12 yeah thats what i asked, what were they *in this case* 2019-04-29 18:40:15 uh let me see 2019-04-29 18:40:57 opal: I mean... that would be consistent for all flags though, right? 2019-04-29 18:41:08 not for all documentation in my experience 2019-04-29 18:41:21 :\ 2019-04-29 18:41:24 some documentation says --flag=arg, some say --flag arg, others say... 2019-04-29 18:41:29 yeah its inconsistent i hate it too 2019-04-29 18:41:49 -t and --virtual do not throw an error when not using NAME 2019-04-29 18:42:04 (imo, it should be `[NAME]` then) 2019-04-29 18:42:14 yeah that would be nice if they denoted it that way 2019-04-29 18:42:20 what program is this again that youre trying to use 2019-04-29 18:42:25 apk 2019-04-29 18:42:40 ah 2019-04-29 18:42:57 also, I don't know what `apk del` does when you don't specify a package 2019-04-29 18:43:05 I guess "all" packages 2019-04-29 18:43:19 it deletes no packages 2019-04-29 18:43:27 same for add 2019-04-29 18:43:35 erm 2019-04-29 18:43:45 it seemed to delete something when I added gcc 2019-04-29 18:43:57 oh wow 2019-04-29 18:44:01 I added gcc with -t flag 2019-04-29 18:44:09 and it deleted it with apk del 2019-04-29 18:44:36 or... hmm 2019-04-29 18:44:43 I am getting inconsistent results 2019-04-29 18:45:49 i see the -static packages are coming through, thats nice 2019-04-29 18:46:16 I guess you're right 2019-04-29 18:46:19 I must have imagined it 2019-04-29 18:46:36 i cant even see any documentation on the -t/--virtual flag for apk lol 2019-04-29 18:46:43 not in `apk --help --verbose`, not on wiki 2019-04-29 18:46:47 anyway... so does --virtual alone actually make it virtual? 2019-04-29 18:46:59 i dont know, never encountered this flag 2019-04-29 18:47:09 opal: ever? 2019-04-29 18:47:16 what documentation are you looking at because i want to read up on it 2019-04-29 18:47:36 there are quite a few Dockerfile's that use it 2019-04-29 18:47:45 um... just `apk add --help` 2019-04-29 18:48:25 I often see `apk --no-cache --virtual .build-deps make g++ autoconf ...` and then `apk del .build-deps` 2019-04-29 18:48:46 ah ok 2019-04-29 18:49:26 so in this case i think -t / --virtual actually takes no args; the NAME is just the normal argument(s) that `apk add` takes 2019-04-29 18:49:38 i think 2019-04-29 18:49:39 oh wait 2019-04-29 18:49:42 no 2019-04-29 18:49:44 no never mind 2019-04-29 18:49:46 yeah 2019-04-29 18:49:49 :) 2019-04-29 18:49:49 yeah NAME is naming the virtual itself 2019-04-29 18:49:51 lol 2019-04-29 18:49:54 indeed 2019-04-29 18:49:59 seems to work without it though 2019-04-29 18:50:13 apk: option requires an argument: t 2019-04-29 18:50:18 not for me 2019-04-29 18:50:19 imo, if it requires a NAME to change the normal behaviour of apk, it should error 2019-04-29 18:50:30 wtf 2019-04-29 18:50:37 did you do `apk add -t gcc` or something 2019-04-29 18:50:40 what version? 2019-04-29 18:50:42 it might have made a virtual package called gcc 2019-04-29 18:50:50 2.10.3 2019-04-29 18:51:09 ^ is the alpine version? 2019-04-29 18:51:19 apk-tools 2.10.3, compiled for x86_64. 2019-04-29 18:52:00 oh, 2019-04-29 18:52:16 yeah, we're using the same tool 2019-04-29 18:52:19 heres the thing 2019-04-29 18:52:27 if it were just making a package named gcc 2019-04-29 18:52:31 then it would show me installs 2019-04-29 18:52:34 and it does 2019-04-29 18:55:02 hmm 2019-04-29 18:55:39 so `apk add -t NAME ...` ignores ... and installs NAME 2019-04-29 18:57:25 however 2019-04-29 18:57:36 `apk -t add ...` installs all ... 2019-04-29 18:57:50 I don't know if there is some default reference for -t that you can delete after 2019-04-29 18:58:20 oh snap; you can do apk del add 2019-04-29 18:58:36 which means it is in fact making a virt package named add 2019-04-29 18:59:02 crazy what you're able to do 2019-04-29 19:19:20 ok so busybox-static installs /bin/busybox.static; is there a way to replace the system busybox install without confusing apk 2019-04-29 19:27:43 misterunknown, i might have found the issue leading to a freeze. i could reproduce it twice stressing the pfSense VM and saw that its root fs was broken. after a fsck i stressed it again, 3 times as long as it previously took to freeze two times, also stressed the cpu additionally and it was fine. would that make sense? a broken vm root fs freezing the entire host? 2019-04-29 20:17:00 fassl: Might be. I didn't experienced a broken fs yet, but heared that it can cause some weird errors. 