2018-09-01 05:52:27 <vkrishn> does installing mongodb-tools still require pkg db? seems it does in al3.7.0
2018-09-01 05:53:04 <vkrishn> but does not show in depends in 'https://pkgs.alpinelinux.org/package/v3.7/community/x86/mongodb-tools'
2018-09-01 06:05:54 <vkrishn> ok, found it its dep of pkg libsasl
2018-09-02 01:46:09 <Bluefoxicy> I need to figure out how to make apk packages for Ansible AWX
2018-09-02 20:08:57 <terra1> Bluefoxicy: There's a quite clear Installation Guide: https://github.com/ansible/awx/blob/devel/INSTALL.md
2018-09-02 20:10:02 <Bluefoxicy> terra:  nod.
2018-09-02 20:10:31 <Bluefoxicy> terra: for some definition of "quite clear"
2018-09-02 20:11:06 <terra> Bluefoxicy: You have to attempt to a manual install for make sure all dependencies are ok
2018-09-02 20:12:14 <terra> After seeing the further, follow same steps in APKBUILD
2018-09-02 20:12:14 <Bluefoxicy> terra:  well yes.  Their install guide is essentially "get a docker container running through a docker clustering manager and have docker images built to run the software in docker"
2018-09-02 20:12:30 <Bluefoxicy> so
2018-09-02 20:12:45 <Bluefoxicy> let's say you want to run nginx in docker
2018-09-02 20:13:03 <Bluefoxicy> you could use the official nginx docker image, which builds nginx from its latest source inside an Alpine system
2018-09-02 20:13:27 <Bluefoxicy> OR you could create a Dockerfile that stars with 'FROM alpine:3.6' or whatnot, apk installs nginx.
2018-09-02 20:13:45 <Bluefoxicy> For that matter, you could ACTUALLY INSTALL ALPINE and install nginx without docker.
2018-09-02 20:14:19 <Bluefoxicy> So AWX runs on CentOS or RHEL or something
2018-09-02 20:14:25 <Bluefoxicy> the Docker image is a CentOS image
2018-09-02 20:14:32 <Bluefoxicy> now I want awx:alpine
2018-09-02 20:15:12 <Bluefoxicy> those "quite clear" build instructions are "sure, that's easy.  Install Alpine and bring up a CentOS container with AWX under Docker."
2018-09-02 20:18:25 <terra> From my understanding, you want pass all docker image steps because you already in Alpine?
2018-09-02 20:18:40 <Bluefoxicy> I want to actually build a package for awx
2018-09-02 20:18:55 <Bluefoxicy> I don't want to install Docker, pull the docker image, and run a container with awx in it.
2018-09-02 20:19:02 <Bluefoxicy> I want to be able to install awx natively
2018-09-02 20:19:26 <Bluefoxicy> Basically, their installation instructions are "we have a virtual appliance image you can run!"
2018-09-02 20:19:30 <terra> Did you check this: https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=awx
2018-09-02 20:19:40 <terra> looks like quite simple
2018-09-02 20:19:58 <Bluefoxicy> hmm
2018-09-02 20:20:36 <terra> This is git version: https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=awx-git
2018-09-02 20:22:05 <Bluefoxicy> setup.py isn't listed anywhere in https://github.com/ansible/awx/blob/devel/INSTALL.md
2018-09-02 20:23:05 <Bluefoxicy> but ok
2018-09-02 20:23:12 <Bluefoxicy> didn't know Arch had pkgbuild
2018-09-02 20:23:14 <terra> you're right.. they seem view the situation on user side.
2018-09-02 20:23:46 <Bluefoxicy> terra:  awx is the open-source version of a product Redhat sells, tied to Redhat services and designed to be a tie-in to support and operating system sales.
2018-09-02 20:24:00 <terra> Tower
2018-09-02 20:24:04 <Bluefoxicy> Yep
2018-09-02 20:24:42 <terra> It is actually otherwise, tower is a branded AWX
2018-09-02 20:24:48 <Bluefoxicy> Given my understanding of Redhat internal business strategies, I would be unsurprised if they intentionally geared all building/installation instructions to attaching the user to Redhat
2018-09-02 20:25:08 <Bluefoxicy> I thought Redhat bought Ansible and Tower, and then open-sourced Tower
2018-09-02 20:25:10 <terra> maybe
2018-09-02 20:25:46 <Bluefoxicy> in any case I'll have a look at this.
2018-09-02 20:26:23 <Bluefoxicy> ... at some point.  Right now I'm more interested in some .Net code I'm writing.  <_<
2018-09-02 20:26:32 <Bluefoxicy> I'm starting to wonder if I had enough fire for all these irons.
2018-09-02 20:27:47 <terra> https://github.com/ansible/awx/blob/devel/README.md -> "It is the upstream project for Tower"
2018-09-02 20:28:23 <Bluefoxicy> Well yes
2018-09-02 20:28:37 <Bluefoxicy> It started life as a closed product
2018-09-03 16:37:13 <wfa> is anyone willing to entertain some questions about Arm DTBs for beaglebone black/green?
2018-09-03 16:38:03 <wfa> I've got the 3.8 generic release almost working, but need to build the DTB for BBB/BBG
2018-09-03 16:39:00 <wfa> if I run my own kernel with my own DTB files that's 100% good, but then I lack the initrd with all the vanilla goodness
2018-09-03 17:57:17 <terra> wfa: did you try mkinitfs ?
2018-09-03 18:06:56 <wfa> I'm cross-compiling and don't have the full arm-specific stuff to do that
2018-09-03 18:07:20 <wfa> i.e. from my intel workstation trying to build an arm image
2018-09-03 18:09:57 <wfa> even if I had the mkinitfs for this Alpine rev, I suspect I will have to add the DTB for bbb/bbg to make it work on my hardware
2018-09-03 18:10:34 <wfa> ** File not found /boot/dtbs/am335x-boneblack.dtb **
2018-09-03 18:10:43 <wfa> ** File not found /boot/dtbs/am335x-bonegreen.dtb **
2018-09-03 18:10:55 <wfa> need to add these to the initrd for me to get a successful boot
2018-09-03 18:13:55 <wfa> sorry, not the initrd, just the tarball that has the kernel, initrd, apks, and dtbs
2018-09-03 18:35:08 <terra> wfa: you could extract all related u-boot files from other distro images like openelec, armbian etc.
2018-09-03 18:35:50 <terra> you just need to change partition UUID from parameters with yours
2018-09-03 18:37:44 <terra> I was using dd method to install u-boot and always worked.
2018-09-03 18:40:12 <terra> if you have egough experince, you just can replace any ARM with Alpine stage3 just by using same kernel and boot.
2018-09-03 18:40:32 <terra> *ARM distro
2018-09-03 18:42:03 <wfa> that might be what I have to do.  I had a different kernel rev than the 3.8
2018-09-03 18:42:54 <wfa> I also can't use the uboot that comes with the distro - built my own from source, substituted the MLO and u-boot.img at the correct dd seek offsets
2018-09-03 18:43:39 <wfa> I guess if I use the same kernel as 3.8 then the Alpine initrd will pick up from there
2018-09-03 18:44:35 <wfa> but I'm very new to Alpine and still wrapping my head around the run-from-ram concepts
2018-09-03 18:47:41 <terra> you can use "overlaytmpfs" boot parameter for easily succeed run-from-ram
2018-09-03 18:48:45 <wfa> :) not smart enough for that yet.  uboot is impenetrable to me.  I can pull off uname_r=mysomething and that's about all I can handle
2018-09-03 18:49:51 <terra> oh I forgot, overlaytmpfs is only makes sense if you use mikitfs generated initramfs, sorry.
2018-09-03 18:51:19 <wfa> I suspect I'll get there eventually.  At this point I'm nearing a full-native built-from source toolchain.  Not what I had hoped, but perhaps I'm better off for it
2018-09-03 18:53:05 <wfa> and thank you for the advice.  I'll try to match the kernel rev
2018-09-03 18:54:36 <wfa> it's late here so I likely won't be interactive, but thanks in advance for any and all advice from people that are running alpine on beaglebone bare metal
2018-09-04 10:56:51 <ncopa> I just want to say a big thank you for everyone helping review and comment on the PRs
2018-09-04 10:56:59 <ncopa> you have no idea how much that means to me :)
2018-09-04 11:02:59 <bernhardgruen> ncopa, you are doing a great job too. It is hard to express that with words.
2018-09-04 11:03:52 <bernhardgruen> it is nice to see that the number of pull requests is down from over 280 to 210 at the moment.
2018-09-04 11:39:22 <ncopa> bernhardgruen: i think the PR queue is going down due to prspkt has stopped to submit new PRs :)
2018-09-04 11:43:38 <unmy> ncopa, thank you ncopa for alpine! and working grub single partition full disk encryption without typing passphrase twice, I will have to find the moment and create some wiki quick how to :P
2018-09-04 11:55:02 <ncopa> so you can do single partision full disk encryption without typing passphrase twice?
2018-09-04 11:56:50 <unmy> ncopa, yep, working great
2018-09-04 11:57:13 <unmy> LVM on LUKS ofc
2018-09-04 11:59:09 <ncopa> cool!
2018-09-04 12:00:23 <unmy> just need add to GRUB_CMDLINE_LINUX "cryptkey=yes" which by default point to /crypto_keyfile.bin + GRUB_ENABLE_CRYPTODISK=y and in mkinitfs.conf add "cryptkey"
2018-09-04 12:01:35 <ncopa> so you sotre the disk excryption key in crypt_keyfile.bin and protect this with a password which you are prompted for
2018-09-04 12:01:43 <ncopa> nice
2018-09-04 12:03:04 <unmy> more like there is no need to create separated boot partition at least in BIOS mode
2018-09-04 12:06:01 <unmy> the key was add "cryptkey=yes" or own "cryptkey=/path/tofile" but if wanna different path then need change it or create own "features" in mkinitfs features.d
2018-09-04 12:07:14 <unmy> but ye, all working like on other linux distributions, good job!
2018-09-04 12:10:09 <unmy> and have to remember to limit access to root only that crypt_keyfile.bin which isnt protected and to initramfs too so normal user wont be abble to get key
2018-09-04 12:15:11 <TBB> I am confused by this
2018-09-04 12:16:54 <unmy> whole / is already encrypted so without password to grub you wont be abble run initramfs, and to avoid type password second type can use key file and store it in initramfs to mount whole root "/"
2018-09-04 12:17:05 <TBB> the tiny bit of code that is grub opens an encrypted root partition to get to load the kernel and the initramfs from it, which then does the exact same open using a key file?
2018-09-04 12:17:25 <unmy> ye
2018-09-04 12:17:29 <TBB> okay
2018-09-04 12:17:42 <TBB> so I wasn't quite as confused as I first thought
2018-09-04 12:18:06 <unmy> is kind of weird, true
2018-09-04 12:19:24 <unmy> or maybe not so much if somebody using different layout of partitions or without lvm or separated /boot
2018-09-04 12:19:50 <TBB> it makes perfect sense. I have this little hobby project I haven't had a chance to advance lately that has a similar aim, but I'm trying to keep the key out of the filesystem entirely
2018-09-04 12:21:58 <TBB> on TPM or smartcard or such
2018-09-04 12:22:33 <TBB> so I could keep the entire disk encrypted, optimally even without any LUKS header on-disk
2018-09-04 12:22:47 <unmy> TBB, ye agree but if somebody already have any access to running machine even as a user then can't do much :\
2018-09-04 12:23:30 <unmy> safe machine is turned off machine :D
2018-09-04 12:24:37 <TBB> yup. it can be quite a depressing task really, if you start considering all possible attack vectors
2018-09-04 12:25:18 <TBB> and even more depressing when you realize that even if you figure out the perfect technical solution, it can still be defeated with a $5 crypto hammer ...
2018-09-04 12:25:23 <unmy> ye for me depressing was to have unencrypted /boot :\
2018-09-04 12:25:40 <AinNero> TBB: alpine diskless
2018-09-04 12:25:54 <AinNero> crypto hammer doesn't work if there is no persistent data to be decrypted
2018-09-04 12:26:16 <TBB> AinNero: that's close to what I'm planning, but unfortunately some of my data needs to persist :)
2018-09-04 12:27:15 <TBB> unmy, it's not really that bad to have /boot unencrypted, you can apply TPM and SecureBoot (on "recent" hardware) to protect it (although it's by no means a trivial task)
2018-09-04 12:29:04 <unmy> there is many approaches to do things, I'm glad that in linux world we still can decide
2018-09-04 12:30:01 <unmy> but ye, time ago I was trying set up alpine to use keyfile and I failed so now I'm happy that could do it
2018-09-04 12:33:48 <unmy> and it was probably before "lemmarathon authored and ncopa committed on May 1" "Add support for keyfiles" in mkinitfs so now I can confirm - it works :)
2018-09-04 12:35:17 <TBB> I made a similar feature for smartcards, but the problem with that approach is you also need a bit of a custom initramfs-init
2018-09-04 12:36:00 <AinNero> usecase for initramfs hooks!
2018-09-04 12:36:21 <TBB> so basically it would be useful to be able to use script snippets you could then introduce to the initramfs via a feature
2018-09-04 12:36:28 <TBB> AinNero: yup
2018-09-04 13:50:49 <ncopa> whee \o/ we are below 200 PRs \o/
2018-09-04 13:51:06 <_ikke_> \0/
2018-09-04 13:51:09 <_ikke_> \o/
2018-09-04 13:52:16 <_ikke_> https://imgur.com/a/s6atg1t
2018-09-04 13:55:01 <clandmeter> \o/\o/\o/
2018-09-04 21:35:25 <sleepycat> is the openrc package something I should depend on if I have a whatever.initd file in my package?
2018-09-04 21:35:55 <_ikke_> sleepycat: no
2018-09-04 21:36:15 <_ikke_> sleepycat: nowadays, those files go into an -openrc subpackage, and it's the other way around
2018-09-04 21:36:21 <_ikke_> it gets installed if openrc is installed
2018-09-04 21:50:14 <sleepycat> 🤔
2018-09-04 21:50:20 <sleepycat> Interesting.
2018-09-04 21:51:24 <sleepycat> I guess I can just do `/etc/init.d/whatever start` and don't really need openrc... is that right?
2018-09-04 21:55:04 <sleepycat> Also does anyone know what's going on with the node-current package? It's not very current.
2018-09-04 21:55:40 <sleepycat> 10.9.0 is the current version.
2018-09-04 21:58:18 <sleepycat> Hmm, after compiling, abuild spits out `fatal: No names found, cannot describe anything.` What names is it looking for?
2018-09-04 22:02:06 <awilfox> sleepycat: git tag
2018-09-04 22:54:17 <sleepycat> awilfox: Ahh, ok. Thanks!
2018-09-05 15:25:51 <Ganwell> I started creating a qemu-system-* VM for (maybe) all the archs alpine linux has. I found out that it can be quite generic if you use netboot, for example no special uboot hacks on arm. I hardcoded most of the qemu networking, because I didn't bother using libvirt. Provisioning the VM is not finished, and at the moment only x86* and arm works. Any comments?
2018-09-05 15:25:53 <Ganwell> https://github.com/ganwell/alpine-multiarch-build/blob/master/boot
2018-09-05 15:27:17 <Ganwell> I saw http://boot.alpinelinux.org/. Is it going to be something similar?
2018-09-05 18:19:57 <ncopa> nice
2018-09-05 18:20:09 <ncopa> its clandmeter that has worked most on the netboot thing
2018-09-05 18:54:47 <clandmeter> Ganwell, not sure i understand your question
2018-09-05 19:13:00 <Ganwell> clandmeter: Ah, I just wanted to show what I am working on. But I actually have a question: How would you typically provision the booted VM? What I am doing is start qemu with -nographic and then run commands via stdin using pexpect. Is there a better way to execute something automatically once the netboot has booted?
2018-09-05 19:20:15 <clandmeter> Ganwell, you are talking about netboot as its only usable via qemu. thats why i didnt understand your question.
2018-09-05 19:21:22 <clandmeter> Ganwell, you can provide it an apkovl and add some initial config to it.
2018-09-05 19:26:34 <Ganwell> clandmeter: Ah, of course netboot works on bare-metal, too. I used docker and qemu-user-static to build and test on other archs, but there are problems with that approach, so I wanted to switch to full VMs. I found a blog post: https://blog.w1r3.net/2018/04/16/automated-provisioning-using-apkovl.html nice!!
2018-09-05 19:27:58 <clandmeter> yes it would be nice if we have some script on online service to provide ovl generation.
2018-09-05 19:29:24 <AinNero> Ganwell: how did you find it?
2018-09-05 19:30:05 <Ganwell> clandmeter: append apkolv, nice. Thanks a lot. Is adding more archs to http://boot.alpinelinux.org easy? Maybe I should help you, instead of doing my own thing.
2018-09-05 19:30:28 <ncopa> what do you think about a contributor checklist when users are creating PRs? : https://github.com/alpinelinux/aports/pull/4862
2018-09-05 19:31:04 <ncopa> i think we need something to point them into right direction
2018-09-05 19:31:19 <ncopa> but i wonder if a checklist is the way to go
2018-09-05 19:31:51 <AinNero> im an avation nerd, and i really appreciate checklists
2018-09-05 19:31:59 <Ganwell> AinNero: Once I knew I want apkovl is googled: apkolv provisioning
2018-09-05 19:32:08 <Ganwell> I
2018-09-05 19:32:12 <Ganwell> googled
2018-09-05 19:32:14 <AinNero> AinNero: funny, because thats my blog
2018-09-05 19:32:22 <Ganwell> I know
2018-09-05 19:32:23 <AinNero> *Ganwell
2018-09-05 19:35:57 <Ganwell> ncopa: Sometimes I find the wrong information, either copying from a bad APKBUILD or reading old wiki entries. So something that is definitely up to date would be nice.
2018-09-05 19:37:02 <Ganwell> ncopa: abuild already gives nice warnings, maybe this could be extended, instead of a checklist.
2018-09-05 19:37:10 <clandmeter> Ganwell, netboot.a.o is based on ipxe
2018-09-05 19:37:21 <clandmeter> so its limited to what ipxe supports
2018-09-05 19:39:35 <Ganwell> ncopa: abuild of course can't check git related mistakes.
2018-09-05 19:55:08 <ncopa> right. so, step 1 would be to have a single source of truth. updated official documentation.
2018-09-05 19:55:18 <ncopa> something that tells how things should be
2018-09-05 19:55:43 <ncopa> step 2 would be to have a linter, that checks the APKBUILD for common things
2018-09-05 19:55:57 <ncopa> we already have some thigns in abuild, but i think it could be extended
2018-09-05 19:56:08 <ncopa> iirc nmeum has started to write a linter
2018-09-05 21:07:24 <nmeum> I did indeed start writing a linter. However, I got demotivated after a while and stopped working on it
2018-09-05 21:07:46 <danieli> I know the feeling of getting demotivated very well
2018-09-05 21:08:21 <nmeum> the code can be found at https://github.com/nmeum/abuild-lint just in case someone wants to take a look at it
2018-09-05 21:44:11 <Ganwell> nmeum: I'll check it out. I started listing common mistakes (independent of how they can be detected). If some core-developers would add mistakes, it would probably get very long. https://wiki.alpinelinux.org/wiki/User:Ganwell/mistakes
2018-09-05 22:49:58 <Ganwell> hmm, git aware tool could also check for missing pkgver bumps, no pkgrel reset, apk that don't start in testing... interesting.
2018-09-05 22:54:04 <ayee> fetch http://dl-cdn.alpinelinux.org/alpine/v3.8/main/x86_64/APKINDEX.tar.gz ERROR: http://dl-cdn.alpinelinux.org/alpine/v3.8/main: temporary error (try again later)
2018-09-05 22:54:15 <ayee> ^my docker build which usually works is failing. Is this related to the new 3.8 release?
2018-09-05 22:55:14 <ayee> I'm just doing an apk update, and apk add. using "FROM python:2-alpine"
2018-09-05 22:56:19 <ayee> It worked a couple weeks ago. *shrugs*
2018-09-05 22:56:36 <awilfox> "temporary error (try again later)" -> Check DNS
2018-09-05 22:57:48 <ayee> dl-cdn.alpinelinux.org is an alias for global.prod.fastly.net. global.prod.fastly.net has address 151.101.40.249
2018-09-05 22:58:46 <ayee> o you're right, it is dns. inside the docker container though.
2018-09-05 22:59:03 <ayee> sudo docker run busybox nslookup google.com; ;; connection timed out; no servers could be reached
2018-09-05 23:11:58 <ayee> thanks!
2018-09-06 10:31:36 <_ikke_> Show should I handle a go project that does a bunch of go gets + 1 git clone in a Makefile?
2018-09-06 11:37:57 <ncopa> _ikke_: good question. I dont have any good answer...
2018-09-06 11:48:46 <clandmeter> Probably break down the makefile
2018-09-06 11:49:41 <clandmeter> If it's related to gitlab that's what I did
2018-09-06 11:53:29 <_ikke_> clandmeter: it is :)
2018-09-06 11:53:35 <_ikke_> I also found jirutka's aports
2018-09-07 13:38:44 <PureTryOut[m]> could anybody review https://github.com/alpinelinux/aports/pull/4479 for me?
2018-09-07 19:55:09 <_ikke_> Anyone knows what provides sys/sysctl.h?
2018-09-07 20:05:55 <_ikke_> glibc apparently :(
2018-09-07 20:07:50 <_ikke_> heh: https://github.com/libressl-portable/portable/issues/96
2018-09-08 21:08:22 <AinNero> have you witnessed the dns issues in the other channel?
2018-09-08 21:08:54 <AinNero> musl 1.1.20 broke something with DNS
2018-09-08 21:09:29 <awilfox> nope
2018-09-08 21:09:42 <awilfox> this is one of 4 channels I'm joined to on freenode
2018-09-08 21:10:01 <AinNero> anyways, i also noticed that nslookup inside of docker containers is slow af with musl
2018-09-08 21:10:32 <awilfox> haven't had any DNS issues with musl
2018-09-08 21:10:51 <awilfox> but Adélie is pretty much vanilla musl, we only have 2 patches
2018-09-08 21:45:43 <AinNero> 
ACTION is angry

2018-09-08 21:45:54 <AinNero> turns out the DNS issue i had is due to an broken dns server
2018-09-08 21:46:08 <AinNero> who sends unsolicited ServFail responses
2018-09-08 21:46:18 <AinNero> well known issue and rich already files issues for it
2018-09-09 02:28:24 <linux_dr> I'm not sure if this is an Alpine or docker question, but I gotta start somewhere... I've got a node webserver running on alpine inside docker, and I'm getting a ENOTFOUND when I try to listen on a port with host IP 0.0.0.0, (which, as I understand, means all interfaces.) Any chance this is a musl issue?
2018-09-09 02:29:10 <awilfox> extremely unlikely bordering on impossible
2018-09-09 02:29:34 <awilfox> ENOTFOUND isn't even something musl will return on bind nor listen afaict
2018-09-09 02:30:11 <linux_dr> awilfox, I think it's being returned from DNS
2018-09-09 02:30:28 <linux_dr> *WHY* 0.0.0.0 is going to DNS is beyond me.
2018-09-09 02:31:13 <linux_dr> In node, it's happening in getaddrinfo()
2018-09-09 02:31:40 <awilfox> "The name does not resolve"
2018-09-09 02:31:44 <awilfox> yes, it's going to DNS :)
2018-09-09 02:32:55 <linux_dr> Yes... I got that much... I don't get WHY it's trying to resolve a numeric IP
2018-09-09 02:33:48 <linux_dr> awilfox, (I do appreciate your help, BTW)
2018-09-09 02:34:18 <linux_dr> so far google's been playing hard to get on this one with me.
2018-09-09 02:35:05 <linux_dr> does musl have a different way of representing "all interfaces" than 0.0.0.0?
2018-09-09 02:35:13 <awilfox> no
2018-09-09 02:35:16 <awilfox> that's a Linux kernel thing
2018-09-09 02:35:21 <awilfox> (and in general, posix)
2018-09-09 02:35:36 <linux_dr> ... if I can get it there...
2018-09-09 02:35:59 <linux_dr> should I add a line to /etc/hosts "0.0.0.0     0.0.0.0"?
2018-09-09 02:36:02 <awilfox> no.
2018-09-09 02:36:07 <awilfox> that will confuse things
2018-09-09 02:36:58 <linux_dr> by things, I think you mean *ME* ... RotFL... You are being very patient, thank you.
2018-09-09 02:39:18 <linux_dr> 
ACTION is reviewing https://github.com/idunham/musl/blob/master/src/network/getaddrinfo.c

2018-09-09 02:39:35 <awilfox> this is very old
2018-09-09 02:39:53 <awilfox> the canonical source of musl is at http://git.musl-libc.org/cgit/musl/tree/src
2018-09-09 02:40:09 <awilfox> so you likely want https://git.musl-libc.org/cgit/musl/tree/src/network/getaddrinfo.c
2018-09-09 02:40:39 <linux_dr> great... thanks..
2018-09-09 02:40:56 <linux_dr> (Like I said, google's been giving me no love on this.)
2018-09-09 02:45:47 <linux_dr> do you expect getaddrinfo() is giving the wrong result, or that it shouldn't be called for this situation?
2018-09-09 02:47:03 <linux_dr> (by wrong, I mean not what node is expecting, not a bug per se)
2018-09-09 02:47:44 <awilfox> I have no idea how well tested node is on alpine
2018-09-09 02:48:37 <linux_dr> This is all pretty stock docker stuff... didn't roll anything myself here... I think I'm configuring something wrong.
2018-09-09 02:51:13 <linux_dr> This is just the built-in nodeJS web server ("expressJS") running docker's official "node:alpine" container
2018-09-09 02:54:01 <awilfox> I can't offer much enlightenment here because I don't use docker at all (have literally never even been to the website, let alone used it)
2018-09-09 02:54:20 <awilfox> nor do I use nodejs (I'm actually an Adélie developer, not an Alpine developer, and we don't ship Node because it requires python2 to build still.)
2018-09-09 02:57:15 <linux_dr> I'm pretty new to docker, so I can understand the sentiment.. I was perfectly happy with vagrant... that's a whole different animal.
2018-09-09 02:57:25 <awilfox> vagrant!  that's something I can do
2018-09-09 03:00:48 <linux_dr> Yeh... I hear you...
2018-09-09 03:01:01 <linux_dr> How did you get involved in Adélie?
2018-09-09 03:01:30 <linux_dr> is it your baby? or did someone pass you the torch?
2018-09-09 03:04:45 <awilfox> yes, my baby.
2018-09-09 03:04:50 <linux_dr> Also, while on the subject, how is Adélie on size? I saw it uses musl, but doesn't seem to have the focus on size and install speed of Alpine. (Standards compliance is nothing to sneeze at. Just curious from a pragmatic POV.)
2018-09-09 03:05:17 <awilfox> https://www.adelielinux.org/info.html#technicalqa the whole messy truth
2018-09-09 03:05:45 <awilfox> linux_dr: depends on your measure
2018-09-09 03:06:03 <awilfox> there is not quite a focus on size; a fully installed system is ~95 MiB
2018-09-09 03:06:16 <awilfox> compared to alpine's almost ridiculous 40 MiB
2018-09-09 03:06:26 <awilfox> but when you boot it up, about the same RAM is used
2018-09-09 03:06:59 <awilfox> adelie focuses a LOT more on RAM vs disk; did a comparison of my firefox profile with 36 tabs open (and loaded, not just sitting there; every single one was clicked on in order)
2018-09-09 03:07:15 <awilfox> kubuntu: 1.8 GB; gentoo: 1.2 GB; alpine: 1.1 GB; adelie: 850 MB
2018-09-09 03:07:32 <awilfox> that's firefox private_dirty
2018-09-09 03:11:40 <linux_dr> I doubt anyone running Alpine in a container has ever seen it running a GUI...lol
2018-09-09 03:12:43 <linux_dr> Where do you hail from? Is Adélie your day job? Do you have financial backing?
2018-09-09 03:12:58 <awilfox> https://i.imgur.com/XNC2k7m.png
2018-09-09 03:13:03 <linux_dr> (mostly out of curiosity)
2018-09-09 03:13:07 <awilfox> this is a screenshot I just took of my own desktop
2018-09-09 03:13:11 <awilfox> a ppc64 running Adélie
2018-09-09 03:13:37 <awilfox> Oklahoma; somewhat; somewhat.
2018-09-09 03:14:23 <danieli> gotta admit, i'm on arch right now
2018-09-09 03:14:34 <danieli> still some things i haven't had time to port to work with musl
2018-09-09 03:14:42 <awilfox> danieli: list them
2018-09-09 03:14:50 <awilfox> danieli: beta2 is our package freeze, there's still time :P
2018-09-09 03:15:11 <danieli> i'm planning to jump over to alpine soon enough™
2018-09-09 03:15:24 <awilfox> we just ported all of kde, then all the /accessories/ of xfce4, then all of elementaryOS' desktop thingy, then the new libreoffice-gtk3
2018-09-09 03:15:25 <danieli> for the most part, security tools are things i need
2018-09-09 03:15:44 <awilfox> oh and I got to port thunderbird personally
2018-09-09 03:15:52 <awilfox> *that* was some fun, even the mozirc people didn't know how to help
2018-09-09 03:16:04 <danieli> that sounds like fun
2018-09-09 03:16:07 <danieli> is the codebase any better now?
2018-09-09 03:16:18 <awilfox> the fun part was writing my first email with User-agent: Thunderbird (Linux ppc64)
2018-09-09 03:16:34 <awilfox> danieli: it's still mozilla-central
2018-09-09 03:16:52 <awilfox> I'd argue the actual logic is better.  but gfx/ and ui/ are much, much worse :s
2018-09-09 03:17:14 <danieli> the crappy chromium/CEF/electron support is also an issue - I wish it wasn't htough
2018-09-09 03:17:34 <awilfox> chromium is being ported to ppc64 (unfortunately)
2018-09-09 03:17:42 <awilfox> ... wait, lol, I thought I was in #talos
2018-09-09 03:17:43 <awilfox> not #alpine
2018-09-09 03:17:46 <awilfox> ppc64 doesn't help you
2018-09-09 03:17:59 <danieli> daniel@daniel-pc [~/projects/alpine/aports/community/chromium]$ ls *.patch | wc -l
2018-09-09 03:18:02 <danieli> 24
2018-09-09 03:18:08 <awilfox> still not as bad as busybox
2018-09-09 03:18:10 <awilfox> :)
2018-09-09 03:18:41 <danieli> 20 for busybox
2018-09-09 03:18:44 <danieli> probably bigger patches?
2018-09-09 03:18:51 <awilfox> ah yeah
2018-09-09 03:19:06 <danieli> 1962 lines of patch files in busybox, 2248 in chromium
2018-09-09 03:19:12 <awilfox> danieli: assuming you are on x86_64
2018-09-09 03:19:17 <danieli> well yeah
2018-09-09 03:19:23 <awilfox> danieli: gcompat might help you with electron
2018-09-09 03:19:55 <danieli> for 5 years, i've been working on/off on a private chromium fork (chromium-hardened)
2018-09-09 03:20:19 <awilfox> honestly, it is kind of... broken on purpose... I don't mean I like wrote bad code or anything, but I purposefully didn't implement all the APIs electron wants, because I was hoping to make a dual-purpose statement: adelie doesn't need electron, and people should write native code for their apps
2018-09-09 03:20:22 <danieli> ripped out the google crap and did an audit of /most/ things, with lots of features disabled - especially in v8
2018-09-09 03:20:34 <danieli> heh
2018-09-09 03:20:35 <danieli> *should*
2018-09-09 03:20:36 <awilfox> but you could likely easily add the necessary APIs
2018-09-09 03:20:37 <danieli> there's the issue
2018-09-09 03:20:41 <danieli> they should, but don't
2018-09-09 03:20:48 <awilfox> it's silly floating point stuff
2018-09-09 03:20:53 <awilfox> that would be a right pain to do correctly
2018-09-09 03:21:04 <awilfox> but likely easy enough to do hax enough to make electron start up
2018-09-09 03:21:10 <awilfox> (our goal is always doing things /right/, hence meh)
2018-09-09 03:21:19 <danieli> start up, sure, we have both chromium and node working
2018-09-09 03:21:21 <danieli> but to work well is another issue
2018-09-09 03:21:29 <danieli> [ hax.ogg ]
2018-09-09 03:23:19 <linux_dr> awilfox, looks like a great project you have going there. You mentioned 95MB? was that a barebones install? (For some reason I thought I remembered Alpine being < 10MB for absolute barebones... Not that you want to go there, but for apples-to-apples comparison.)
2018-09-09 03:23:34 <awilfox> linux_dr: that was adelie-base
2018-09-09 03:23:41 <danieli> the alpine rootfs is ~5 MB
2018-09-09 03:23:42 <awilfox> linux_dr: adelie-core is closer to 50 MB but I don't think that even has an init system
2018-09-09 03:23:46 <danieli> (the one you use in docker)
2018-09-09 03:23:50 <awilfox> adelie-core is for like docker
2018-09-09 03:24:09 <danieli> yeah, it's meant to be the absolute barebones
2018-09-09 03:24:11 <danieli> no init, nada
2018-09-09 03:24:21 <danieli> lxc however does better with an init system
2018-09-09 03:24:40 <awilfox> linux_dr: adelie-base-posix is around 115 MB and adds stuff like vim (which is needed for posix ex; funny enough, nvi's ex is not posix compliant, even though it shipped in 4.4BSD and some SysVs)
2018-09-09 03:24:55 <linux_dr> awilfox, a week or so ago, I discovered you don't even need pretty much ANY OS inside the container, if your binaries are self-contained.
2018-09-09 03:25:20 <awilfox> linux_dr: that's correct
2018-09-09 03:25:29 <awilfox> though you do need a /bin/sh to be strictly conformant; otherwise system(3) will fail
2018-09-09 03:25:40 <awilfox> we ship dash for that, 600kb
2018-09-09 03:25:47 <awilfox> you can also use zsh, 4 MB, or bash, 7 MB
2018-09-09 03:26:11 <danieli> I am currently creating rust-nightly for a friend... bleh
2018-09-09 03:26:13 <linux_dr> awilfox, sorry.. context... within docker you don't need even alpine... you can have a 100% empty filesystem, and throw your binaries in there...
2018-09-09 03:26:46 <danieli> you /could/ but that wouldn't be the best of ideas
2018-09-09 03:26:47 <awilfox> linux_dr: right that's what I meant
2018-09-09 03:26:56 <awilfox> linux_dr: you still need a /bin/sh or the system(3) C library call will fail
2018-09-09 03:27:08 <linux_dr> see, you know more about docker than you think you do.. lol
2018-09-09 03:27:10 <awilfox> if you have a static linked dash, you could still get by < 1 MB though
2018-09-09 03:27:21 <danieli> *some* common sense still applies :)
2018-09-09 03:27:23 <awilfox> lol no I just know a lot about musl
2018-09-09 03:27:38 <awilfox> managed to get a lot of patches pushed through into 1.1.20
2018-09-09 03:27:43 <linux_dr> awilfox, why would you call system() if there's no OS out there?
2018-09-09 03:27:57 <awilfox> linux_dr: who knows.  lots of libraries do for weird reasons
2018-09-09 03:28:30 <awilfox> qt calls system() on signals (abrt/ill, not segv) to "pretty dump" core
2018-09-09 03:29:28 <linux_dr> awilfox, Understood... I'm not the language's biggest fan, but been doing a lot recently with go, and it is odd/refreshing/curious/etc. that it's binaries (typically) are 100% static with no libc dependancy even.
2018-09-09 03:29:52 <awilfox> yeah
2018-09-09 03:31:10 <linux_dr> I'm a bit too addicted to inheritance to be 100% on the go bandwagon yet... also really think it needs a few things, like templates and sane error handling still.
2018-09-09 03:31:55 <danieli> to rehash anunpopular opinion, it kind of feels like C++'s little brother, with some features the creators wanted thrown onto it
2018-09-09 03:32:00 <danieli> the error handling also isn't "quite there"
2018-09-09 03:33:34 <linux_dr> danieli, To me it feels more like what they LEFT OUT more then added.
2018-09-09 03:34:19 <danieli> C++ is huge
2018-09-09 03:34:21 <linux_dr> Yeh... the error handling feels like "watch where you step"
2018-09-09 03:34:28 <danieli> http://shitcode.net/112
2018-09-09 03:35:52 <linux_dr> C++ has plenty of warts... a lot I like about it... but I'm not sure I've found any language I'm 100% happy with.. but then again, there are several still on my bucket-list.
2018-09-09 03:37:24 <danieli> wonder if i should create a rust-nightly package in testing?
2018-09-09 03:37:43 <danieli> i just ported the patches and it is currently building on a scaleway dedi
2018-09-09 03:37:59 <linux_dr> I've heard good things about rust, but haven't gotten to it yet.
2018-09-09 03:38:11 <danieli> "rust evangelism strike force"
2018-09-09 03:38:19 <danieli> that's why you've heard good things :P
2018-09-09 03:38:28 <danieli> i'm doing this for a friend though, i don't use it much myself
2018-09-09 03:38:51 <linux_dr> awilfox, do you have X running in Adélie?
2018-09-09 03:39:08 <danieli> they have a GUI running so I would assume so, with strong certainty
2018-09-09 03:39:17 <awilfox> linux_dr: did you miss the screenshot?  hehe
2018-09-09 03:39:28 <awilfox> I linked one to you above
2018-09-09 03:39:47 <awilfox> linux_dr: https://i.imgur.com/XNC2k7m.png
2018-09-09 03:40:00 <linux_dr> Saw the screenshot... didn't realize your parent host was Adelie as well...
2018-09-09 03:40:07 <awilfox> oh, yes
2018-09-09 03:40:07 <linux_dr> go you! :)
2018-09-09 03:40:07 <danieli> i just use the default TTY login -> startx if I am "daniel" and on TTY1 -> xinit -> i3
2018-09-09 03:40:09 <awilfox> it is
2018-09-09 03:41:12 <linux_dr> awilfox, any chance Blender compiles in Adélie?
2018-09-09 03:41:36 <danieli> https://pkgs.alpinelinux.org/package/edge/testing/x86_64/blender
2018-09-09 03:41:49 <danieli> should work in adelié too?
2018-09-09 03:42:04 <awilfox> would assume so
2018-09-09 03:42:09 <danieli> only requires a single and fairly small musl patch
2018-09-09 03:42:42 <linux_dr> danieli, has it been submitted upstream?
2018-09-09 03:42:48 <awilfox> ah hm
2018-09-09 03:42:51 <awilfox> ilmbase
2018-09-09 03:42:53 <awilfox> no
2018-09-09 03:42:55 <danieli> not sure, I don't have any insight there
2018-09-09 03:42:57 <danieli> I don't think so
2018-09-09 03:42:59 <linux_dr> is it still compatible (patched) with glibc?
2018-09-09 03:43:12 <awilfox> we don't have ilmbase.  it is broken on all non-x86_64 CPUs
2018-09-09 03:43:22 <danieli> should be
2018-09-09 03:43:34 <danieli> it just adds some clauses to if statements
2018-09-09 03:43:42 <danieli> adds `defined(HAVE_EXECINFO_H))` in a couple spots
2018-09-09 03:43:55 <linux_dr> danieli, You don't happen to have much docker experience... do you?
2018-09-09 03:43:59 <danieli> just awareness if a compability layer is present I suppose
2018-09-09 03:44:07 <danieli> linux_dr: honestly, yes, but I stay slightly off that scene
2018-09-09 03:44:16 <danieli> ncopa works at docker though, so he's your best bet
2018-09-09 03:44:20 <danieli> but he's usually swamped in work
2018-09-09 03:45:06 <linux_dr> Just trying to get something working for work.. google hasn't been much help so far.... awilfox did what she could but I'm out of ideas.
2018-09-09 03:45:19 <danieli> shoot
2018-09-09 03:45:30 <awilfox> yeah I think something may be up in node or docker
2018-09-09 03:45:32 <awilfox> but not sure
2018-09-09 03:45:40 <danieli> after staring into the void for a bit, I could help out
2018-09-09 03:45:50 <danieli> okay, what did you say the issue is?