2019-04-29 20:18:12 It's quite a strong argument, if you can't reproduce this anymore... 2019-04-29 20:27:25 misterunknown, i just really hope it was the issue 2019-04-29 20:54:47 fassl: I know that feel... We have a similar issue on one of our customers server, which occasional freezes, and we only see some nullbyte characters in the syslog. 2019-04-29 20:55:16 Reboot helps, but it's not reproducable. 2019-04-29 21:05:51 misterunknown, ouch :( such things suck big time 2019-04-29 21:07:39 a friend of mine once cut his pc case for installing custom water cooling and sometimes the system would just crash... any stress ran fine for hours, normal use ran for weeks sometimes... turns out a tiny piece of the case ended up in one ram slot 2019-04-29 21:07:54 Ooooof. 2019-04-29 21:37:16 Hi all. I'm trying to get a thin client that I found to do something fun/useful with alpine. I'm having trouble getting the wifi to work and none of ways I'v read help. 2019-04-29 21:39:12 Does your ethernet still work? 2019-04-29 21:40:13 dunno... it seems to load up eth0 ok, but I don't have a place to plug it in 2019-04-29 21:40:51 I could plug it in to this lappy I'm on, but I can't get the 'bridge" to work 2019-04-29 21:41:15 Ah, I see. 2019-04-29 21:41:41 Does the network interface show up in `ip link`? 2019-04-29 21:41:53 lemme see 2019-04-29 21:43:15 can you even bridge? i had to forward traffic from/to wifi iirc 2019-04-29 21:44:12 it says I have lo: and eth0 2019-04-29 21:44:49 But no wireless device. 2019-04-29 21:45:03 correct 2019-04-29 21:45:28 Hmm. I wonder whether there's some screwy config in /etc/network/interfaces 2019-04-29 21:45:42 It's a Ralink chip 2019-04-29 21:46:26 To be honest, that means nothing to me. I don't know any troubleshooting steps specific to ralink chips. 2019-04-29 21:47:03 thats cool 2019-04-29 21:47:45 But yeah, have you correctly populated /etc/network/interfaces? 2019-04-29 21:48:23 with setup-interfaces? 2019-04-29 21:49:17 Is it fair to assume that setup-interfaces works correctly? I've always done it by hand. 2019-04-29 21:50:26 If you could help me bridge my wifi to my eth0 on this lappy I could probably work it out 2019-04-29 21:51:33 If you cat the file, is there at least a listing for a wireless device? If so, we could probably find out more information from init's startup logs. 2019-04-29 21:51:42 Honestly that's not my repertoire. 2019-04-29 21:52:08 which file am I 'cat' ing? 2019-04-29 21:52:32 `/etc/network/interfaces` 2019-04-29 21:52:37 aight 2019-04-29 21:53:14 only lists eth0 2019-04-29 21:53:56 Well, I suppose that's your problem then. Without a listing for a wireless interface, init doesn't know to set one up. 2019-04-29 21:54:20 sounds good 2019-04-29 21:57:55 does lspci -nnnk show any driver for your wifi device? (i am not sure if this command works in alpine) 2019-04-29 21:59:11 it worked 2019-04-29 22:00:48 any kernel driver in use? 2019-04-29 22:01:01 how do I tell that 2019-04-29 22:01:03 As in, you have a network connection? 2019-04-29 22:01:21 I have no connection 2019-04-29 22:01:43 it should list you devices, and then indented some information about it, some lines could contain "Kernel driver in use" 2019-04-29 22:01:56 Ah, you mean `lspci` works. My apologies for the misunderstanding. 2019-04-29 22:03:13 yeah lspci works 2019-04-29 22:04:17 it doesn't show anything about wifi, or wlan, or eth0, or Network 2019-04-29 22:04:35 those are all the things that I sorted through 'grep' 2019-04-29 22:13:35 Okay. On my system, `lspci -nnnk` contains an entry with the model of my wireless chip. Maybe try grepping with Ralink. 