2018-09-09 03:46:12 <linux_dr> I'm trying to run nodeJS on Alpine (node:alpine official image).. with node running an expressJS webserver...
2018-09-09 03:47:28 <linux_dr> trying to listen on "all network interfaces", denoted by 0.0.0.0, and instead of sending that to the kernel as a numeric IP, it gets sent to DNS and musl shoots back ENOTFOUND.
2018-09-09 03:47:48 <linux_dr> So I figure I'm doing something silly wrong.
2018-09-09 03:48:12 <danieli> "it gets sent to DNS"?
2018-09-09 03:48:13 <danieli> that's really odd
2018-09-09 03:48:18 <danieli> it tries to resolve 0.0.0.0?
2018-09-09 03:48:53 <danieli> what is the exact line of code you use to make your application listen?
2018-09-09 03:48:58 <danieli> http://tpaste.us/paste
2018-09-09 03:49:02 <linux_dr> specifically it's getting passed to getaddrinfo()
2018-09-09 03:49:09 <danieli> I'd like the nodejs line of code
2018-09-09 03:49:44 <linux_dr> Give me a minute... Can't cut n paste... it's on a different machine.
2018-09-09 03:50:11 <danieli> you could do `app.listen(8080);` or `app.listen(8080, "0.0.0.0");`
2018-09-09 03:50:35 <danieli> the first one implicitly listens on all interfaces
2018-09-09 03:53:02 <linux_dr> const http = require('http'); .... const HTTP_ADDR = process.env.HTTP_ADDR || '0.0.0.0'; .... const server = http.createServer(app); server.listen(HTTP_PORT, HTTP_ADDR, () => { console.log("stuff"); });
2018-09-09 03:53:25 <linux_dr> danieli, the second.. what's the difference?
2018-09-09 03:54:14 <danieli> they effectively do the same thing
2018-09-09 03:54:21 <linux_dr> I could weed out "0.0.0.0" as a special case, and ommit it.
2018-09-09 03:54:50 <linux_dr> danieli, any idea why getaddrinfo() gets called on it?
2018-09-09 03:55:06 <danieli> what does `server.listen` actually get as arguments?
2018-09-09 03:55:20 <danieli> does it get 80 and "0.0.0.0" for port and host respectively?
2018-09-09 03:56:04 <linux_dr> const express = require('express)const app = express();
2018-09-09 03:56:15 <linux_dr> const express = require('express'); ... const app = express();
2018-09-09 03:57:30 <danieli> try this, see if it runs at all https://hastebin.com/raw/newacezufa
2018-09-09 03:57:46 <linux_dr> danieli, not using 80.. that requires root...
2018-09-09 03:57:50 <danieli> also, do you have `127.0.0.1       hostname` in /etc/hosts?
2018-09-09 03:57:54 <danieli> well, try with >1000
2018-09-09 03:58:00 <danieli> oh
2018-09-09 03:58:03 <danieli> I had a typo there, my bad
2018-09-09 03:58:18 <danieli> https://hastebin.com/raw/qemegaceko
2018-09-09 03:58:21 <danieli> corrected
2018-09-09 03:58:30 <danieli> I was a bit distracted
2018-09-09 03:58:31 <linux_dr> not sure where the port number came from but it's 28585
2018-09-09 03:58:56 <danieli> if HTTP_PORT is not supplied, I believe it would use a random port
2018-09-09 03:59:22 <danieli> anyway, try the code in that hastebin link, see if it runs for a few second without error
2018-09-09 04:00:18 <linux_dr> danieli, no... it's specified in the docker-compose.yml
2018-09-09 04:01:29 <danieli> set up an identical docker container by hand (docker run --rm ... alpine:3.8 sh), `apk add nodejs-current`, npm install express, and run it
2018-09-09 04:01:35 <linux_dr> module.parent? my Javascript must be rustier than I thought.
2018-09-09 04:02:45 <danieli> err, that's just something I do
2018-09-09 04:02:53 <danieli> it only runs that block of code if it's ran directly from the shell
2018-09-09 04:03:00 <danieli> old habits die hard
2018-09-09 04:07:30 <danieli> brb, back in a few minutes
2018-09-09 04:07:36 <linux_dr> k
2018-09-09 04:17:39 <danieli> all right
2018-09-09 04:17:41 <danieli> did you test it?
2018-09-09 04:21:50 <linux_dr> banging my head against a wall...
2018-09-09 04:22:04 <linux_dr> container exits as soon as it starts..
2018-09-09 04:22:34 <danieli> erm, don't execute the node file directly
2018-09-09 04:23:44 <danieli> docker run --rm -it alpine:3.8
2018-09-09 04:23:59 <danieli> daniel@daniel-pc [~]$ docker run --rm -it alpine:3.8
2018-09-09 04:24:00 <danieli> # apk add nodejs-current
2018-09-09 04:24:32 <linux_dr> i was missing "-it"
2018-09-09 04:24:59 <linux_dr> Well... you said to use the same container... so I am using node:10-alpine
2018-09-09 04:25:32 <danieli> I meant an identical one in terms of what's installed, the settings (if any)
2018-09-09 04:26:03 <linux_dr> Yeh... so, node is already installed.
2018-09-09 04:27:55 <linux_dr> decided it would probably just be easiest to build the container and run the shell in it like you said
2018-09-09 04:28:05 <linux_dr> building now....
2018-09-09 04:28:05 <danieli> that's what I'm doing right now
2018-09-09 04:28:09 <danieli> building?
2018-09-09 04:28:20 <danieli> can't you just `docker pull alpine:3.8` and do it that way?
2018-09-09 04:28:36 <linux_dr> It's one piece in a puzzle.. had to fetch some npm stuff..etc.
2018-09-09 04:28:36 <danieli> nodejs:10-alpine should be practically that, with node and npm installed
2018-09-09 04:29:33 <danieli> that's what I did and it works https://hastebin.com/raw/erekowacar
2018-09-09 04:30:17 <linux_dr> Ok.... in the shell now...
2018-09-09 04:30:45 <danieli> I suspect your other application is doing the listen logic wrong somehow
2018-09-09 04:31:00 <linux_dr> Yeh... no error here.
2018-09-09 04:31:03 <danieli> a lacking env variable, some arguments backwards, something else perhaps
2018-09-09 04:31:09 <danieli> it's hard to say without being able to troubleshoot it myself
2018-09-09 04:33:13 <linux_dr> it seems happy as anything when run directly
2018-09-09 04:34:43 <danieli> try console.log-ing HTTP_PORT and HTTP_HOST so you're sure they're correct, also make sure you parseInt(HTTP_PORT, 10)
2018-09-09 04:34:53 <danieli> env vars are strings
2018-09-09 04:35:34 <linux_dr> I think I figured it out... you'll both probably slap me when I tell you where I goofed..
2018-09-09 04:35:45 <danieli> i knew it was a goof somewhere :)
2018-09-09 04:36:01 <danieli> that's why I forced you through the trouble of testing it and considering it
2018-09-09 04:36:16 <linux_dr> The double-quotes were being sent as part of the value..
2018-09-09 04:36:31 <awilfox> hehe, at least it isn't anything major to fix.
2018-09-09 04:36:51 <danieli> well, crap, lol
2018-09-09 04:36:52 <linux_dr> let me confirm that...
2018-09-09 04:37:30 <danieli> right, you're specifying env vars from the docker-compose.yml
2018-09-09 04:41:02 <linux_dr> what's a docker "repository?
2018-09-09 04:41:51 <danieli> that's really something for #alpine-offtopic, but they're practically images, similar to github
2018-09-09 04:42:04 <danieli> it's the source for building docker containers
2018-09-09 04:42:36 <linux_dr> yeh... no i was trying to figure out an error
2018-09-09 04:43:39 <danieli> #docker perhaps, unless it's related to alpine
2018-09-09 04:44:05 <linux_dr> no... was trying to figure out an error so we could test what we were just discussing.
2018-09-09 04:44:19 <linux_dr> I got it solved... building the whole ball of mud now.
2018-09-09 04:44:28 <danieli> all right, let's take it to #alpine-offtopic if you need something more regarding that
2018-09-09 04:44:34 <danieli> we already littered this channel quite a bit
2018-09-09 04:45:18 <linux_dr> maybe just wait for an ack that all's well.. but yes... definitely off topic.
2018-09-09 04:47:07 <linux_dr> danieli, Yep... that issue's fixed.. thanks for all your help.
2018-09-09 04:47:24 <linux_dr> awilfox, and thanks for your help also.
2018-09-09 04:48:09 <awilfox> np
2018-09-09 04:48:52 <danieli> i've been in your place, struggling with silly, stupid mistakes
2018-09-09 04:49:02 <danieli> especially in the area of docker and node
2018-09-09 04:49:08 <danieli> it's better now though
2018-09-09 04:55:24 <linux_dr> awilfox, Is Adélie source code based, like Gentoo, too?
2018-09-09 04:56:17 <linux_dr> (With APK, I'm guessing not?)
2018-09-09 04:56:50 <danieli> you mean building locally rather than using prebuilt packages?
2018-09-09 04:57:00 <linux_dr> yeh
2018-09-09 05:00:00 <linux_dr> I never played with Gentoo, but remember poking at Sourcere Linux when it came out... disappointed it died out.
2018-09-09 05:00:22 <danieli> mess with freebsd and portsnap
2018-09-09 05:05:13 <linux_dr> if I was in the market for a new distro or os, I'd be more interested... as it stands, I'm stretched rather thin
2018-09-10 00:33:32 <okeuday_bak> the s390x build machine is building with hercules emulation of s390x, right?
2018-09-10 00:35:43 <awilfox> okeuday_bak: no, it is a real IBM System/390 which is donated by IBM, iirc
2018-09-10 00:38:58 <okeuday_bak> awilfox: is it possible to get access to it for testing a package? I have a test that is probably failing due to a timeout, but it would be nice to test it repeatedly.
2018-09-10 00:39:42 <awilfox> I don't know.  I don't have any access to it nor am I an Alpine developer
2018-09-10 00:40:02 <okeuday_bak> awilfox: ok, I can ask when more people are around
2018-09-10 00:53:59 <danieli> tmh1999, ncopa, possibly clandmeter and breno leitao have insight into that
2018-09-10 07:23:49 <ncopa> okeuday_bak: hi, what do you need to test?
2018-09-10 09:33:53 <vkrishn> snownews has new releases for awhile, https://github.com/kouya/snownews/releases
2018-09-10 09:36:25 <vkrishn> seems like nice tool, http://p2f.tuxfamily.org/
2018-09-10 14:51:36 <okeuday_bak> ncopa: wanted to fix https://github.com/alpinelinux/aports/commit/e129d95ea81ee0b1776580b20e4c3a7f79eb023d#diff-60845915858e8dfa820127d84eb60aa1
2018-09-10 14:52:24 <okeuday_bak> so the community/cloudi package for s390x use
2018-09-10 21:13:34 <okeuday> ncopa: didn't see response if you had one before due to disconnect, regarding using s390x build machine for community/cloudi testing, if the machine actually just a hercules emulator that would explain slowness
2018-09-10 21:14:53 <clandmeter> okeuday, its bare metal no emulation at all.
2018-09-10 21:15:43 <okeuday> clandmeter: ok, is it possible to get access to use it for testing?  I am not aware of a different way to get access to that architecture for open-source testing
2018-09-10 21:16:13 <okeuday> unless it would be hercules emulation locally
2018-09-10 21:17:14 <clandmeter> okeuday, yes its difficult to get access to this arch.
2018-09-10 21:17:34 <clandmeter> im not sure if we can provide it, ncopa is the one with access.
2018-09-10 21:18:05 <okeuday> clandmeter: ok
2018-09-10 21:18:40 <clandmeter> there are only a handful of ppl with access to these kind of machines.
2018-09-10 21:18:47 <clandmeter> mostly banks iirc
2018-09-10 21:19:22 <okeuday> clandmeter: yes, I know, but it is a nice use-case I would like to make sure remains valid, as long as the community/erlang package doesn't have issues
2018-09-10 21:19:25 <awilfox> https://linuxone20.cloud.marist.edu/cloud/#/register
2018-09-10 21:19:39 <awilfox> there is 120 day trial for the LinuxONE (s390x) cloud
2018-09-10 21:20:16 <awilfox> maybe you can email / contact IBM somehow and ask if they would give you access if you are genuinely interested in maintaining s390x
2018-09-10 21:21:07 <okeuday> awilfox: Ok, thanks, that might work
2018-09-10 21:42:27 <Gottox> e
2018-09-10 22:10:57 <[[sroracle]]> f
2018-09-10 22:15:44 <mepholic> g
2018-09-10 23:58:44 <danieli> ncopa: check that bug, it's important ^
2018-09-10 23:58:52 <danieli> ncopa: #9408 that is
2018-09-11 05:40:54 <mps> algitbot: :) bug 9410 is not a bug, Debian stretch works under LXC an AL
2018-09-11 06:03:59 <ncopa> okeuday: i was to recommend the timelimited linuxone thing for you
2018-09-11 06:05:28 <ncopa> danieli: thanks. i will look into it
2018-09-11 06:26:24 <danieli> ncopa: good morning :) really just needs your blessing, as far as I can see
2018-09-11 06:43:46 <clandmeter> is it just me or is github having some issues?
2018-09-11 06:54:42 <azarus> clandmeter: i can confirm github is not really responsive
2018-09-11 06:56:56 <okeuday> ncopa: ok, I will try that
2018-09-11 07:27:28 <danieli> clandmeter: someone other friends of mine were having issues with it yesterday
2018-09-11 07:58:45 <_ikke_> clandmeter: heh, github as lots of issues :P
2018-09-11 07:59:11 <_ikke_> couldn't resist
2018-09-11 07:59:22 <clandmeter> similar like  your keyboard?
2018-09-11 07:59:29 <clandmeter> :p
2018-09-11 08:00:35 <_ikke_> ?
2018-09-11 08:00:45 <_ikke_> ah, the 'h' is missing
2018-09-11 08:47:43 <ncopa> anyone else having issues with xfwm4?
2018-09-11 08:47:54 <ncopa> i rebooted my machine and xfwm4 does not start
2018-09-11 08:52:33 <TBB> I haven't had that specific problem but it's always been a bit unstable
2018-09-11 09:06:07 <AinNero> ncopa: i just noticed that the linux kernel may also parse the ip= parameter when CONFIG_IP_PNP is enabled
2018-09-11 09:06:22 <AinNero> which could replace the configure_network code from the initramfs
2018-09-11 09:06:45 <AinNero> but, its enabled for -virt, but not for -vanilla, so it needs to make constant at first
2018-09-11 09:07:11 <AinNero> also im not entirely sure if we want the kernel to do DHCP on its own
2018-09-11 09:08:10 <AinNero> s/constant/consistent/
2018-09-11 09:10:23 <ncopa> AinNero: can you please ping me tomorrow and I'll have a look at the initramfs
2018-09-11 09:10:36 <ncopa> i saw you have a few PRs too
2018-09-11 09:10:41 <ncopa> iirc some was a WIP
2018-09-11 09:10:47 <ncopa> i'd lke to merge those
2018-09-11 09:11:00 <ncopa> but now i have a couple more urgent issues
2018-09-11 09:11:05 <AinNero> ah, those are different issues, relax
2018-09-11 09:11:37 <ncopa> seems like there is a problem with recent apk-tools fix
2018-09-11 09:34:09 <nmeum> ncopa: any chance you could take a look at https://github.com/alpinelinux/aports/pull/5120 this week? This is a first step towards (at least partly) running the go testsuite in check()
2018-09-11 09:40:44 <ncopa> nmeum: i will try
2018-09-11 09:40:59 <ncopa> i would like to have a look at the busybox upgrade too
2018-09-11 09:41:32 <nmeum> that would be nice as well, though the busybox upgrade still needs a lot of work
2018-09-11 09:41:44 <nmeum> might get around to start working on it again this week
2018-09-11 09:44:25 <ncopa> bumped into an apk-tools regression :-/
2018-09-11 10:14:38 <ncopa> phew, its not a regression
2018-09-11 10:32:09 <AinNero> >:(
2018-09-11 10:32:18 <AinNero> the built-in dhcp of the kernel is crap
2018-09-11 10:32:43 <AinNero> it drops the dns settings and stuff into some /proc file, but forgets to mention which device it did DHCP on
2018-09-11 10:36:27 <AinNero> doesn't look like we can relieve the initramfs here..
2018-09-11 11:10:59 <nmeum> fabled: did you get around to rebasing the fbsplash patch? If you haven't started yet I could do it myself as well
2018-09-11 11:26:25 <ncopa> xfwm4 is broken with musl 1.1.20
2018-09-11 11:26:31 <ncopa> i have no clue how to debug it
2018-09-11 11:28:02 <nmeum> what exact is broken?
2018-09-11 11:28:05 <nmeum> *exactly
2018-09-11 11:29:09 <ncopa> xfwm4 exits with error with musl 1.1.20.
2018-09-11 11:30:08 <ncopa> (xfwm4:17839): xfwm4-WARNING **: 13:30:03.224: Could not find a screen to manage, exiting
2018-09-11 11:31:44 <ncopa> ok, its running in valgrind
2018-09-11 11:32:48 <fabled> nmeum, no not yet. though, i could look at it now i think. unless you want to do it
2018-09-11 11:34:58 <nmeum> fabled: since I didn't not write the patch initially I think it would be preferable if you rebase it and, if convinient, also submit it upstream :)
2018-09-11 11:35:57 <fabled> i've sent it upstream couple of times long ago
2018-09-11 11:37:05 <nmeum> did you get any comments on the patch? did they consider integrating it?
2018-09-11 11:38:37 <nmeum> I just think that it would be preferable if our downstream patches which add new features to busybox would be integrated upstream since this would make it easier to update our downstream version
2018-09-11 12:16:12 <fabled> clandmeter, i'm building chromium and I get tons of:
2018-09-11 12:16:12 <fabled> warning: unknown warning option '-Wno-enum-compare-switch'; did you mean '-Wno-enum-compare'? [-Wunknown-warning-option]
2018-09-11 12:16:12 <fabled> warning: unknown warning option '-Wno-null-pointer-arithmetic'; did you mean '-Wno-null-arithmetic'? [-Wunknown-warning-option]
2018-09-11 12:16:12 <fabled> warning: unknown warning option '-Wno-ignored-pragma-optimize'; did you mean '-Wno-ignored-pragma-intrinsic'? [-Wunknown-warning-option]
2018-09-11 12:16:15 <fabled> any ideas?
2018-09-11 12:16:52 <clandmeter> which arch?
2018-09-11 12:17:00 <fabled> all
2018-09-11 12:17:07 <clandmeter> oh hmm
2018-09-11 12:17:12 <fabled> seems clang++ vs g++ difference
2018-09-11 12:17:23 <fabled> apparently we use now clang++ to build chromium
2018-09-11 12:17:26 <clandmeter> i havent touched chromium for some time.
2018-09-11 12:17:29 <clandmeter> ncopa took over.
2018-09-11 13:22:49 <fabled> nmeum, so i got http://tpaste.us/zeNW which is untested, and changes the IMAGE_ALIGN name to IMG_ALIGN to since IMG_POSX is named like that
2018-09-11 13:57:29 <ncopa> i am tagging 3.8.1 now
2018-09-11 14:12:48 <danieli> ncopa: releasing it today?
2018-09-11 14:12:51 <danieli> or just readying up?
2018-09-11 14:15:41 <ncopa> yes, releasing now
2018-09-11 14:16:00 <danieli> neat!
2018-09-11 14:39:45 <clandmeter> ncopa, can we add support for signed modloop?
2018-09-11 14:40:21 <clandmeter> most bits are already in place
2018-09-11 14:47:05 <ncopa> er.. sure
2018-09-11 14:47:12 <ncopa> i just tagged it...
2018-09-11 14:50:15 <AinNero> does this imply an mkinitfs release, clandmeter?
2018-09-11 14:51:34 <clandmeter> ncopa, i think 1 min between anounce and actual release is a bit short :)
2018-09-11 14:52:21 <clandmeter> AinNero, it will mean a new release or just a patch to aports mkinitfs
2018-09-11 14:52:31 <clandmeter> but im not sure we can still change it now.
2018-09-11 14:53:23 <clandmeter> i was hoping to update boot.a.o with edge releases and modloop verification.
2018-09-11 14:59:58 <ncopa> in general, make bugreports, set target 3.8.x for those things that you want in stable release
2018-09-11 15:00:28 <ncopa> i use that as checklist to verify that i havent forgotten anythign important before i tag release
2018-09-11 15:01:16 <ncopa> so if there is something you want included in next stable, file an issue
2018-09-11 15:24:50 <ncopa> does this release notes look ok? http://wwwtest.alpinelinux.org/posts/Alpine-3.8.1-released.html
2018-09-11 15:24:57 <_ikke_> checking
2018-09-11 15:47:16 <ncopa> no complains. I think I just push it.
2018-09-11 15:47:32 <_ikke_> right, it's pretty trivial\
2018-09-11 19:02:55 <ncopa> i pushed alpine 3.7.1 too
2018-09-11 19:03:07 <ncopa> and i think i will do 3.6.3
2018-09-11 19:03:16 <ncopa> we need the minirootfs images
2018-09-11 19:06:48 <_ikke_> \0/
2018-09-11 19:06:51 <danieli> \o/
2018-09-11 19:07:01 <danieli> ncopa: that secfix?
2018-09-11 19:07:36 <unmy> \o/
2018-09-11 19:08:27 <ncopa> yup
2018-09-11 19:13:40 <HRio> \o/
2018-09-11 19:13:51 <ncopa> not sure i should do 3.5...
2018-09-11 19:16:27 <duncan^> For users using -hardened kernel in 3.7, is there some message e.g. in /etc/motd or /etc/issue telling folks to switch to the -vanilla kernel or upgrade to 3.8.x?
2018-09-11 19:16:56 <duncan^> Or is 3.7 entirely out of support essentially and users should just upgrade?
2018-09-11 19:17:35 <_ikke_> 3.7 is still supported
2018-09-11 19:17:45 <ncopa> kernel is not
2018-09-11 19:17:46 <ncopa> i mean
2018-09-11 19:17:58 <_ikke_> right
2018-09-11 19:18:26 <ncopa> kernel in 3.7 is entirely out of support and users should just upgrade
2018-09-11 19:18:48 <ncopa> the reason i am doing the older releases now is that we also ship a minirootfs that is needed for the docker images
2018-09-11 19:19:27 <ncopa> for solving this one: https://github.com/gliderlabs/docker-alpine/issues/413
2018-09-11 19:19:42 <ncopa> i think we need a way to do "light" releases in future
2018-09-11 19:19:59 <ncopa> where we only generate minirootfs images, and not the full isos
2018-09-11 19:20:39 <_ikke_> And also communicate to users that it's only the minirootfs images that get upgraded
2018-09-11 19:23:35 <ncopa> im not announcing the older releases
2018-09-11 19:26:50 <ncopa> ok, i also tagged 3.5.3
2018-09-11 19:27:07 <ncopa> i did it before i reached to regret :)
2018-09-11 19:27:25 <_ikke_> hehe
2018-09-11 19:27:27 <_ikke_> :D
2018-09-11 19:35:33 <PureTryOut[m]> could any Matrix admin in this room add the +freenode:matrix.org community? that way IRC users should get a `#` flair besides their name, making it a bit clearer who is using what
2018-09-11 19:44:38 <ncopa> clandmeter: ^^^
2018-09-11 20:04:07 <clandmeter> PureTryOut[m], i have no idea what you mean
2018-09-11 20:04:28 <clandmeter> i stopped using matrix so im not on par with the tech
2018-09-11 20:06:10 <danieli> who /are/ alpine matrix admins?
2018-09-11 20:06:13 <unmy> 
ACTION reading about matrix, before I thought [m] mean some kind of 'moderator' status :D

2018-09-11 20:06:23 <danieli> I thought it was mobile
2018-09-11 20:06:28 <unmy> :D
2018-09-11 20:06:31 <danieli> at first, anyway
2018-09-11 20:07:09 <clandmeter> me and ncopa are
2018-09-11 20:07:15 <clandmeter> but we dont use it anymore
2018-09-11 20:07:22 <clandmeter> because of broken bridge
2018-09-11 20:07:31 <danieli> right, but other people use it
2018-09-11 20:07:57 <clandmeter> i know
2018-09-11 20:08:41 <clandmeter> if one of the alpine matrix users want to take over thats fine by me.
2018-09-11 20:09:33 <clandmeter> another option would be to remove the server and switch back to using matrix.org
2018-09-11 20:09:46 <ncopa> i support that
2018-09-11 20:11:00 <clandmeter> i need to check which steps
2018-09-11 20:13:34 <PureTryOut[m]> well, Matrix has a thing called communities now. basically the only thing the admin has to do is add one of those to this room in the roomsettings, just like you add a room alias to it
2018-09-11 20:19:48 <PureTryOut[m]> no the room will keep existing
2018-09-11 20:19:57 <PureTryOut[m]> just the alias would be gone #alpine-devel:alpinelinux.org
2018-09-11 20:20:04 <PureTryOut[m]> (which would be a shame 😉 )
2018-09-11 20:22:24 <clandmeter> well, that doesnt add up to the amount of resources we need to reserve for keeping it running/
2018-09-11 20:22:47 <danieli> and maintain it
2018-09-11 20:23:07 <danieli> nevermind that - "keeping it running" covers it
2018-09-11 20:23:14 <clandmeter> i actaully wrote another msg on matrix and it didnt pop up here...
2018-09-11 20:23:19 <danieli> been up for way too long, I'm off for some shut-eye
2018-09-11 20:23:24 <danieli> lol, yeah.. that too.
2018-09-11 20:24:08 <clandmeter> ok lets set next week to stop our local matrix instance.
2018-09-11 20:24:16 <danieli> works for me
2018-09-11 20:24:29 <clandmeter> if somebody has good reasons to keep it let us know.
2018-09-11 20:24:42 <clandmeter> and i mean *good* reasons :)
2018-09-11 20:25:26 <PureTryOut[m]> ooh you're shutting it down? damn 😕
2018-09-11 20:25:48 <PureTryOut[m]> well, maybe you guys can try again once Dendrite or another homeserver is usable
2018-09-11 20:26:08 <danieli> I've never even heard of that to be honest
2018-09-11 20:26:15 <PureTryOut[m]> anyways, I'd still like you to add that community 😉
2018-09-11 20:26:30 <danieli> can bring this up tomorrow some time for a discussion, I'm off for now
2018-09-11 20:26:37 <danieli> also, may I suggest #alpine-infra for that?
2018-09-11 20:27:06 <clandmeter> i know dendrite
2018-09-11 20:27:09 <clandmeter> but dev is slow
2018-09-11 20:27:20 <clandmeter> its made in go
2018-09-11 20:54:46 <PureTryOut[m]> could you add that community to the room though? +freenode:matrix.org
2018-09-11 20:55:15 <andypost> somehow I can't write from riot via both ways now - server & matrix bridge
2018-09-11 21:35:49 <awilfox> funny timing; the IRC network that Adélie is on just removed its matrix bridge too because messages weren't going through properly and the homeserver would crash if multiple users using quaternion joined the same room at the same time...
2018-09-12 13:05:09 <danieli> how come we don't have python 3.7?
2018-09-12 15:26:10 <_ikke_> useless answer: no one bothered to update it yet
2018-09-12 15:33:36 <danieli> i thought there was more to it
2018-09-12 15:33:48 <danieli> something dependent on it breaks with 3.7, or.. something
2018-09-12 17:00:04 <ncopa> havent had time to llok at python 3.7 yet
2018-09-12 17:00:11 <ncopa> i think we should update gcc first
2018-09-12 17:00:19 <danieli> probably a good idea
2018-09-12 17:00:26 <danieli> i'm in the process of testing a py3.7 build
2018-09-12 17:00:38 <danieli> builds fine with just a version bump, just need to make sure some tests run
2018-09-12 17:01:59 <danieli> i got erlang/otp 21.x (21.0.9 atm) in my todo too
2018-09-12 17:02:14 <danieli> and to finish rust-nightly for a good friend
2018-09-12 17:21:06 <danieli> worth noting: a test in py 3.7 requires TLS 1.3, which openssl just released iirc
2018-09-12 17:27:54 <danieli> openssl 1.1.1 specifically
2018-09-12 17:39:05 <PureTryOut[m]> could somebody review this small PR of mine? https://github.com/alpinelinux/aports/pull/4479
2018-09-12 17:39:19 <_ikke_> I can review it, but cannot merge it
2018-09-12 17:46:58 <ncopa> PureTryOut[m]: im looking at it
2018-09-12 17:47:04 <ncopa> sorry for taking so long time
2018-09-12 17:47:25 <ncopa> i think i will push it as is, but i think its unnecessarily bloaty
2018-09-12 17:48:10 <ncopa> flatc seems to link static to libflatbuffers
2018-09-12 17:48:39 <ncopa> i wonder if we want to ship flatc in a subpackage
2018-09-12 18:55:30 <craftyguy> is there a mailing list where security announcements (e.g. CVEs, etc) are made? Wiki says 'alpine-security' list is no longer used :(
2018-09-12 18:55:58 <craftyguy> or, how do folks that use Alpine know when a CVE that affects a component in Alpine is disclosed?
2018-09-12 21:13:59 <PureTryOut[m]> ncopa: I guess it could be put into `-dev`?
2018-09-12 21:14:00 <PureTryOut[m]> thanks for merging it for now though
2018-09-13 06:16:19 <ncopa> craftyguy: you can get them from bugs.a.o, category "security"
2018-09-13 06:17:31 <ncopa> set filter to also include closed ones
2018-09-13 06:17:51 <ncopa> at the bottom you have a link to an atom feed
2018-09-13 06:19:32 <craftyguy> ncopa: awesome, yea i just discovered the atom feed :D
2018-09-13 06:20:17 <craftyguy> basically I just want to be notified when there are updates to packages for security reasons
2018-09-13 06:20:40 <craftyguy> rather than updates for packages just because some new feature or non-security fix
2018-09-13 06:20:58 <craftyguy> this may work though, I'll have to play with it
2018-09-13 06:21:11 <ncopa> you could watch the stable branch too
2018-09-13 06:22:11 <craftyguy> do commits for security updates have any special tag in the commit message, or commit summary?
2018-09-13 06:22:47 <ncopa> normally we have seomthing like "security" or "CVE" in the commit message yes
2018-09-13 06:23:39 <ncopa> sometimes we are not made aware that an update had security implications long time after
2018-09-13 06:23:50 <ncopa> we dont change the git log in those cases :)
2018-09-13 06:24:02 <craftyguy> nice. watching the stable branch for commits with that in the message might be a better solution
2018-09-13 06:24:16 <craftyguy> sure, can't expect to catch 'em all :P
2018-09-13 06:24:26 <ncopa> i think so too
2018-09-13 06:25:02 <craftyguy> ncopa: thanks for the help, I appreciate it!
2018-09-13 06:25:14 <ncopa> in 3.8-stable we have some non-security fixes, in 3.7-stable and older it is almost only security fixes
2018-09-13 06:26:11 <ncopa> i know there is room for improvement to announce security fixes, so thank you for you patience :)
2018-09-13 06:28:22 <craftyguy> good to know, I believe we're currently using 3.7 (but would have to double check)
2018-09-13 07:00:52 <_ikke_> craftyguy: The problem is that some bugfixes aren't marked as security fix but later discoverd to fix an actual security issue
2018-09-13 07:01:02 <_ikke_> for example the recent openssh user enumeration
2018-09-13 07:01:21 <_ikke_> so only focussing on security fixes might cause you to miss out on important fixes
2018-09-13 11:37:25 <clandmeter> kunkku, is it normal for awall to have issues when iptables has an empty ruleset?
2018-09-13 11:37:30 <clandmeter> like for intial installation.
2018-09-13 11:37:35 <clandmeter> install/setup
2018-09-13 11:37:46 <fabled__> clandmeter, i usually have needed to enable iptables manually
2018-09-13 11:38:07 <clandmeter> enable as in start it from openrc?
2018-09-13 11:41:15 <fabled__> load stuff once before 'awall activate' starts working
2018-09-13 11:41:19 <fabled__> mostly get some kernel module right
2018-09-13 11:41:37 <fabled__> and yes, enable iptables manually
2018-09-13 11:45:20 <clandmeter> ah yes that was what i did before
2018-09-13 11:45:24 <clandmeter> load iptables module
2018-09-13 11:47:35 <fabled__> kunkku, would be kinda nice if 'awall activate' or initialize would setup openrc+iptables in kernel
2018-09-13 11:47:48 <AinNero> oh, RE: the ModSecurity issues on alpine, i really dont want to encourage putting "firewalls" infront of your applications instead of fixing them
2018-09-13 11:48:55 <AinNero> _ikke_: also anther thing, the IT Security field is the only place where the SSH thing is an "enumeration"
2018-09-13 11:49:23 <AinNero> and this is likely because the discoverers of that bug never studied math
2018-09-13 11:50:12 <AinNero> more accurate name would be "oracle" probably
2018-09-13 11:51:31 <_ikke_> AinNero: yeah, I copied that term from another source, but didn't think it was a good term
2018-09-13 11:51:44 <AinNero> its nothing personal against you
2018-09-13 11:51:48 <_ikke_> nod
2018-09-13 11:51:56 <_ikke_> You cannot enumerate the users, you still have to guess
2018-09-13 11:52:11 <AinNero> im mostly annoyed that people use wrong terms and never correct it
2018-09-13 11:52:19 <AinNero> so the wrong use becomes popular
2018-09-13 11:52:31 <_ikke_> Welcome to language 101
2018-09-13 11:53:08 <AinNero> Like, everytime when i tell people that i have an VPN, they ask me which provider
2018-09-13 11:53:18 <_ikke_> heh
2018-09-13 11:53:24 <AinNero> and im like, no, i just have an literal virtual private network
2018-09-13 11:53:28 <AinNero> literally literally
2018-09-13 12:42:37 <Michcioperz> what ssh thing is/isn't enumeration?
2018-09-13 12:44:29 <_ikke_> Michcioperz: There was a bug fixed in ssh where ssh would give a different response when a user was present on the system than when a user wasn't and a malformed packet was sent
2018-09-13 12:46:29 <Michcioperz> ah you mean it's not enumeration cause you can't bend it to throw the whole list
2018-09-13 12:46:38 <Michcioperz> you can only check every combination
2018-09-13 12:46:39 <_ikke_> correct
2018-09-13 12:59:06 <danieli> iirc it measured the time taken to respond with an error message to deduce whether the username exists or not
2018-09-13 12:59:57 <_ikke_> no, this was not a timing attack
2018-09-13 13:01:30 <_ikke_> In one cases it returns an error message, in the other case it closes the connection
2018-09-13 19:14:13 <anoldhacker>  I'm looking for something along the lines of authbind (https://en.wikipedia.org/wiki/Authbind) -- I want a non-root user to be able to bind to port 80 without going crazy in an AWS Fargate deploy.  Thoughts?
2018-09-13 19:27:28 <anoldhacker> Looks like Alpine Wall can do a port forward from 80 to a higher port.  If only I had done such a thing before...
2018-09-13 19:56:13 <ncopa> https://justi.cz/security/2018/09/13/alpine-apk-rce.html
2018-09-13 19:56:43 <_ikke_> Was this the one fixed in 3.8.1?
2018-09-13 19:56:56 <ncopa> yes
2018-09-13 19:56:58 <_ikke_> ok
2018-09-13 19:57:18 <ncopa> and it was the reason that we did 3.7.x, 3.6.x, 3.5.x releases too
2018-09-13 19:57:28 <_ikke_> Yeah, I understood
2018-09-13 20:13:53 <_ikke_> Is this bsd-3-clause? https://raw.githubusercontent.com/ProgVal/Limnoria/master/LICENSE.md I see some minor differences, which could be an older revision
2018-09-13 20:14:33 <_ikke_> http://tpaste.us/5YB7
2018-09-13 20:15:58 <ncopa> https://news.ycombinator.com/item?id=17981452
2018-09-13 20:16:38 <_ikke_> "I'm not convinced they fixed it.
2018-09-13 20:16:40 <_ikke_> "
2018-09-13 20:17:21 <ncopa> "Reading the *commit message* ... I'm not convinced they fixed it."
2018-09-13 20:17:35 <_ikke_> right
2018-09-13 20:18:28 <ncopa> so, he is not convinced just because fabled didn't write in the commit message "this fixes the issue that will soon be posted in a blog article and hackernews"
2018-09-13 20:19:36 <_ikke_> That's not how I read what he said
2018-09-13 20:21:47 <_ikke_> Not that his just his opinion matters a lot
2018-09-13 20:22:46 <ncopa> I would have been more worried if he said "reading the code changes in the commit..."
2018-09-13 20:24:06 <_ikke_> nod
2018-09-13 20:28:29 <mps> is upgrading apk-tools enough to be safe from that vulnerability?
2018-09-13 20:29:42 <ncopa> yes
2018-09-13 20:29:50 <ncopa> should be
2018-09-13 20:31:45 <mps> ah, I upgraded yesterday (when 3.8.1 is released) but forgot one box :(
2018-09-13 20:32:18 <mps> it is good to lurk IRC
2018-09-13 20:42:18 <mps> if there are no /etc/apk/commit_hooks.d/ dir then the nachine is not attacked?
2018-09-13 20:42:41 <_ikke_> I don't think there are any active exploits for this
2018-09-13 20:42:51 <mps> or the attacker could delete this dir in his code?
2018-09-13 20:42:53 <_ikke_> someone already would need to mitm you
2018-09-13 20:42:56 <_ikke_> yes
2018-09-13 20:43:57 <mps> I hope so. But wanted to hear what devs think about that issue
2018-09-13 20:44:08 <ncopa> mps: if there are no /etc/apk/commit_hooks.d/ then you are likely not attacked
2018-09-13 20:44:53 <ncopa> i bet there are lots of other ways to autoexecute the code once its injected
2018-09-13 20:45:14 <_ikke_> yes, it would just be the initial attack vector
2018-09-13 20:46:06 <mps> aha, time to think about GPG/PGP signing apk repos
2018-09-13 20:46:18 <_ikke_> They are already signed
2018-09-13 20:47:04 <ncopa> they are signed but not with gpg
2018-09-13 20:47:21 <mps> _ikke_: I see RSA keys in the /etc/apk/keys
2018-09-13 20:47:25 <ncopa> the problem at hand is that they found a way to bypass the signature
2018-09-13 20:47:55 <ncopa> its time to redesign the apk format and index format
2018-09-13 20:48:04 <_ikke_> It's that apk already modifies the filesystem before the hash is hcecked
2018-09-13 20:48:09 <_ikke_> checked
2018-09-13 20:48:34 <ncopa> we have been working on redesign of the apk and index format for a while (more than a year at least)
2018-09-13 20:49:14 <ncopa> this issue just gave us a bit motivation to speed up the work :)
2018-09-13 20:49:33 <_ikke_> heh
2018-09-13 20:50:00 <ncopa> i better get some sleep
2018-09-13 20:50:04 <ncopa> good night
2018-09-13 20:50:05 <_ikke_> sleep tight
2018-09-13 20:50:38 <mps> good night
2018-09-14 06:47:00 <danieli> ncopa: that license (progval/limnoria) was bsd-3-clause, yes
2018-09-14 06:59:38 <_ikke_> danieli: that was me :)
2018-09-14 07:00:22 <danieli> early morning and lots to do, sorry about that
2018-09-14 07:09:08 <_ikke_> no worry
2018-09-14 10:20:08 <Fusl> jirutka[m]: are you on?
2018-09-14 10:51:53 <Fusl> or maybe someone else on who can help me make this PR compliant to the aports repo? https://github.com/alpinelinux/aports/pull/4438
2018-09-14 10:52:54 <_ikke_> So jirutka provided some feedback, did you make those changes?