2019-04-29 22:15:05 seems like it just understands m and k 2019-04-29 22:16:41 lspci -mk | grep -e 02 -e 0e should give you network controllers and wireless controllers, but i don't know if wireless controllers in this case would mean something like a xbox controller :D 2019-04-29 22:16:59 0d sorry 2019-04-29 22:17:46 s/0d/0d instead of 0e/ 2019-04-29 22:18:01 fail 2019-04-29 22:20:25 s/0d/0d\ instead\ of\ 0e/ 2019-04-29 22:25:11 sorry I had a customer 2019-04-29 22:25:29 Neat. What business are you in? 2019-04-29 22:25:47 sandwiches 2019-04-29 22:26:43 doesn't grep for ralink 2019-04-29 22:26:51 Very cool. 2019-04-29 22:28:12 I'm gonna tap out. Have a good one. 2019-04-30 00:13:51 does anyone know if it's possible to use separate separate routing tables for interfaces using DHCP? 2019-04-30 00:14:11 s/separate // 2019-04-30 01:18:19 doggone: not by default, but you can hook into dhclient script to do just about anything you like on DHCP responses. 2019-04-30 09:34:56 Running Alpine version 3.9 but need an older version of openjdk8-jre-base than the default version 8.201... How can I achieve this? 2019-04-30 09:47:34 <_ikke_> Depends a bit on the dependencies of openjdk8-jre-base 2019-04-30 09:49:43 I surely don't have to install an older Alpine version to achieve this, right? 2019-04-30 09:51:07 <_ikke_> Well, if it depends on older versions of libraries, it's going to be tricky 2019-04-30 09:51:27 <_ikke_> You could try to build the package yourself on 3.9 in that case 2019-04-30 09:57:05 I read that apk can keep packages back as part of an upgrade. Just a thought: can I add use older repositories and force apk to use older repo? 2019-04-30 09:57:50 <_ikke_> You would basically downgrade to an older alpine version 2019-04-30 10:22:57 Changed repos to 3.4, ran apk and it did the trick. 2019-04-30 10:24:01 pvdp: the problem with downgrading is leaving /etc in a state that can break things. 2019-04-30 10:25:18 OTOH, reinstalling an older Alpine happens in a matter of seconds 2019-04-30 10:27:21 OK...so what would be the right way to go about this? 2019-04-30 10:28:24 well it depends what you want to do. 2019-04-30 10:28:54 you can pin an older repo and add the package and pray to whatever god that it works. 2019-04-30 10:28:55 Just need an older jdk8 version < 8.152 2019-04-30 10:29:27 3.4 is pretty old 2019-04-30 10:29:38 i think it will break things 2019-04-30 10:29:50 Yes, but it has an old enough version. 2019-04-30 10:30:11 + seems to work ok. 2019-04-30 10:30:51 if you downgrade your whole system that would proably work 2019-04-30 10:30:54 but its not supported by us 2019-04-30 10:31:12 so we will remember your nickname and put you on ignore :p 2019-04-30 10:31:19 I understand. 2019-04-30 10:31:31 you can also use docker or lxc 2019-04-30 10:32:20 Yes, i know. 2019-04-30 10:32:37 and then there's always the way of packaging what you want and using that package to install. a bit more work yes, but you're more likely to get what you want 2019-04-30 10:33:57 I probably have to got that path at some stage, yes. Just taking baby steps at the time. 2019-04-30 10:34:05 yes, thats the only sensible way around it. 2019-04-30 10:34:48 and it's not quite as difficult as it may sound at first 2019-04-30 10:35:13 like swimming 2019-04-30 10:35:38 Can I quote you on that one when I get stuck TBB ? 2019-04-30 10:37:18 go ahead. it's a safe statement if you start with the presupposition that packaging is _insanely_ hard work 2019-04-30 10:38:05 hahaha so true 2019-04-30 10:38:52 Thx for the help. Appreciated 2019-04-30 10:39:46 if you go down that path eventually, you can bother me about it and I'll help if possible 2019-04-30 10:40:06 Thx 2019-04-30 13:20:55 hello, i installed linux-firmware-rtl_nic but still i get a direct firmware load error message with error code -2. the file it tries to load is in /lib/firmware, what am i doing wrong? 2019-04-30 13:21:28 <_ikke_> DId you check dmesg? 2019-04-30 13:24:23 yep: Direct firmware load for rtl_nic/rtl8168e-3.fw failed with error -2 2019-04-30 13:30:35 you must be on 3.9 2019-04-30 13:31:07 3.9 has the firmware. hmm 2019-04-30 13:33:54 <_ikke_> what does uname -r return? 