2018-09-14 10:53:01 <Fusl> i did
2018-09-14 10:53:13 <_ikke_> You did not push them yet apparently
2018-09-14 10:53:24 <Fusl> i did not, i tried the changes locally and they dont work as expected
2018-09-14 10:53:46 <_ikke_> ah ok
2018-09-14 10:53:53 <Fusl> more specifically, when i install the package on another system, the group of the kismet_capture file is not correctly reassigned to the `kismet` group
2018-09-14 10:55:11 <Fusl> removed `$pkgname.post-install`, replaced `-g root` with `-g kismet` for the `install` command and added `pkggroups="kismet"`
2018-09-14 10:55:43 <Fusl> on a freshly installed alpine system, the file gets created with an unknown group: -r-xr-x---    1 root     103         465352 Sep 14 10:36 /usr/bin/kismet_capture*
2018-09-14 10:56:31 <danieli> oh yeah, kismet, I was working on an apkbuild for that
2018-09-14 10:59:08 <Fusl> i can push the changes i made locally to the repo so you can try it out on your side, _ikke_
2018-09-14 11:02:28 <_ikke_> I have only time later
2018-09-14 11:09:28 <andypost> Needs to backport to 3.8 http://dup.pw/aports/3f15c0b2 http://dup.pw/aports/d3f643b6
2018-09-14 13:38:27 <kunkku> fabled: regarding easy initial awall setup
2018-09-14 13:38:40 <kunkku> I have recently implemented: https://git.alpinelinux.org/cgit/aports/tree/main/awall/setup-firewall
2018-09-14 13:39:14 <kunkku> have to think what is the correct responsibility split between that and
2018-09-14 13:39:19 <kunkku> awall activate
2018-09-14 19:16:11 <AinNero> the wiki entry on PXE boot is still referring to the alpine_dev and NFS functionality
2018-09-14 19:16:16 <AinNero> which has been removed from mkinitfs
2018-09-14 19:16:43 <AinNero> is NFS boot media support something maintain-worthy?
2018-09-14 21:00:14 <PureTryOut[m]> could somebody give voice to MartijnBraam and then remove it again to work around the permission issue on the Matrix side? it was done to me in the past as well, but I can't remember who did it
2018-09-14 22:35:09 <Fusl> andypost[m]: are you on?
2018-09-14 22:38:34 <Fusl> andypost[m]: regarding your comment at https://github.com/alpinelinux/aports/pull/5172#issuecomment-421462230, we can definitely split the init.d directory into its own -openrc subpackage (i guess its as easy as just adding subpackages="$pkgname-openrc" like with -doc?) but wouldnt that cause the package installation to no longer include the init.d file for systems that need it?
2018-09-14 23:50:31 <[[sroracle]]> Fusl: the openrc default split function includes an install_if rule so that if openrc is installed the -openrc subpackage automatically gets installed
2018-09-15 12:04:26 <Ganwell> Recently I had few python packages breaking due bad dependencies. certbot needs idna<2.7, we install idna==2.7, we might or might not be able to downgrade idna, because of other packages. Python does not allow multiple version of the same packages anyways. In theory stop delivering python packages at all, let people use virtual-envs. That thats a silly solution too.
2018-09-15 12:06:57 <_ikke_> What is preventig you from using virtenvs right now?
2018-09-15 12:07:00 <_ikke_> preventing*
2018-09-15 12:09:44 <Ganwell> I can solve the problem. Thats not the thing. I am concerted that we deliver broken packages to the user.
2018-09-15 12:10:09 <Ganwell> concerned
2018-09-15 12:11:21 <Ganwell> On my system I did pip install "idna<2.7" -> solved for me, but the office 3.8 certbot pacakge is still broken.
2018-09-15 12:11:33 <Ganwell> offical
2018-09-15 12:11:42 <_ikke_> on 3.8 or only on edge?
2018-09-15 12:11:47 <Ganwell> 3.8
2018-09-15 12:13:35 <_ikke_> How / when is it broken? Seems to work for me
2018-09-15 12:14:03 <Ganwell> https://gist.github.com/ganwell/9169d24e6da4dbeba0c49c5252685b5b
2018-09-15 12:14:50 <Ganwell> it has to be broken, becuase it requires idna<2.7 but idna==2.7 is installed
2018-09-15 12:15:42 <_ikke_> right, so certbot 0.25 says it requries it
2018-09-15 12:15:55 <_ikke_> edge has 0.27 which does not have the issue anymore
2018-09-15 12:16:30 <Ganwell> so upgrading to certbot 0.27 is also an option
2018-09-15 12:18:13 <_ikke_> Please report a bug at bugs.a.o
2018-09-15 12:20:10 <Ganwell> My problem is not current concrete problem, there are usually ways to resolve that, but fact, that these things happen on current stable. I am interested in the correct strategy to avoid that. Probably adding a smoke test to all python packages would help?
2018-09-15 12:20:43 <_ikke_> yes
2018-09-15 12:20:50 <_ikke_> definetely
2018-09-15 12:21:05 <_ikke_> We are pushing to have more tests in packages, but a lot of projects do not offer them yet
2018-09-15 12:22:35 <Ganwell> I usually add a smoke test, if there are not tests, it will trigger these "pkg_resources.ContextualVersionConflict" errors.
2018-09-15 12:23:07 <Ganwell> I'll do a PR adding a test to certbot and requesting a backport
2018-09-15 17:50:39 <_ikke_> hmm: https://travis-ci.org/alpinelinux/aports/builds/429037878#L557
2018-09-15 17:50:58 <_ikke_> It says no packages found to be build, even though the diff shows a diff
2018-09-17 07:37:39 <azarus> I'd like to know how the data was generated in https://git.alpinelinux.org/cgit/apk-tools/tree/src/fetch.c#n40
2018-09-17 07:38:04 <azarus> How was it compressed, and how can one do that to arbitrary text?
2018-09-17 07:40:32 <Michcioperz> `#include <zlib.h>`
2018-09-17 07:46:06 <Michcioperz> what's that text btw?
2018-09-17 07:47:48 <azarus> a coffee cup
2018-09-17 07:48:41 <Michcioperz> oh
2018-09-17 07:48:43 <Michcioperz> lol
2018-09-17 07:49:08 <azarus> (now, i know it's compressed using zlib, but how can I use zlib on arbitrary files?)
2018-09-17 07:49:30 <_ikke_> azarus: what do you want to achieve exactly>
2018-09-17 07:49:32 <_ikke_> ?
2018-09-17 07:49:58 <azarus> _ikke_: compress text so I can do something akin to what is done in fetch.c
2018-09-17 07:52:37 <azarus> I'd just like to keep compressed text in a .c file.
2018-09-17 07:53:11 <_ikke_> just make a small c program dat reads the data from a file, compresses it, and then outputs each byte in a c-array fashion
2018-09-17 07:53:16 <azarus> OK
2018-09-17 07:53:26 <azarus> thanks!
2018-09-17 07:53:44 <_ikke_> naturally it doesn't have to be c
2018-09-17 07:53:56 <_ikke_> any language should be capable of doing that
2018-09-17 07:53:59 <azarus> but to use zlib you likely have to use C, right
2018-09-17 07:54:26 <_ikke_> no
2018-09-17 07:54:35 <_ikke_> many languages have bindings for it
2018-09-17 07:54:48 <Michcioperz> if I understand this is basically gzip?
2018-09-17 08:24:34 <ncopa> morning
2018-09-17 08:24:46 <ncopa> yes, this is basically gzip
2018-09-17 08:29:29 <azarus> morning
2018-09-17 09:11:58 <AinNero> the fuck
2018-09-17 09:12:05 <_ikke_> ?
2018-09-17 09:12:19 <AinNero> did i just see Torvalds adopt Coraline Ada's CoC?
2018-09-17 09:18:58 <azarus> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8a104f8b5867c682d994ffa7a74093c54469c11f
2018-09-17 09:19:12 <_ikke_> --> #alpine-offtopic :-)
2018-09-17 09:19:45 <azarus> (sorry
2018-09-17 09:19:46 <azarus> )
2018-09-18 09:08:14 <vkrishn> is templates/lxc-alpine script maintained in git ?
2018-09-18 11:08:15 <xsteadfastx> could it be that there was a ssl update in edge? bitlbee in the edge repo cant connect to anything... with 3.8 no problem. i guess bitlbee needs to be recompiled
2018-09-18 11:11:59 <_ikke_> hmm, I don't see any substansive changes on libressl
2018-09-18 11:12:03 <_ikke_> since 3.8.0
2018-09-18 11:12:43 <_ikke_> the openssl package has been renamed to openssl1.0
2018-09-18 11:13:00 <danieli> openssl 1.1.1 supports tls 1.3, shame some things depend on that
2018-09-18 11:13:07 <danieli> some *newer* things
2018-09-18 11:14:59 <AinNero> i know people who had problems with DNS with musl 1.1.20
2018-09-18 11:15:13 <danieli> i heard chatter about that too
2018-09-18 13:58:10 <danieli> and there it is
2018-09-18 15:55:02 <_ikke_> What strategies do people use here to check / get notified on new versions of upstream projects? I cannot imagine everyone is constantly checking each package if there are updates available
2018-09-18 15:55:21 <_ikke_> (I'm aware of flagging in pkgs.a.o, but that doesn't cover everything)
2018-09-18 16:15:49 <awilfox> _ikke_: repology
2018-09-18 16:17:05 <_ikke_> ah, yeah, came across that, but forgot about it
2018-09-18 16:23:44 <_ikke_> "meson.build:127:2: ERROR:  Native dependency 'systemd' not found" :-/
2018-09-18 16:24:45 <AinNero> _ikke_: which upstream?
2018-09-18 16:25:36 <_ikke_> colord
2018-09-18 16:25:52 <_ikke_> you can disable it apparently (meson -Dsystemd=false)
2018-09-18 16:39:44 <_ikke_> Anyone know how to specify the installation location with meson / ninja?
2018-09-18 16:39:59 <_ikke_> The readme doesn't mention the install step at all
2018-09-18 16:40:05 <awilfox> -Dprefix=/foo/bar
2018-09-18 16:40:18 <_ikke_> I mean for packaging
2018-09-18 16:40:20 <awilfox> oh wait
2018-09-18 16:40:25 <awilfox> DESTDIR doesn't work?
2018-09-18 16:40:41 <awilfox> all the things I've packaged with meson and ninja support DESTDIR
2018-09-18 16:40:46 <_ikke_> ninja install DESTDIR=../pkg/
2018-09-18 16:40:49 <awilfox> no
2018-09-18 16:40:53 <_ikke_> ninja: error: unknown target 'DESTDIR=../pkg'
2018-09-18 16:40:55 <_ikke_> ok
2018-09-18 16:40:56 <awilfox> DESTDIR=../pkg/ ninja install
2018-09-18 16:41:00 <_ikke_> ah ok
2018-09-18 16:41:10 <_ikke_> that's a Make thing then :)
2018-09-18 18:07:47 <_ikke_> Fun, a test suite of a project failing due to an assert failing
2018-09-18 18:43:25 <_ikke_> Seems like it's trying to connect to udev to get some information
2018-09-18 19:11:53 <_ikke_> udev_hwdb
2018-09-19 06:19:54 <xsteadfastx> _ikke_: it was just a guess. because bitlbee cant connect to anything ssl if i install it in alpine edge
2018-09-19 06:20:28 <xsteadfastx> to test i use the same config with bitlbee and its deps from 3.8 and it works like it used to last week... until i rebuild my container
2018-09-19 06:22:08 <xsteadfastx> AinNero: so there was a musl update in edge?
2018-09-19 06:31:36 <clandmeter> xsteadfastx, there was
2018-09-19 06:31:48 <clandmeter> xsteadfastx, is there no log?
2018-09-19 06:57:55 <xsteadfastx> clandmeter: of bitlbee? nope. even in verbose mode. i need to find out how to start a debugger...
2018-09-19 06:59:06 <xsteadfastx> bitlbee always tries to connect and it fails. network itself works
2018-09-19 06:59:12 <xsteadfastx> i can ping the servers
2018-09-19 06:59:26 <clandmeter> without log its hard to tell. alternative is to strace the pid.
2018-09-19 07:02:00 <xsteadfastx> i google around how to strace
2018-09-19 07:51:07 <xsteadfastx> this should be the problem in a strace: "07:50:03 read(3, 0x7ffc8bcc6c68, 16)    = -1 EAGAIN (Resource temporarily unavailable)"
2018-09-19 07:51:13 <xsteadfastx> this comes up everytime it tries to connect
2018-09-19 09:31:07 <lskillen> Hey all - Is there any appetite for a package management SaaS that supports Alpine Linux (for hosted public/private repositories)? There's nothing in the official wiki about hosted repository providers, so I thought it would be better to ask here. ;)
2018-09-19 10:23:02 <AinNero> what do they do?
2018-09-19 10:24:00 <AinNero> ping lskillen
2018-09-19 10:26:24 <AinNero> like, one can safely host alpine repos on one-click static webhosts
2018-09-19 10:27:02 <AinNero> a company with too much money could push it into an S3 bucket, maybe?
2018-09-19 10:27:40 <lskillen> Basically an easy way to create your own custom alpine-compatible repository (i.e. works with "apk install"), for public and/or private packages, but it is hosted/managed for you in the Cloud. Some people use custom repositories for their own private assets, or as a cache against public/upstream repositories so that your pipelines still work when they're down.
2018-09-19 10:28:22 <_ikke_> The only thing you need for hosting your own repositories is storage + a webserver
2018-09-19 10:28:31 <AinNero> ^
2018-09-19 10:28:53 <AinNero> i have my own private repositories on an shared hosting provider that runs of donations
2018-09-19 10:30:38 <lskillen> Well, possibly. :) Everybody has different requirements - That's fine if you're OK with managing and building it yourself. Usually people start to think about managed repositories when they're using more than just one technology (e.g. Debian packages in addition to Alpine, or is that an evil word here?), or you're looking for org/team management, security controls, automatic handling for package
2018-09-19 10:30:44 <lskillen> signing, etc.
2018-09-19 10:33:17 <lskillen> That probably makes it obvious that it tends to be companies (or at least, groups of devs) that use this sort-of thing. ;)
2018-09-19 10:34:08 <AinNero> hm.
2018-09-19 10:35:14 <lskillen> In saying that, the service has a generous amount of free storage/bandwidth/features on the free plan, as long as the repositories are public, and if the user/org is hosting open-source then they get a huge amount of storage/bandwidth for free.
2018-09-19 10:35:29 <AinNero> i disapprove of this kind of software development culture fetishizing fully featured magic boxes, but thats my personal opinion
2018-09-19 10:35:54 <lskillen> Aye, I get that - I'm an open-source maintainer and developer myself :)
2018-09-19 10:37:53 <lskillen> The service is built to be as transparent as possible, offers data portability in/out (so easy to get a dump of all repositories if you've looking to "move out"), tries to help the community with free/open-source offerings, and isn't about capitalising on other technologies alone (i.e. contributions to open-source, builds all integrations as open-source, and future plans to move actual software to
2018-09-19 10:37:59 <lskillen> open-core for full visibility).
2018-09-19 10:40:43 <AinNero> i guess it would be hip with startups
2018-09-19 10:41:19 <lskillen> Anyway, not here to tout or spam - I think Alpine is awesome and we're fans of it (as containerisation users). If we ever migrate from fledgling startup to "made it" I'll make sure we can give back to the technologies we support in other ways (e.g. financial sponsorship).
2018-09-19 10:41:33 <lskillen> Suppose so ;)
2018-09-19 10:41:58 <AinNero> like, im not too happy with this
2018-09-19 10:42:50 <AinNero> there is that crowd where "it works" seems to be the scale of success, without much understanding of the details
2018-09-19 10:43:34 <AinNero> and like, after two years of github i understand why people disable the issues on their projects
2018-09-19 10:43:41 <lskillen> hehe
2018-09-19 10:44:19 <lskillen> Is this experience from the maintainership/developer angle?
2018-09-19 10:45:10 <AinNero> sh4rm4^bnc might remember that one person who wanted to make an container for sabotage linux
2018-09-19 10:45:28 <AinNero> and they didn't get what self-hosted and self-compiling meant
2018-09-19 10:45:59 <AinNero> and after quite some work, they learned that their expectations of something smaller than alpine were unrealistic
2018-09-19 10:46:17 <AinNero> because like, sabotage is an compiled distro, so you need a toolchain, which is usually bug
2018-09-19 10:46:21 <AinNero> *big
2018-09-19 10:47:26 <AinNero> i dont want to point at others and call them dumb and act like im not making mistakes, but like, i felt sort of disgusted
2018-09-19 10:47:43 <lskillen> Well not to worry here, this isn't about reinventing or packaging anything up differently, or anything of that sort
2018-09-19 10:48:28 <lskillen> It isn't even about Alpine alone, there are all sorts of other packaging formats in the mix - It's much closer to something like ECR on AWS
2018-09-19 10:48:34 <_ikke_> I guess the idea is to offer a solution for people who want to have their own packages, but don't want to bother with the details
2018-09-19 10:49:03 <AinNero> its not like you can compile something and then wrap it up for both alpine and debian\
2018-09-19 10:49:11 <AinNero> except for statically compiled thing like go or java
2018-09-19 10:49:53 <AinNero> but yeah, this are probably details modern engineers dont want to bother with
2018-09-19 10:50:47 <lskillen> That's not what we see people do - They might be building alpine packages to go into their Docker containers, which are then utilised with Python packages to build their service. They may be distributing their software as vendors as deb/rpm packages to others, etc. They might be an individual who just "wants a repo" quickly, or a large org that has to worry about teams of people, of
2018-09-19 10:50:53 <lskillen> security/provenance/auditing/etc. It's tricky. :)
2018-09-19 10:51:58 <AinNero> Not sure if im conservative or fed up with fundamentally broken software, or the first as result of the latter
2018-09-19 10:52:37 <lskillen> S'ok, I'm liberal and also fed up ;)
2018-09-19 10:52:39 <AinNero> anyways, feel free to ignore my ramblings. I have to work with docker each day.
2018-09-19 10:53:26 <lskillen> I hear you
2018-09-19 11:27:31 <lskillen> brb, odd issue with running irssi inside tmux :)
2018-09-19 12:31:48 <Shiz> lskillen: think i know some people who may be interested in it :p
2018-09-19 12:42:33 <lskillen> Awesome :)
2018-09-19 12:43:18 <lskillen> People have asked us to implement Alpine repositories, so hopefully that means it is useful or they were trolling us. ;)
2018-09-19 12:45:44 <lskillen> Happy to provide a link either here or privately, with permission
2018-09-19 12:46:50 <lskillen> It'd also be great to talk to someone who has access to the wiki to see if it would be appropriate to establish a new section for package manager services that support Alpine as a package format (free or otherwise) :)
2018-09-19 13:21:10 <danieli> llvm 7 released
2018-09-19 13:21:33 <_ikke_> saw it on HN
2018-09-19 13:22:42 <danieli> mm
2018-09-19 13:22:59 <danieli> I saw it on twitter, then hn just after
2018-09-20 07:47:21 <ollieparanoid[m]> hi folks! the armhf, aarch64, ppc64le buildbots are stuck on mariadb for some reason, and the logs can't be accessed on http://build.alpinelinux.org/ -- could someone look into this? (ncopa maybe?)
2018-09-20 07:48:16 <ncopa> in a sec
2018-09-20 07:51:46 <ollieparanoid[m]> thanks!
2018-09-20 11:00:26 <ncopa> i think we need upgrade gcc. i wonder if i should jump directly to gcc8 and skip gcc7
2018-09-20 11:51:33 <danieli> ncopa: we can try
2018-09-20 11:51:43 <danieli> i'll have a look at llvm 7
2018-09-20 12:24:42 <ncopa> we have a patch that enables --as-needed by default
2018-09-20 12:24:47 <ncopa> i wonder if we still need it
2018-09-20 12:25:30 <raph_ael> hi, how much space is needed to set an alpine  mirror and what is the procedure to set an official mirror ? thanks
2018-09-20 12:26:34 <_ikke_> raph_ael: https://wiki.alpinelinux.org/wiki/How_to_setup_a_Alpine_Linux_mirror
2018-09-20 12:26:34 <AinNero> there is an wiki page for that
2018-09-20 12:26:45 <_ikke_> ^
2018-09-20 12:27:14 <raph_ael> _ikke_: thanks I missed this one
2018-09-20 12:27:17 <_ikke_> There is an e-mail address where you can send a request to be listed as an official mirror
2018-09-20 12:28:29 <_ikke_> alpine-mirrors@alpinelinux.org
2018-09-20 12:28:32 <raph_ael> ok, I wanted above all to see if we had space enough to add an alpine mirror
2018-09-20 12:28:51 <_ikke_> The wiki page mentions the size per repo
2018-09-20 12:29:09 <raph_ael> I've just seen thanks :) the page is very informative
2018-09-20 12:29:52 <raph_ael> _ikke_: is it mandatory to have all distrib versions ?
2018-09-20 12:30:06 <_ikke_> no
2018-09-20 12:30:14 <raph_ael> great
2018-09-20 12:30:18 <_ikke_> https://mirrors.alpinelinux.org/
2018-09-20 12:30:39 <_ikke_> Here you see an overview of what other mirrors typically do
2018-09-20 12:31:59 <raph_ael> most of them seem to have everything
2018-09-20 13:12:44 <raph_ael> ok, it's on its way
2018-09-20 13:32:26 <Myhro> Hi. I'm trying to find some information about the official Alpine Vagrant boxes - https://app.vagrantup.com/alpine/boxes/alpine64 - but so far the only thing I found it is this issue - https://bugs.alpinelinux.org/issues/8902 - which complains exactly that there are no docs for them. Does someone know how they are built? Who maintains them? I'd like to send a few fixes/updates for it.
2018-09-20 13:36:30 <danieli> good questin
2018-09-20 13:36:33 <danieli> question*
2018-09-20 17:00:19 <danieli> assigned to wrong person, so they make a new issue?
2018-09-20 17:00:21 <danieli> that's odd
2018-09-20 17:01:34 <_ikke_> thats people
2018-09-20 17:01:59 <_ikke_> someone already marked is as dupe
2018-09-20 17:03:57 <AinNero> odd
2018-09-20 17:06:58 <_ikke_> even
2018-09-21 00:35:41 <Thalheim> so Julia's apparently been of interest; I have 1.0.0 building on Alpine 3.8, and passes most tests (http://paste.debian.net/hidden/a253a865/). Anyone interested in working with me on this?
2018-09-21 00:42:23 <danieli> Thalheim: Shiz?
2018-09-21 00:42:40 <danieli> which tests /don't/ pass?
2018-09-21 00:43:10 <Shiz> yes, interested in working with it on you, but not now :)
2018-09-21 00:43:19 <Shiz> i need to get up in ~2.5h and will be mostly gone for the weekend
2018-09-21 00:43:22 <Shiz> after that def though
2018-09-21 01:38:53 <danieli> no worries, was thinking you'd be interested in cooperating eventually
2018-09-21 04:22:26 <ollieparanoid[m]> ncopa: would there be any advantage in going with gcc7 vs gcc8?
2018-09-21 04:23:15 <ollieparanoid[m]> also, can I help with unblocking build-edge-aarch64 somehow? there are still no logs for why mariadb is failing to build. thanks for fixing it on armhf though!
2018-09-21 06:04:37 <ncopa> ollieparanoid[m]: i think it hangs on some of the tests. the only thing i did with armhf was to reboot it
2018-09-21 06:07:19 <ncopa> 42124 buildoze  0:00 ctest -E test-connect
2018-09-21 06:07:19 <ncopa> 42444 buildoze  4:38 /home/buildozer/aports/main/mariadb/src/mariadb-10.3.9/unittest/sql/my_apc-t
2018-09-21 06:07:35 <ncopa> that is what hangs
2018-09-21 06:48:41 <Thalheim> danieli: Shiz: no rush; take your time.
2018-09-21 06:54:37 <ollieparanoid[m]> ncopa: I guess you also need to reboot it then
2018-09-21 06:54:49 <ollieparanoid[m]> I've just compiled mariadb for aarch64 with qemu and that ran through
2018-09-21 06:56:43 <ncopa> i ran it in a container here too
2018-09-21 06:57:08 <ncopa> i think its a race or similar that makes it lockup on random
2018-09-21 06:57:13 <ollieparanoid[m]> did it work?
2018-09-21 06:57:37 <ncopa> yes, it passed
2018-09-21 06:57:42 <ncopa> so i dont know why it happens
2018-09-21 06:57:56 <ollieparanoid[m]> yeah, it sounds like a race condition :\ I hate those
2018-09-21 06:57:56 <ncopa> i did reboot the aarch64 builder yesterday
2018-09-21 06:58:32 <ncopa> its probably somethign that is unlikely to happen, but may occour if you have 96 cores...
2018-09-21 06:58:41 <ollieparanoid[m]> what are the next steps then? bug report and rebooting again, hoping it goes through?
2018-09-21 06:59:44 <ncopa> im gonna try disable running tests in fakeroot
2018-09-21 07:00:06 <ollieparanoid[m]> ok
2018-09-21 07:00:26 <ncopa> re gcc7 vs gcc8
2018-09-21 07:00:38 <ncopa> the only benefit with gcc7 would be that its more mature
2018-09-21 07:00:52 <ncopa> more software is patched and fixed to work for it
2018-09-21 07:00:56 <ncopa> work with it
2018-09-21 07:01:07 <ncopa> but i dont think gcc8 will be any problem
2018-09-21 07:01:09 <ollieparanoid[m]> that makes sense
2018-09-21 07:01:18 <ncopa> so im aiming for gcc8 and see how it goes
2018-09-21 07:01:58 <ollieparanoid[m]> okay. postmarketOS is still based on Alpine edge (I have it high up on my list to base it on Alpine stable) so we'll have fun with the gcc update either way I guess
2018-09-21 07:02:17 <ollieparanoid[m]> maybe we could coordinate the time a bit
2018-09-21 07:03:16 <ncopa> we need to do it asap
2018-09-21 07:03:25 <ncopa> hopefully next week
2018-09-21 07:03:46 <ollieparanoid[m]> okay. as long as you don't do it on a friday, that would be good already :p
2018-09-21 07:04:06 <ollieparanoid[m]> also what's the reason again for why the gcc update is necessary, is there a program that doesn't compile with gcc6?
2018-09-21 07:49:27 <ncopa> yes, I have bumped into a couple apps that require c++17
2018-09-21 07:49:42 <ncopa> i think ceph was one that i cannot upgrade til gcc is updated
2018-09-21 07:51:49 <clandmeter> ollieparanoid[m], you build aarch64 on qemu?
2018-09-21 07:54:03 <ollieparanoid[m]> clandmeter: with "pmbootstrap build --arch=aarch64" :p (which is qemu + distcc + cross compiler running in native chroot, so it's not super slow)
2018-09-21 07:55:26 <ollieparanoid[m]> ncopa: in case you didn't see this, mariadb failed now with one test
2018-09-21 07:55:54 <ollieparanoid[m]> 
ACTION needs to go though, thanks for looking into this

2018-09-21 08:09:08 <Myhro> Is it OK to fix a package and add myself as a contributor? Or should I ask the maintainer before doing that?
2018-09-21 08:12:06 <ncopa> ollieparanoid[m]: it was me who killed it
2018-09-21 08:12:25 <ncopa> Myhro: its ok to fix a package and add yourself as contributor
2018-09-21 08:12:49 <ncopa> its also ok to fix a package and not add yourself as contributor. your name will be in commit log
2018-09-21 08:13:39 <Myhro> ncopa: Great. Thanks.
2018-09-21 08:18:08 <ncopa> ollieparanoid[m]: #9441
2018-09-21 09:08:32 <Myhro> Done: https://github.com/alpinelinux/aports/pull/5213 - that's my first PR, so please let me know if I did something dumb.
2018-09-21 09:16:29 <AinNero> Myhro: reusing the pkgname isnt strictly necessary, but thats only some minor nitpick
2018-09-21 09:17:07 <_ikke_> Myhro: I noticed that py-setuptools sometimes also is a runtime dependency. Did you check that?
2018-09-21 09:17:56 <Myhro> AinNero: you mean avoiding the $pkgname variable usage?
2018-09-21 09:18:08 <_ikke_> only where it makes sense
2018-09-21 09:18:12 <Myhro> _ikke_: yes, I installed using pip and it only installed attrs and six as dependencies.
2018-09-21 09:18:27 <_ikke_> ok
2018-09-21 09:18:54 <_ikke_> Myhro: Imagine someone creating an alternative version of the same package, so a different packagename
2018-09-21 09:19:13 <_ikke_> Myhro: Then ask yourself where it makes sense that $pkgname is used, and where doesn't it make sense
2018-09-21 09:19:35 <_ikke_> source urls don't change in this case, so that would be broken
2018-09-21 09:19:44 <_ikke_> (I know it's done a lot in packages)
2018-09-21 09:20:54 <Myhro> Hummm. Now I got what you meant. I followed the example of bigger packages like https://github.com/alpinelinux/aports/blob/master/main/py-flask/APKBUILD and https://github.com/alpinelinux/aports/blob/master/main/py-django/APKBUILD - they do the same for the source URL.
2018-09-21 09:21:19 <Myhro> And actually I found it quite clever.
2018-09-21 09:21:55 <Myhro> But, it can be indeed a bad practice. I'll update it
2018-09-21 09:25:30 <Myhro> Actually, it doesn't fit this use case: the URLs are composed of "$_pkgname" (the PyPI one), not the actual "$pkgname" (for Alpine).
2018-09-21 09:25:50 <_ikke_> projectname != packagename in everycase, but it's often conflated
2018-09-21 09:28:17 <Myhro> Ok, now I'm a little bit confused. You mean the source url would be broken if there's an alternatice Alpine or PyPI package with a different name?
2018-09-21 09:33:04 <_ikke_> Say you want to make a package with a certain thing enabled (which isn't by default)
2018-09-21 09:33:14 <_ikke_> you could copy the APKBUILD, change the pkgname
2018-09-21 09:33:34 <_ikke_> but the source url shouldn't change in that case
2018-09-21 09:44:01 <Myhro> Yes, but $pkgname (Alpine: py-automat) and $_pkgname (PyPI: Automat) have different values. If the first one changes, this won't impact in the source url whatsoever (as it is solely based on the PyPi name).
2018-09-21 09:48:27 <_ikke_> Right, I notice
2018-09-21 09:51:58 <Myhro> I realise it is confusing. Maybe it's better if $_pkgname is changed to something like $pypiname ?
2018-09-21 09:52:23 <_ikke_> No, is not necessary
2018-09-21 09:52:36 <Myhro> Ok.
2018-09-21 11:22:16 <ScrumpyJack> how often do we run secfixes.lua?
2018-09-21 11:33:26 <vkrishn> rnalrd: could you pls run checksum for main/ulogd/musl-fixes.patch, seems a mismatch
2018-09-21 11:42:22 <rnalrd> vkrishn, edge?
2018-09-21 11:42:57 <rnalrd> yep
2018-09-21 11:52:16 <vkrishn> in v3.8.1 also
2018-09-21 11:52:20 <rnalrd> yeah
2018-09-21 11:52:33 <_ikke_> did upstream change the source?
2018-09-21 11:53:40 <rnalrd> nope, a "git commit -a" inadvertently included a change in the patch in an unrelated commit, it was my mistake
2018-09-21 11:53:53 <rnalrd> ulogd did not build at all
2018-09-21 11:54:06 <_ikke_> ah ok
2018-09-21 11:55:50 <rnalrd> 3.8-stable fixed as well
2018-09-21 11:59:40 <vkrishn> maybe upgrade it too, ulogd-2.0.7
2018-09-21 11:59:57 <vkrishn> https://git.netfilter.org/ulogd2/
2018-09-21 12:31:26 <Thalheim> danieli: the tests that don't pass are 'cmdlineargs' and parts of 'libgit2', and I'm not sure what "broken" in the above trace means, but 'ranges' is apparently quite broken
2018-09-22 10:44:26 <_ikke_> Ganwell: I just ran into the certbot issue as well
2018-09-22 13:11:03 <AinNero> hm, i discovered a puzzle piece that is missing on alpine
2018-09-22 13:11:24 <AinNero> either an getpeername command or some environment variable for nc =-l
2018-09-22 13:11:31 <AinNero> * nc -l
2018-09-22 13:12:28 <AinNero> busybox has an udpsvd and tcpsvd that does set an environment variable, but aside that the functionality is not different from nc
2018-09-23 15:49:42 <Thalheim> really? https://i.imgur.com/Fc87M2I.png
2018-09-23 16:00:02 <terra> Thalheim: afaik, openrc would deal with pidfile if initscript has definition of pidfile="/run/$SVCNAME.pid"
2018-09-23 16:00:57 <terra> it makes senses that busybox hasn't configured with pidfile feature.
2018-09-23 16:06:29 <Thalheim> terra: the mail message headers, not the body
2018-09-23 16:14:39 <_ikke_> terra: but that means that busybox will need to background the process as well
2018-09-23 16:15:07 <_ikke_> lol
2018-09-23 16:21:30 <terra> :\
2018-09-23 16:50:54 <_ikke_> did someone accidentally use the wrong e-mail adres? :D
2018-09-23 16:56:47 <danieli> context? @ _ikke_
2018-09-23 16:57:03 <Thalheim> danieli: that email I linked above...
2018-09-23 16:57:07 <awilfox> danieli: Thalheim's link
2018-09-24 05:32:12 <ollieparanoid[m]> ncopa: good morning! would be nice if you could post the progress of upgrading to gcc8 here once you're starting with it
2018-09-24 06:41:18 <ncopa> morning
2018-09-24 06:41:51 <_ikke_> good morning
2018-09-24 07:01:39 <ncopa> ollieparanoid[m]: im gonna work on gcc8 today, after that i'll try make a separate gcc6 package so we can build openjdk7
2018-09-24 07:02:09 <ncopa> then i'll try update som package with lots of deps to rebuild, for example ffmpeg 4
2018-09-24 07:02:32 <ncopa> and then i'll look at switch libressl -> openssl
2018-09-24 07:04:07 <ollieparanoid[m]> sounds good. I guess adding gcc8 as new "gcc8" package would make it more complicated than keeping the old gcc as "gcc6" around?
2018-09-24 07:12:05 <ollieparanoid[m]> ncopa: anyway, I'm heading off to my day job. But I'll be online afterwards and see how it goes. good luck!
2018-09-24 07:23:42 <ncopa> yes, gcc will be upgraded to gcc 8
2018-09-24 07:23:55 <ncopa> and i will try add gcc6 as separate package
2018-09-24 08:54:09 <PureTryOut[m]> oh exciting changes. thanks for your efforts ncopa !
2018-09-24 09:58:19 <nmeum> regarding basesystem upgrades I am also planning to finish the busybox 1.29 upgrade this week
2018-09-24 10:34:22 <ncopa> nmeum: that would be great
2018-09-24 13:31:43 <ncopa> ollieparanoid[m]: we are adding armv7 and may want discontinue with armv6 (armhf) in the future
2018-09-24 13:32:25 <ncopa> currently they both run on same hardware, which makes it slower than necessary
2018-09-24 13:33:27 <ncopa> we do have a spare x-gene board, but nowere to run it and nobody to set it up
2018-09-24 13:41:56 <nmeum> ncopa: the busybox upgrade is pretty much done. I will continue testing it on my laptop the following days. Additionally, it would be nice if someone could review the changes I made. See: https://github.com/alpinelinux/aports/pull/4938#issuecomment-423976104
2018-09-24 13:42:46 <ncopa> nmeum: nice!
2018-09-24 14:10:58 <AinNero> nmeum: built it, installed, my machine still boots, so it cant be that bad
2018-09-24 14:12:41 <nmeum> that test already passed on my machine as well :) let me know if you find any additional problems
2018-09-24 14:23:59 <TBB> silly question about apk: could apk cache sync be used for unattended downloading of updates on the background?
2018-09-24 14:33:21 <TBB> I mean, if I update the indices and then issue apk cache sync, does the sync act on the updated indices or the list of installed packages?
2018-09-24 14:34:31 <TBB> if it works on the former, then sync would work as a "download but not install yet" mechanism, and it could be completed in the course of several sessions in case the system has a really slow link
2018-09-24 14:43:57 <ncopa> not sure
2018-09-24 14:46:39 <TBB> and of course I managed to delay asking just enough to miss fabled :)
2018-09-24 19:01:48 <AinNero> are the md5sums and sha256sums deprecated, as in, should i remove them when touching a package?
2018-09-24 19:02:13 <AinNero> abuild checksum currently does it
2018-09-24 19:03:14 <_ikke_> if abuild checksum removes them, I guess you can assume they are deprecated
2018-09-24 19:03:20 <AinNero> alright
2018-09-24 20:54:07 <AinNero> oh crap, 250 PR's again
2018-09-24 20:55:17 <_ikke_> I do contribute to that as well
2018-09-24 20:57:12 <AinNero> we could auto-close PR's that have no activity for 3 months or more
2018-09-24 20:58:00 <_ikke_> yeah, that long-tail really adds upp
2018-09-24 20:58:27 <_ikke_> some PRs from 2016 are still open
2018-09-24 20:58:35 <_ikke_> wondering how relevant they still are
2018-09-24 21:41:43 <clandmeter> AinNero, there is an auto close bot, but i think it doesnt really do what it suppose to do.
2018-09-24 21:41:51 <clandmeter> not sure how to debug it.
2018-09-24 21:42:33 <st1cky> hello
2018-09-24 21:43:10 <st1cky> any reason on why alpine's busybox doesn't have pidfiles enabled in the config?
2018-09-24 21:43:47 <clandmeter> busybox pidfiles?
2018-09-24 21:44:36 <st1cky> I think it was CONFIG_FEATURE_PIDFILE
2018-09-24 21:44:40 <st1cky> let me look up
2018-09-24 21:44:48 <clandmeter> ah you are the one with the weird email address on the ML?
2018-09-24 21:44:54 <st1cky> yes
2018-09-24 21:45:12 <clandmeter> i dont think that really appreciated around here
2018-09-24 21:45:50 <st1cky> the mail or the pidfile thing?
2018-09-24 21:46:01 <clandmeter> your email address
2018-09-24 21:46:52 <st1cky> *email address
2018-09-24 21:46:54 <st1cky> sry, laggy connection
2018-09-24 21:48:16 <st1cky> well, regarding the mail content (I'll look to do smth about the email address), any reason why it's disabled, or how init.d scripts could be modified?
2018-09-24 21:48:39 <st1cky> current state renders them unusable, as you lose the ability to reload daemons
2018-09-24 21:49:13 <st1cky> it can be a "no problem" in cases like inetd, where you can pkill -HUP it to reread the config file
2018-09-24 21:49:44 <st1cky> but haven't checked yet if, for example, dnsd or httpd provide a similar mechanism
2018-09-24 21:50:03 <st1cky> let alone the fact of stopping the process from openrc
2018-09-24 21:51:26 <clandmeter> openrc can generate a pidfile for you. is that not enough?
2018-09-24 21:52:30 <st1cky> I need to check that. If so, would a patch for busybox-initscripts be welcome?
2018-09-24 21:54:28 <clandmeter> what do you want to modify in bb-initscripts?
2018-09-24 21:54:44 <AinNero> oh, which email address?
2018-09-24 21:55:17 <AinNero> st1cky: uhm
2018-09-24 21:55:23 <clandmeter> http://lists.alpinelinux.org/alpine-devel/6288.html
2018-09-24 21:55:29 <AinNero> st1cky: https://github.com/alpinelinux/aports/pull/5226
2018-09-24 21:55:51 <AinNero> oh, cockmail :3 met him in person last year
2018-09-24 21:56:07 <AinNero> got some horsefucker.org address myself, but i dont use it for professional setting
2018-09-24 21:57:37 <st1cky> AinNero: cool, that might work
2018-09-24 21:58:02 <AinNero> i teested all of the initscripts except watchdog and some other where i lacked an prerequisite
2018-09-24 21:58:20 <AinNero> * of busybox-initscripts
2018-09-24 21:59:39 <st1cky> well, with that PR, my work (?) here is done
2018-09-24 21:59:41 <st1cky> bye
2018-09-24 22:00:15 <AinNero> mh. not sure if this is good or bad
2018-09-25 05:53:56 <ollieparanoid[m]> PureTryOut is right, a lot of messages did not appear here. Here it looked like _ikke_ meant contribution to the armv7 efforts, not to creating a big backlog of PRs :p
2018-09-25 05:54:09 <ollieparanoid[m]> PureTryOut: see http://dev.alpinelinux.org/irclogs/%23alpine-devel-2018-09.log
2018-09-25 06:10:30 <_ikke_> The irc-matrix bridge doesn't seem stable
2018-09-25 06:11:44 <ollieparanoid[m]> yeah not really :\
2018-09-25 09:39:44 <PureTryOut[m]> that's because it isn't stable. it has had this issue for ages, still no permanent fix. also, being written in NodeJS, what else did we expect 😢
2018-09-25 12:02:05 <fabled> postfix is broke?