2019-04-30 13:39:49 fassl: modprobe r8169 ? 2019-04-30 13:49:58 tmhoang, yeah on 3.9 2019-04-30 13:50:00 ikke 4.19.29-0-vanilla 2019-04-30 13:50:20 but i installed zfs-vanilla or whatever it is called 2019-04-30 13:50:32 # modprobe r8169 would load the firmware for you 2019-04-30 13:51:29 isn't it already loaded as i got that error message?= 2019-04-30 13:53:07 anyway i issued that command, nothing new came up in dmesg 2019-04-30 14:01:02 hm, some people suggest CONFIG_FW_LOADER_USER_HELPER_FALLBACK=y 2019-04-30 14:01:43 <_ikke_> That would mean you would need to recompile the kernel 2019-04-30 14:02:12 yeah 2019-04-30 14:03:16 but i am not sure it would help me 2019-04-30 14:03:49 "Enabling this option forces a sysfs userspace fallback mechanism to be used for all firmware requests which explicitly do not disable a a fallback mechanism. Firmware calls which do prohibit a fallback mechanism is request_firmware_direct()." 2019-04-30 14:04:10 my error message got "Direct firmware load" 2019-04-30 14:09:17 do i maybe need to put it into initfs? 2019-04-30 14:27:40 hey I'm learning how to maintain headless and I installed eudev, which is supposed to be added via rc-update add udev. that service fails though with "Error relocating /sbin/udevd: name_to_handle_at: symbol not found; start-stop-daemon: failed to start /sbin/udevd. Is this something to ask in #gentoo-udev or am I a lazy bastard for not searching the wiki well enough? 2019-04-30 14:28:50 <_ikke_> zava: Did you combine repos from different versions (stable / edge)? 2019-04-30 14:29:17 I undid all the # from /etc/apk/repositories 2019-04-30 14:30:24 <_ikke_> can you paste it somewhere? 2019-04-30 14:31:19 I had them combined, just commented out the main and left the edge 2019-04-30 14:32:26 now im upgrading and there is some purging going on. Error relocating /sbin/udevadm 2019-04-30 14:35:23 I did rc-update del udev, now syslogd failed 2019-04-30 14:37:17 ok I believe with my level of knowledge I better use the stable repos 2019-04-30 14:37:43 <_ikke_> yes, that would be advisable 2019-04-30 14:38:05 <_ikke_> remove the edge repos, and then apk upgrade --update --available 2019-04-30 14:39:38 Thanks for your guidance 2019-04-30 14:43:47 _ikke_: One day I might live in a place where the people say your nickname colloquially 2019-04-30 14:45:07 Hello, I was wondering how you'd go about increasing the file descriptor limit for a daemon. 2019-04-30 15:13:38 _ikke_, tmhoang thanks for helping, stupid me was loading the module in a container which did not have the firmware file 2019-04-30 15:37:52 :D 2019-04-30 16:59:03 x86_64/wine package can only run 64bit windows apps? 2019-04-30 17:07:55 ncopa: how does apk handle files that existed in an old version of the package that don't exist in a newer version? 2019-04-30 17:13:21 <_ikke_> unless they are in /etc, they will probably be deleted 2019-04-30 17:14:26 _ikke_: does it check if the file changed and delete if they didn't? 2019-04-30 17:24:22 <_ikke_> afaik it doesn't for non-config files 2019-04-30 17:24:52 Yes yes I meant for stuff in /etc 2019-04-30 17:24:56 Sorry I wasn't clear enough 2019-04-30 17:25:49 <_ikke_> it will rename the file to $name.apksave 2019-04-30 17:26:58 is that new? 2019-04-30 17:27:36 i only know apk-new 2019-04-30 17:28:26 <_ikke_> sorry 2019-04-30 17:28:28 <_ikke_> that one yes 2019-04-30 17:28:31 <_ikke_> or 2019-04-30 17:28:33 <_ikke_> hmm 2019-04-30 17:28:46 <_ikke_> Maybe I confure arch 2019-04-30 17:29:13 arch definitely uses pacsave 2019-04-30 17:29:21 :) 2019-04-30 17:29:29 (and also pacnew) 2019-04-30 17:29:41 thats when you dont eat your own dogfood ;-) 2019-04-30 17:31:28 And .xbps-new I think 2019-04-30 19:15:06 Each time I upgrade the linux kernel, I get this 2019-04-30 19:15:07 https://imgur.com/a/1EFtMVF 2019-04-30 19:15:12 How can I fix that ? 2019-04-30 19:20:15 this is "normal" for LVM, nothing to worry about 2019-04-30 20:02:28 Thanks :)