2018-09-25 12:02:17 <fabled> mail.crit postfix[14472]: fatal: /usr/lib/postfix/postfix-script: No such file or directory
2018-09-25 12:02:18 <_ikke_> when / where/
2018-09-25 12:02:20 <fabled> it's in libexec
2018-09-25 12:02:23 <fabled> edge
2018-09-25 12:02:25 <_ikke_> ah ok
2018-09-25 12:02:54 <fabled> ncopa, i think it's moving stuff to libexec
2018-09-25 12:03:45 <fabled> oh
2018-09-25 12:03:53 <fabled> it needs daemon_directory change in master.cf
2018-09-25 12:46:49 <ncopa> there was an issue about moving it to libexec
2018-09-25 12:59:26 <AinNero> 9471 is not a bug, right?
2018-09-25 13:05:01 <AinNero> that the guy is doing sysctl from within the containers seems like they are doing something wrong
2018-09-25 13:06:52 <_ikke_> and containers don't boot anyway
2018-09-25 13:07:06 <_ikke_> (personal gripe :) )
2018-09-25 13:07:30 <AinNero> also correct...
2018-09-25 16:48:49 <ncopa> $ gcc --version
2018-09-25 16:48:49 <ncopa> gcc (Alpine 8.2.0) 8.2.0
2018-09-25 16:59:37 <AinNero> .a is for static linking, which is not supported?
2018-09-25 16:59:48 <AinNero> yeah, it is
2018-09-25 17:08:29 <AinNero> policy is that we dont support static linking against alpine libs?
2018-09-25 17:09:39 <_ikke_> AinNero: https://pkgs.alpinelinux.org/contents?file=*.a&path=&name=&branch=edge&arch=x86_64
2018-09-25 17:09:55 <AinNero> i guess no.
2018-09-25 17:10:46 <AinNero> so should static linking support added to pciutils?
2018-09-25 17:27:24 <mps> AinNero: I think it should
2018-09-25 17:27:53 <mps> if the -dev package is installed, of course
2018-09-25 17:38:18 <ncopa> ollieparanoid[m]: i will most likely push gcc8 tomorrow morning
2018-09-25 17:50:54 <geekodour08> Hello I am new to alpine linux and absolute beginner for alpine linux development, If this is a low priority issue then I'd like to work on it but looking at the description it seems will need changes in the libpci-dev package?
2018-09-25 17:51:01 <geekodour08> https://bugs.alpinelinux.org/issues/9472
2018-09-25 18:01:21 <_ikke_> geekodour08: You need to find out why that file is not generated during build process
2018-09-25 18:07:50 <awilfox> what is the proper way to drop maintainership of packages?
2018-09-25 18:13:00 <_ikke_> I guess it's either finding someone else to maintain it, or move it to unmaintained, but depending on what package, that might not be that trivial
2018-09-25 19:21:08 <clandmeter> awilfox, which pkgs
2018-09-25 19:21:32 <TBB> ncopa: regarding that question I had some time ago about apk cache sync, I've initially tested the patch that was committed earlier today to github (extra switches to apk cache) and it seems to work great
2018-09-25 19:24:08 <awilfox> clandmeter: libpipeline and man-db can probably be moved to unmaintained, ipmiutil was already blank Maintainer: before I took it, I don't actually remember signing up to maintain twm but it appears to have happened anyway, etc...
2018-09-25 19:24:20 <awilfox> https://pkgs.alpinelinux.org/packages?name=&branch=edge&arch=ppc64le&maintainer=A.+Wilcox
2018-09-25 19:24:29 <awilfox> I don't have the time nor the resources to adequately test bumps
2018-09-25 19:24:37 <awilfox> and I don't want to contribute to the problem of "push stuff and don't even test it first"
2018-09-25 19:25:41 <awilfox> I literally do not own any x86_64 systems any more, beyond one that is used for building adelie packages.  my ppc64 is BE, so I'd need to get a LE VM going on it, which is possible, but it won't catch x86_64 bugs.
2018-09-25 19:25:52 <awilfox> and my experience in adelie is that there are a lot more x86_64 bugs than there are others.
2018-09-25 19:34:28 <clandmeter> awilfox, can you submit a PR with those pkgs and remove yourself? we can see if one of us can take over.
2018-09-25 19:34:28 <Michcioperz> e
2018-09-25 19:34:56 <awilfox> clandmeter: so just make Maintainer: line empty?  I can do that some time this week
2018-09-25 19:35:02 <awilfox> if you don't see it by friday, poke me
2018-09-25 19:35:11 <awilfox> neck-deep in firefox code right now
2018-09-25 19:40:40 <clandmeter> awilfox, ok
2018-09-25 20:04:34 <AinNero> could we open the PR section so that everyone can review pr's ?
2018-09-25 20:04:48 <AinNero> to encourage users to test the changes of other users
2018-09-25 20:07:22 <_ikke_> Everyone *can* review PRs
2018-09-25 20:08:27 <AinNero> *sighs*
2018-09-25 20:08:34 <AinNero> did they move the button for it?
2018-09-25 20:08:40 <AinNero> at least i cant find it anymore..
2018-09-25 20:08:46 <_ikke_> What button?
2018-09-25 20:09:04 <AinNero> ah, in files changed..
2018-09-25 20:09:08 <_ikke_> yes
2018-09-25 20:58:14 <AinNero> whats the difference between community and main?
2018-09-25 20:59:14 <AinNero> im currently testing hiredis and i would have expected it in testing or community seems it seems rather rarely used and fresh
2018-09-25 20:59:18 <_ikke_> support duration
2018-09-25 20:59:31 <AinNero> mh
2018-09-25 21:00:03 <_ikke_> the packaged versions in main should at least get fixes for 2 years I believe
2018-09-25 21:00:17 <AinNero> yeah, i figured that from your short answer
2018-09-25 21:01:05 <_ikke_> But that split is fairly recent, and sometimes due to dependencies more packages are in main than you would expect based on that rule
2018-09-25 21:21:41 <PureTryOut[m]> the GCC8 upgrade will be interesting for postmarketOS. most kernels will probably not compile anymore. Please don't let that hold you back though, we need this upgrade for other stuff 😛
2018-09-25 21:34:33 <AinNero> openrc scripts should be in -openrc subpackages now?
2018-09-25 21:38:23 <PureTryOut[m]> Could any Matrix admin set the required permission level to talk back to 0 from 1? It's preventing most Matrix users to talk
2018-09-25 21:41:56 <AinNero> also, im not sure what to do with community/network-manager-applet
2018-09-25 21:42:01 <AinNero> it depends on stuff in testing
2018-09-25 21:42:11 <AinNero> is that valid?
2018-09-25 21:52:10 <AinNero> alright, i just saw the submitter took care of this.
2018-09-26 05:29:18 <ollieparanoid[m]> ncopa: thanks for the heads up!
2018-09-26 05:32:55 <ollieparanoid[m]> also +1 for the matrix permissions
2018-09-26 12:56:29 <ncopa> ok, here comes gcc8
2018-09-26 19:58:57 <AinNero> ncopa: do you have specific criteria for merging PRs?
2018-09-26 20:01:01 <ncopa> the PR needs to be good :)
2018-09-26 20:01:01 <ncopa> whats "good" is a longer story I suppose
2018-09-26 20:02:10 <ncopa> for things like: https://github.com/alpinelinux/aports/pull/5181
2018-09-26 20:03:00 <ncopa> xpecex is a relatively new contributor, and I don't trust that he has verified that the check does not break anything
2018-09-26 20:03:05 <ncopa> eg breaks ABI
2018-09-26 20:03:11 <ncopa> so that is one thing i check
2018-09-26 20:03:21 <ncopa> I use checkapk to help me with that
2018-09-26 20:03:55 <AinNero> not sure if its useful if i build and test stuff myself and do an review
2018-09-26 20:04:05 <AinNero> if you need to do the same anywats
2018-09-26 20:04:09 <AinNero> *anyways
2018-09-26 20:05:06 <ncopa> not always
2018-09-26 20:05:29 <ncopa> if i see "nero" give a thumbs up, i may end up just push if the commit looks good
2018-09-26 20:05:39 <ncopa> the hiredis PR
2018-09-26 20:06:01 <ncopa> was not good
2018-09-26 20:06:15 <ncopa> he did a sed to fix the .pc after the build
2018-09-26 20:06:31 <ncopa> you normally can set the prefix from configure script or via variable
2018-09-26 20:06:38 <ncopa> so i checked what fedora does
2018-09-26 20:07:05 <ncopa> https://src.fedoraproject.org/cgit/rpms/hiredis.git/tree/hiredis.spec#n28
2018-09-26 20:07:12 <ncopa> they set PREFIX at build time
2018-09-26 20:07:28 <mepholic> god specfiles are gross
2018-09-26 20:08:17 <AinNero> ah i see, merged but additional fixup commit
2018-09-26 20:08:22 <ncopa> AinNero: i fixed hiredis with c124a200753098a4b17500af5c0fcb80b0bcc169
2018-09-26 20:08:27 <AinNero> i noticed
2018-09-26 20:08:41 <ncopa> yeah, its somewhat non-optimal doing that
2018-09-26 20:08:55 <ncopa> its normally better to give feedback and let the contributor do the fix
2018-09-26 20:09:04 <ncopa> but that is timeconsuming
2018-09-26 20:09:14 <ncopa> its better because then we educate them
2018-09-26 20:10:18 <ncopa> jirutka was very good at that, (but could be better at give constructive/positive feedback)
2018-09-26 20:10:43 <ncopa> i think what would been even better, if we can point users to some documentation
2018-09-26 20:10:57 <ncopa> for example the hiredis sed thing
2018-09-26 20:11:05 <AinNero> im often at loss when searching for policies on certain stuff
2018-09-26 20:11:09 <AinNero> like the sed thing
2018-09-26 20:11:26 <ncopa> i should have written a doc that we prefer patches over sed
2018-09-26 20:11:32 <ncopa> and the link to that doc
2018-09-26 20:11:57 <ncopa> i think part of the problem is that we have no clear policy where those docs should be :)
2018-09-26 20:12:07 <ncopa> we currently have lots in wiki
2018-09-26 20:12:25 <ncopa> but wiki is better for community docs than for official docs
2018-09-26 20:13:12 <AinNero> toplevel file in aports? like README ?
2018-09-26 20:13:17 <ncopa> there is a .github/CONTRIBUTING.md doc which is pretty nice
2018-09-26 20:13:19 <ncopa> yes!
2018-09-26 20:13:31 <ncopa> toplevel README is exactly what we need
2018-09-26 20:13:35 <ncopa> but it shoudl be short
2018-09-26 20:13:48 <ncopa> give the big overview
2018-09-26 20:13:57 <ncopa> and point to details in other docs
2018-09-26 20:14:08 <ncopa> i'm thinking of things like docs/CONTRIBUTING.md
2018-09-26 20:14:14 <ncopa> for general guidelintes
2018-09-26 20:14:28 <ncopa> maybe have specific docs for github PRs
2018-09-26 20:14:43 <ncopa> sending patches via git send-email
2018-09-26 20:14:52 <_ikke_> I think AinNero is in the perfect spot atm to document those kinds of things
2018-09-26 20:14:59 <AinNero> weeee
2018-09-26 20:15:03 <ncopa> :D
2018-09-26 20:15:44 <ncopa> i think we should have a big-plan first, a plan how we want organize the docs
2018-09-26 20:16:05 <ncopa> i think contribute docs should be kept directly in aports tree
2018-09-26 20:16:54 <ncopa> im really not godo at writing docs :-(
2018-09-26 20:16:55 <AinNero> _ikke_: but i need to know the information first, which means i'll be bothering core people for weeks
2018-09-26 20:17:09 <AinNero> ncopa: IMHO my grammar is awful sometimes
2018-09-26 20:17:20 <ncopa> thats no problem, really
2018-09-26 20:17:29 <_ikke_> We can review that :P
2018-09-26 20:17:29 <ncopa> we can have grammar experts proofread
2018-09-26 20:18:00 <ncopa> what we need is a documenting genious, who has a good idea how to organize the docs
2018-09-26 20:18:00 <AinNero> i was more referring to the manpages in mkinitfs
2018-09-26 20:18:08 <_ikke_> AinNero: Problem is that the people that know about it don't know what to document
2018-09-26 20:18:57 <ncopa> manpages are for reference, but does not give the big-picture view
2018-09-26 20:19:41 <_ikke_> I kinda like sphinx + read-the-docs skin
2018-09-26 20:20:54 <AinNero> im quite into manpages
2018-09-26 20:21:08 <AinNero> i feel more comfortable in the terminal than in the browser
2018-09-26 20:21:16 <ncopa> good
2018-09-26 20:21:33 <ncopa> i think we have a PR for manpages to abuild
2018-09-26 20:21:37 <_ikke_> haha :D
2018-09-26 20:22:27 <ncopa> ok, i need to sleep now
2018-09-26 20:22:32 <ncopa> good night
2018-09-26 20:22:48 <AinNero> gn8
2018-09-26 20:33:20 <_ikke_> This is an example of sphinx + readthedocs theme: https://cryptography.io/en/latest
2018-09-26 20:36:27 <AinNero> this invokes negative memories of docs of hipster projects
2018-09-26 20:40:23 <_ikke_> :-(
2018-09-26 20:40:53 <AinNero> im working with python crapware at work...
2018-09-27 06:48:48 <Bodenhaltung> Troll? https://nopaste.xyz/?0f508748d41338f1#NZFOeUxfgSKJJNw8GXrjdSd5ojBAXuobmE3qvxes/qo=
2018-09-27 06:52:31 <mepholic> Bodenhaltung: what happens if you link it?
2018-09-27 06:52:35 <mepholic> lol
2018-09-27 06:52:53 <Bodenhaltung> Maybe fire? :D
2018-09-27 06:53:26 <mepholic> also, does the config specify "/etc/powerdns/pdns.d"
2018-09-27 06:53:29 <mepholic> somewhere?
2018-09-27 06:56:15 <Bodenhaltung> mepholic: omg, ah, lol. Yes.
2018-09-27 06:56:53 <mepholic> occam's razor
2018-09-27 06:56:55 <mepholic> :^)
2018-09-27 06:57:10 <Bodenhaltung> Indeed. :D
2018-09-27 07:02:27 <Bodenhaltung> Hmm, if i want a guardian process, i have to change the init script?
2018-09-27 07:03:35 <Bodenhaltung> Ah, i c, https://bugs.alpinelinux.org/issues/7885
2018-09-27 07:03:40 <mepholic> > When launched with guardian=yes, pdns_server wraps itself inside a ‘guardian’. This guardian monitors the performance of the inner pdns_server instance which shows up in the process list of your OS as pdns_server-instance
2018-09-27 07:04:20 <mepholic> Bodenhaltung: is that path crap in the default alpine pdns package?
2018-09-27 07:04:47 <Bodenhaltung> mepholic: Nope, my crap config from another server.
2018-09-27 07:05:00 <mepholic> ah lmao
2018-09-27 07:05:02 <mepholic> ok
2018-09-27 07:05:12 <Bodenhaltung> ^_^
2018-09-27 10:29:34 <ncopa> i just bushed busybox update
2018-09-27 10:29:41 <ncopa> watch out for breakages
2018-09-27 11:35:30 <vitronic> Hello everyone, I have more than 40 packages for alpine, someone could tell me what steps to follow to send them.
2018-09-27 11:37:16 <ncopa> hi vitronic, this may be of help: https://github.com/alpinelinux/aports/blob/master/.github/CONTRIBUTING.md
2018-09-27 11:38:09 <ncopa> please note that new packages should got to testing initially
2018-09-27 11:38:47 <vitronic> ncopa: Greetings, thank you, I've had them for a long time, it's time I sent them, they are all tested since version 3.6
2018-09-27 11:38:51 <vitronic> fine
2018-09-27 11:39:07 <ncopa> can you test them in edge?
2018-09-27 11:39:13 <ncopa> in git master
2018-09-27 11:39:56 <vitronic> ncopa: only stable
2018-09-27 11:41:38 <vitronic> that is to say all are tested in stable, not use edge
2018-09-27 13:56:46 <ScrumpyJack> where is cgi-bin/dl.cgi ? can't find it in alpine-mksite
2018-09-27 14:24:45 <ncopa> i think its on wiki.a.o iirc
2018-09-27 19:03:17 <ollieparanoid[m]> ncopa: I am not sure if we'll need it. but is a "gcc6" package still planned?
2018-09-27 19:03:25 <ollieparanoid[m]> ...and how long will it be around?
2018-09-27 19:33:00 <ollieparanoid[m]> ncopa: after researching, it seems we will also need a gcc6 package around for the old android kernels. at least one where it was tested won't boot after being compiled with gcc8.
2018-09-27 19:33:25 <ollieparanoid[m]> I'm interested in how you would tell abuild to use gcc6 for the packages where it is needed. modify the PATH maybe?
2018-09-27 19:34:10 <_ikke_> ollieparanoid[m]: Perhaps by defined CC in /etc/abuild.conf?
2018-09-27 19:34:19 <_ikke_> s/defined/defining/
2018-09-27 19:34:35 <_ikke_> or in the APKBUILD I guess
2018-09-27 19:34:48 <_ikke_> make CC=/path/to/gcc6
2018-09-27 19:35:16 <ollieparanoid[m]> _ikke_: but that would be set globally. yeah, in the APKBUILD would make more sense
2018-09-27 21:27:51 <Bodenhaltung> Hmm, where i find "galera_new_cluster"?
2018-09-27 22:03:07 <duncan^> Isn't tha related to mariadb?
2018-09-27 22:11:34 <Bodenhaltung> Yeah, i think i have to add "--wsrep-new-cluster" to the start init script for the first time
2018-09-28 06:11:25 <ncopa> ollieparanoid[m]: gcc6 is only for bootstrapping openjdk7, so we will likely keep it for a while
2018-09-28 06:19:47 <ollieparanoid[m]> ncopa: good morning!
2018-09-28 06:21:53 <ollieparanoid[m]> I see. do you know when you will add it? I'm adding a (rather hackish) gcc6 aport right now in postmarketOS based on the latest version of alpine. When Alpine also ships gcc6, we might switch to that version instead.
2018-09-28 06:23:24 <ollieparanoid[m]> I've used /usr/gcc6 as prefix and !tracedeps, so it libgcc doesn't conflict with the existing gcc8. also I've disabled bootstrap and c++ and all other languages and libraries except for C, because that should be enough for compiling kernels, and now the gcc package compiles in 10 minutes instead of 30 on my laptop
2018-09-28 06:26:10 <Bodenhaltung> Can anyone guide me to setup a mariadb galera cluster? The mariadb package does not have wsrep_provider galera binary?
2018-09-28 06:28:25 <Bodenhaltung> Hmm, https://bugs.alpinelinux.org/issues/4646 .... okay?
2018-09-28 07:50:59 <Bodenhaltung> No one here with a Galera Cluster? oO :D *spam*
2018-09-28 08:50:11 <Bodenhaltung> Checking for C++ header file boost/shared_ptr.hpp... no ..... no suitable shared_ptr header found
2018-09-28 08:50:15 <Bodenhaltung> Hmm, how can i solve this?
2018-09-28 09:24:20 <Bodenhaltung> I try to build galera "plugin", but it failed: https://pastebin.com/CPfLQgc3
2018-09-28 09:32:51 <Bodenhaltung> Ah, boost-dev solved this.
2018-09-28 10:55:13 <_ikke_> Something went wrong when building py-pycryptodome. The buidlog says it was already up-to-date (http://build.alpinelinux.org/buildlogs/build-edge-x86_64/testing/py-pycryptodome/py-pycryptodome-3.6.6-r0.log) and the subpackages don't seem to have been built
2018-09-28 11:15:36 <AinNero> could someone neutralize PR #5251 please?
2018-09-28 11:16:07 <AinNero> otherwise it will be generating notifications for the next two years
2018-09-28 11:16:14 <AinNero> or everyone needs to manually unsubscribe
2018-09-28 11:16:44 <AinNero> https://github.com/alpinelinux/aports/pull/5251
2018-09-28 11:17:10 <AinNero> why does github even allow PR's source branches that people have no access to
2018-09-28 11:24:08 <Shiz> ncopa: are we vulnerable to https://www.reddit.com/r/sysadmin/comments/9jfckm/redhat_kernel_zeroday_cve201814634/ ?
2018-09-28 11:35:56 <AinNero> lol, how do people even find these bugs
2018-09-28 12:11:26 <ncopa> Shiz: looks like 4.14.y has da029c11e6b12f321f36dac8771e833b65cec962 exec: Limit arg stack to at most 75% of _STK_LIM
2018-09-28 12:12:11 <ncopa> so v3.8 is not vulnerable
2018-09-28 12:13:54 <ncopa> same with linux-4.9.y
2018-09-28 12:14:29 <ncopa> since 4.9.39
2018-09-28 12:16:59 <ncopa> 4.4.77
2018-09-28 12:17:11 <ncopa> sorry. 4.4.78 has the fix too
2018-09-28 12:17:28 <ncopa> this means that alpine v3.6 and v3.5 could be vulnerable
2018-09-28 12:18:55 <ncopa> they are vulnerable
2018-09-28 16:54:13 <leo-unglaub> i am giving alpine linux a try on the new hetzner cloud. but for some reason networking does not work out of the box. i get the correct ip via dhcp, but no connection is possible. it always ends up in a "network unreachable"
2018-09-28 16:55:36 <_ikke_> leo-unglaub: hmm
2018-09-28 16:55:39 <leo-unglaub> i think ther was a change back in the dhcp code that disable the default routing
2018-09-28 16:55:57 <_ikke_> leo-unglaub: Did you check if you got a default route?
2018-09-28 16:56:20 <leo-unglaub> no, i did not get a default route. thats why i am looking for the commit that changed that
2018-09-28 16:56:48 <_ikke_> right
2018-09-28 16:56:54 <_ikke_> Why would that be disabled though?
2018-09-28 16:57:33 <leo-unglaub> i have no idea, but thats a good question
2018-09-28 17:01:46 <_ikke_> did you try setting a static IP, or is that not possible?
2018-09-28 17:01:51 <_ikke_> (just to see if it works)
2018-09-28 17:02:12 <leo-unglaub> i tryed to set a static ip with the same result. the problem is that i cannot set the route myself because i get an error
2018-09-28 17:02:33 <leo-unglaub> route: ioctl 0x890b failed: Network unreadable
2018-09-28 17:02:49 <_ikke_> unreadable or unreachable?
2018-09-28 17:02:58 <leo-unglaub> in this case unreadable
2018-09-28 17:03:00 <_ikke_> ok
2018-09-28 17:03:10 <_ikke_> But that might be preventing dhcp to fail as well, right?
2018-09-28 17:03:38 <leo-unglaub> according to dhcp it works fine. it exists correctly and assigns an ip, a subnet and the gateway. only the routes are missing
2018-09-28 17:03:56 <_ikke_> isn't the gateway the default route?
2018-09-28 17:04:07 <leo-unglaub> yeah, but its not listet as a route
2018-09-28 17:04:13 <leo-unglaub> there are no rountes in there at all
2018-09-28 17:04:29 <_ikke_> right, so it's trying to set a gateway, but fails, it seems
2018-09-28 17:04:53 <leo-unglaub> yes, it seams like it. but i dont get an error message. the error must be hidden somewhere
2018-09-28 17:04:59 <_ikke_> right
2018-09-28 17:05:04 <leo-unglaub> when i try it by hand i get the error messge i posted above
2018-09-28 17:05:12 <_ikke_> How are you tring to add the default route?
2018-09-28 17:05:19 <_ikke_> With route?
2018-09-28 17:05:34 <leo-unglaub> but i am not very familiar with the route utility from busybox, maybe i am doing it wrong
2018-09-28 17:05:48 <leo-unglaub> route add default gw 172.31.1.1
2018-09-28 17:06:10 <_ikke_> I use iproute2 nowadays
2018-09-28 17:06:36 <leo-unglaub> is that in the extended image?
2018-09-28 17:06:46 <leo-unglaub> if not, i cannot use it in this case
2018-09-28 17:07:09 <_ikke_> https://git.alpinelinux.org/cgit/alpine-iso/tree/alpine-extended.packages#n49
2018-09-28 17:07:11 <_ikke_> yes
2018-09-28 17:08:02 <leo-unglaub> i only have iproute from busybox in my install
2018-09-28 17:08:07 <leo-unglaub> is that the one you are using as well?
2018-09-28 17:08:16 <_ikke_> No
2018-09-28 17:08:20 <_ikke_> though
2018-09-28 17:08:23 <_ikke_> probably yes
2018-09-28 17:08:34 <_ikke_> it's probably a slimmed down version
2018-09-28 17:09:00 <_ikke_> ip route add default via 172.31.1.1
2018-09-28 17:12:30 <leo-unglaub> i rebootet the server to get a clean live image 3.8.1 and did the config static
2018-09-28 17:12:50 <leo-unglaub> when i enter your command i get the follwoing: ip: RTNETLINK nswers: Network unreadable
2018-09-28 17:13:02 <_ikke_> Did you add an ip already?
2018-09-28 17:13:12 <leo-unglaub> yes, i did via interfaces file
2018-09-28 17:13:25 <leo-unglaub> confirmed via ifconfig
2018-09-28 17:13:27 <_ikke_> ok
2018-09-28 17:13:42 <_ikke_> I've never seen that error message
2018-09-28 17:13:47 <leo-unglaub> i tested the infrastructure with an ubuntu and a debian image. the datacenter stuff works correctly
2018-09-28 17:13:53 <_ikke_> both ddg and google neither
2018-09-28 17:14:14 <leo-unglaub> ubuntu and debian get the ip and set the route and it works, so i can rule out a server or hardware problem
2018-09-28 17:15:11 <_ikke_> To be sure, that default gateway is in the subnet of your ip address?
2018-09-28 17:15:30 <unmy> ye, what is your subnet mask?
2018-09-28 17:15:42 <leo-unglaub> no, thats the thing with Hetzner
2018-09-28 17:16:00 <leo-unglaub> the gateway is always outside of my own network. thats hacky, but valid behavour
2018-09-28 17:16:13 <leo-unglaub> i also needed a lot of patches to get it to work in openbsd ...
2018-09-28 17:16:36 <leo-unglaub> unmy: the subnet mask is 255.255.255.255
2018-09-28 17:16:47 <_ikke_> that's weird
2018-09-28 17:17:08 <leo-unglaub> _ikke_: yeah, but sadly a lot more hosters are doing this setup
2018-09-28 17:17:31 <leo-unglaub> in openbsd i had to add the route but specify the interface
2018-09-28 17:17:42 <_ikke_> you can try that as well
2018-09-28 17:17:52 <_ikke_> ip route add default via 172.31.1.1 dev <dev>
2018-09-28 17:19:11 <leo-unglaub> already did, no change
2018-09-28 17:19:13 <unmy> leo-unglaub, did you read that? https://wiki.hetzner.de/index.php/CloudServer/en
2018-09-28 17:20:12 <leo-unglaub> unmy: yes, but there is no mention of it. i stuck to https://wiki.hetzner.de/index.php/Cloud_IP_static
2018-09-28 17:20:55 <_ikke_> leo-unglaub: I tried the same, the message is "network unreachable"
2018-09-28 17:21:00 <_ikke_> not unreadabl
2018-09-28 17:21:03 <_ikke_> unreadable
2018-09-28 17:21:12 <unmy> and as you could read it isnt just typical IP configuration but absed on some service floating IP
2018-09-28 17:21:24 <leo-unglaub> _ikke_: unreachable
2018-09-28 17:21:29 <unmy> s/absed/based
2018-09-28 17:21:30 <leo-unglaub> did is misstype?
2018-09-28 17:21:34 <_ikke_> leo-unglaub: yes, you did
2018-09-28 17:21:42 <leo-unglaub> _ikke_: upppps, sorry about that
2018-09-28 17:22:17 <unmy> so prety sure that simple busybox dhcp client dont have support for that
2018-09-28 17:22:31 <_ikke_> Let me check iproute2
2018-09-28 17:22:32 <leo-unglaub> unmy: no, the floating ip stuff is disabled because it costs extra
2018-09-28 17:22:42 <leo-unglaub> so it should not be the problem
2018-09-28 17:23:31 <unmy> so got normal private single IP?
2018-09-28 17:24:04 <leo-unglaub> unmy: yes sir
2018-09-28 17:24:15 <leo-unglaub> normal ip 159.69.180.56
2018-09-28 17:24:42 <_ikke_> nope
2018-09-28 17:24:56 <leo-unglaub> the problem is that the gateway is in a different subnet ... that seams to be to much for the busybox networking tool
2018-09-28 17:25:08 <_ikke_> leo-unglaub: iproute2 doesn't accept it either
2018-09-28 17:25:21 <leo-unglaub> _ikke_: do you have a hetzner vm as well?
2018-09-28 17:25:22 <_ikke_> and its RTNETLINK that returns the error message, that appears to be the kernel
2018-09-28 17:25:27 <leo-unglaub> or where did you try?
2018-09-28 17:25:30 <_ikke_> leo-unglaub: No, I just started a VM
2018-09-28 17:25:36 <_ikke_> Not networked
2018-09-28 17:25:50 <_ikke_> But just try to add a default gateway outside of the interface subnet
2018-09-28 17:25:57 <leo-unglaub> _ikke_: but if its the kernel, why does fedora, debian, ubuntu, .. work with it
2018-09-28 17:26:07 <_ikke_> That's a good question, I'm not sure
2018-09-28 17:26:53 <unmy> or maybe some sysctl setting hmm
2018-09-28 17:27:05 <_ikke_> https://unix.stackexchange.com/questions/98344/does-the-gateway-have-to-be-on-the-subnet
2018-09-28 17:27:54 <leo-unglaub> _ikke_: yes, but that answer is more than 5 years old
2018-09-28 17:30:57 <_ikke_> leo-unglaub: try ip route add 172.31.1.1 dev eth0
2018-09-28 17:32:05 <leo-unglaub> _ikke_: with that i can now poing the gateway 172.31.1.1
2018-09-28 17:32:15 <_ikke_> Can you now add the default route?
2018-09-28 17:32:53 <leo-unglaub> yeah, now i can ping 8.8.8.8
2018-09-28 17:33:11 <leo-unglaub> nice!
2018-09-28 17:33:12 <_ikke_> https://serverfault.com/questions/547344/openbsd-gateway-outside-subnet-works-in-linux/769010
2018-09-28 17:33:16 <_ikke_> found it there
2018-09-28 17:34:07 <_ikke_> So you need an post-up rule to your /etc/network/interfaces that adds that
2018-09-28 17:34:13 <unmy> weird is his mask 255.255.255.255 which means that there is only his address and nothing else so how that should work with anything
2018-09-28 17:34:32 <leo-unglaub> _ikke_: yes, awesome!
2018-09-28 17:34:36 <leo-unglaub> thank you so much
2018-09-28 17:34:48 <_ikke_> unmy: apparently it's enought to have a route to that specific address
2018-09-28 17:35:03 <_ikke_> a route that links the gw to the interface
2018-09-28 17:35:04 <unmy> ye is weird as fuck but I see they doing that, eh
2018-09-28 17:35:45 <leo-unglaub> yes, its wired as hell ... but thats the largest hoster in germany, so knowing how it works there is key for alpine
2018-09-28 17:36:16 <_ikke_> But that means you can never use normal dhcp in that netowrk
2018-09-28 17:36:18 <_ikke_> network
2018-09-28 17:36:31 <AinNero> leo-unglaub: Hetzner Cloud?
2018-09-28 17:36:37 <_ikke_> leo-unglaub: I wonder how this works on ubuntu
2018-09-28 17:36:38 <AinNero> i patched udhcpd there
2018-09-28 17:36:58 <unmy> ovh doing same so... they all got twisted way to do some things :P
2018-09-28 17:37:08 <leo-unglaub> AinNero: yes, Hetzner
2018-09-28 17:37:24 <AinNero> i tried messing there with diskless and apkovl
2018-09-28 17:37:58 <leo-unglaub> _ikke_: i tryed 5 different distros and in all of them it works with dhcp. even in windows ... yes, i tryed a windows image just out of interrest :P
2018-09-28 17:38:11 <AinNero> unmy: with the extra route, the kernel knows that it needs to do ARP on that specific interface
2018-09-28 17:38:29 <unmy> leo-unglaub, but tried different dhcp client in alpine if abble to install it?
2018-09-28 17:38:38 <AinNero> leo-unglaub: i once had a patch, but afaik it never went production-ready
2018-09-28 17:39:00 <leo-unglaub> unmy: no, i did not ... just managed to get it working with _ikke_ help one minute ago ;)
2018-09-28 17:39:21 <unmy> gj :)
2018-09-28 17:39:28 <leo-unglaub> AinNero: maybe you have time some day to post the patch somewhere
2018-09-28 17:39:40 <_ikke_> So apparently the dhcp client knows how to deal with that situation I guess
2018-09-28 17:39:51 <_ikke_> leo-unglaub: Would be interesting to see the ip route output on ubuntu
2018-09-28 17:40:49 <leo-unglaub> _ikke_: sure, debian or ubuntu or fedora? what do you prefer?
2018-09-28 17:41:00 <_ikke_> doesnt matter
2018-09-28 17:42:23 <leo-unglaub> # ip route
2018-09-28 17:42:23 <leo-unglaub> default via 172.31.1.1 dev eth0
2018-09-28 17:42:23 <leo-unglaub> 172.31.1.1 dev eth0 scope link
2018-09-28 17:43:05 <_ikke_> right, so it does add that extra route
2018-09-28 17:43:46 <leo-unglaub> yes, it must be checking if the gateway is on a different subnet and then add that route
2018-09-28 17:44:02 <leo-unglaub> but the question is, what RFC does define this behavour or is this just lucky guessing?
2018-09-28 17:44:37 <_ikke_> I guess it's a matter of: I get a default route via dhcp on interface X, so we can assume that route is reachable on interface X
2018-09-28 17:45:13 <AinNero> turns out chisa also had the same issue with hetzner
2018-09-28 17:45:43 <leo-unglaub> AinNero: i guess a few people have that issue. there in germany they are a big player
2018-09-28 17:45:52 <AinNero> im definitely in for fixing it
2018-09-28 17:46:05 <AinNero> or at least provide some sort of workaround
2018-09-28 17:46:09 <AinNero> or custom installer
2018-09-28 17:46:15 <leo-unglaub> and like they say here in germany: "nit schlecht fir ane bayern" ;)
2018-09-28 17:46:43 <_ikke_> do bayern people have a bad name?
2018-09-28 17:46:55 <AinNero> leo-unglaub: open an ticket for it?
2018-09-28 17:47:02 <AinNero> maybe i'll look into it
2018-09-28 17:47:19 <leo-unglaub> _ikke_: only if you ask people who not live there ;)
2018-09-28 17:47:32 <_ikke_> haha :D
2018-09-28 17:48:05 <_ikke_> isn't that tautological?
2018-09-28 17:48:53 <leo-unglaub> well, i can say that because my family is from there ;) but hetzner is so close to the border to the rest of germany, that hey hate beeing called "bayern"
2018-09-28 17:49:01 <leo-unglaub> never call them that, they propobly kill you
2018-09-28 17:49:02 <leo-unglaub> haha
2018-09-28 19:32:09 <z3ntu> -dev subpackages don't depend on the non-dev package, right?
2018-09-28 19:33:12 <_ikke_> https://pkgs.alpinelinux.org/package/edge/community/x86/libnumbertext-dev
2018-09-28 19:33:24 <_ikke_> (Just picked a random -dev package)
2018-09-28 19:33:31 <_ikke_> Look at the depends
2018-09-28 19:34:39 <z3ntu> _ikke_: This is done automatically?
2018-09-28 19:36:29 <z3ntu> Because I have several packages here some -dev packages don't depend on their base for some reason
2018-09-28 19:36:50 <_ikke_> No, usually not
2018-09-28 19:37:03 <_ikke_> At least, it's not done automatically I mean
2018-09-28 19:37:22 <_ikke_> https://pkgs.alpinelinux.org/package/edge/main/aarch64/ppp-dev
2018-09-28 19:38:07 <z3ntu> Okay so I have to add both example and example-dev as dependency to packages that depend on them?
2018-09-28 19:42:38 <_ikke_> -dev packages are usually just build-time deps
2018-09-28 19:42:46 <_ikke_> (makedepends)
2018-09-28 19:44:48 <z3ntu> The package I want to build also wants the contents in the non-dev package during build time
2018-09-28 19:49:15 <leo-unglaub> whats the correct way to enter my ip route commands in the interfaces file?
2018-09-28 19:49:34 <leo-unglaub> post-up ip route add does not seam to be triggered
2018-09-28 19:50:19 <z3ntu> Also, I have a feeling that lttng-tools needs a rebuild
2018-09-28 19:50:20 <z3ntu> ERROR: unsatisfiable constraints:
2018-09-28 19:50:29 <z3ntu> so:liblttng-ust-ctl.so.2 (missing):
2018-09-28 19:50:29 <z3ntu>     required by: lttng-tools-2.9.6-r0[so:liblttng-ust-ctl.so.2] lttng-tools-2.9.6-r0[so:liblttng-ust-ctl.so.2]
2018-09-28 20:16:51 <_ikke_> Is it normal that the v3.8.1-extended images on boot does not have /etc/network/interfaces?
2018-09-28 20:17:09 <_ikke_> (ie, booting the image itself, no network seems to come up)
2018-09-29 21:46:41 <mps> _ikke_: last night someone asked about adding static route
2018-09-29 21:47:09 <mps> could that be set in /etc/conf.d/staticroute?
2018-09-30 00:01:44 <AinNero> may we grab patches from debian and put them into the aports repo?
2018-09-30 00:01:49 <AinNero> how is the licensing there?
2018-09-30 00:02:11 <AinNero> for pciutils, libpci.a, per default it does not have static library binaries
2018-09-30 00:02:19 <AinNero> and debian people patched something in...
2018-09-30 00:08:12 <duncan^> AinNero: You should expect patches to be the same license as the software that is being patched but there may be caveats to that.
2018-09-30 00:08:39 <duncan^> Just because a patch exists for a piece of software doesn't mean that it's suitable to apply it or that it will solve the issue in question
2018-09-30 00:09:30 <duncan^> Since you're thinking of taking patches from Debian, the license of the software would depend on Debian policy or might be package-specific; you should check that
2018-09-30 00:10:13 <pardis> Debian packages have a file called debian/copyright which should document any package-specific oddities
2018-09-30 07:23:01 <GlemSom> Anyone else having issues with login to bugs.alpinelinux.org ?  I've reset my password twice, and I still get invalid user or password when I try to login :/
2018-09-30 07:49:27 <clandmeter> GlemSom, nope, just logged in correctly
2018-09-30 08:21:39 <z3ntu> Has anyone ever seen an error like `Error relocating lib/libmirserver.so: _ZN5boost4asio6detail15keyword_tss_ptrINS1_10call_stackINS1_14thread_contextENS1_16thread_info_baseEE7contextEE6value_E: symbol not found` ?
2018-09-30 08:22:12 <z3ntu> The .so is built from source on an Alpine VM and this I get this output when running ldd on the .so (plus the linked libraries)
2018-09-30 12:48:53 <z3ntu> Fixed my error. It was caused by using the gold linker instead of the default linker.
2018-09-30 15:55:01 <z3ntu> ncopa: QtLinuxAccessibilitySupport is missing from qt5-qtbase
2018-09-30 17:07:53 <z3ntu> ncopa: I got the information that `-accessibility` should help with that. I'll try it in the next days if you don't do so first
2018-09-30 18:17:42 <z3ntu> ncopa: I opened a PR for it: https://github.com/alpinelinux/aports/pull/5281