2018-07-01 08:53:52 <danieli> nmeum: I thought so
2018-07-01 08:54:17 <danieli> security is a large concern for many alpine users, and it relies on SIGs/WGs
2018-07-01 10:06:36 <mepholic> hey all, I've noticed something strange.
2018-07-01 10:07:02 <mepholic> why do apkbuilds with postinstalls and preinstalls not have any entries in sha512sums
2018-07-01 10:07:09 <mepholic> that seems like a mistake, and dangerous
2018-07-01 10:07:30 <mepholic> I know if you put them in sources=, abuild complains
2018-07-01 10:07:42 <mepholic> and only sources= have sha512sums= entries
2018-07-01 10:08:28 <mepholic> it seems like at the very minimum, abuild should detect the presence of those files and add/check their checksums
2018-07-01 10:14:56 <awilfox> it does detect presence: https://git.alpinelinux.org/cgit/abuild/tree/abuild.in#n166
2018-07-01 10:17:50 <mepholic> awilfox pointed out to me that there's a commit for this: https://git.alpinelinux.org/cgit/abuild/commit/abuild.in?id=df8e468f0e527ca9158d3c27fc80f91262834b7b
2018-07-01 10:18:07 <mepholic> i guess back in 2009, when install scripts were a subpackage?
2018-07-01 14:38:41 <azarus> nmeum: closed the other PRs for GIMP, https://github.com/alpinelinux/aports/pull/4427 is now the only one with all the necessary commits
2018-07-01 14:39:44 <nmeum> thanks!
2018-07-01 14:50:02 <azarus> nmeum: the new commits should contain your proposed change for prepare()
2018-07-01 14:50:32 <azarus> (well "new" just rebased and edited)
2018-07-01 14:54:17 <azarus> if you want me to squash/reorder/whatever them, happy to do it
2018-07-01 15:06:35 <azarus> (also deleted prepare() in gegl now)
2018-07-01 15:17:20 <mps> in which branch of local aports repo should I add patch for package which is in main but I'm not maintainer
2018-07-01 15:17:38 <mps> master or create new git branch
2018-07-01 15:21:36 <azarus> mps: don't make changes in master, that's the job of the commiters
2018-07-01 15:21:43 <azarus> make the changes in your own branch
2018-07-01 15:22:52 <mps> hmm, switch to often between branches. ok, can live with that ;)
2018-07-01 15:25:01 <azarus> that's the point of git! branching is cheap
2018-07-01 15:25:20 <mps> what is best practice, create new branch for every package or one branch for sending patches and new aports
2018-07-01 15:25:36 <azarus> whatever suits you best, but I use 1 branch per change
2018-07-01 15:25:52 <azarus> (well, changes that depend on each other go into the same branch)
2018-07-01 15:26:03 <azarus> there are however many ways to go about it
2018-07-01 15:26:21 <mps> aha, I thought about that, but will consider what fits my habits better
2018-07-01 15:27:31 <mps> btw, do anyone looked at pijul.org , patch based colaborative work
2018-07-01 15:27:43 <mps> s/do/did/
2018-07-01 15:28:43 <azarus> i have, don't see how it benefits us much
2018-07-01 15:29:30 <azarus> git is simply the standard, and more than good enough for our purposes
2018-07-01 15:31:03 <mps> it is new approach. it looks interesting 'in my eyes' but I'm not expert in the field and would like to hear opinions from people more experienced in collaborative work
2018-07-01 15:31:57 <azarus> and also, it's written in rust, which is currently not in a happy state on Alpine
2018-07-01 15:33:10 <mps> yes, I cannot bootstap it in arm64 because cargo/rust doesn't well work on ARM with musl
2018-07-01 15:33:45 <skarnet> if I had a dollar for every "new approach" which is actually "same designs as existing software but half-assed by lack of experience but we call it new because it's the flavour-of-the-month language and it's hype hype hype"
2018-07-01 15:34:24 <mps> skarnet: well, everything was at some time new
2018-07-01 15:34:32 <skarnet> yes
2018-07-01 15:34:39 <skarnet> survivor bias
2018-07-01 15:34:41 <azarus> The Pijul Book states that "Git deals only with snapshots (or versions)" which is fundamentally wrong
2018-07-01 15:35:03 <azarus> Git deals with changesets.
2018-07-01 15:35:22 <azarus> it's like they're confusing git with subversion
2018-07-01 15:36:37 <mps> anyway, I have 'gut feeling' that the pijul is interesting, and I'm proud of my gut feeling about computer technology
2018-07-01 15:37:39 <mps> I'm using Linux from 1993 and nothing else for my work and private use of computers
2018-07-01 15:37:57 <azarus> it would also mean reeducating basically everyone to a new VCS.
2018-07-01 15:38:54 <mps> I had a 'gut feeling' in 1992 that the Linux is 'right thing' when everyone laughed at me when I talked about it
2018-07-01 15:39:37 <TBB> did your gut feeling tell you about the BSD lawsuit? :)
2018-07-01 15:39:42 <azarus> well, feel free to import the entire aports history to pijul, experimenting with it; and then reporting back
2018-07-01 15:39:53 <azarus> i'm genuinely interested in seeing how it goes
2018-07-01 15:40:22 <mps> azarus: maybe I will but not now, I have better plan for this summer :)
2018-07-01 15:40:40 <azarus> we only have 63'395 commits, won't be a challenge at all ;)
2018-07-01 15:42:59 <mps> TBB: I looked at BSD and all around 'culture' and what not. I don't think anything bad about BSD but somehow 'higher consciousness' told me to stay with linux
2018-07-01 15:43:26 <azarus> ... BSD isn't going anywhere, and it's still perfectly good to use
2018-07-01 15:43:31 <mps> azarus: I'm not such masochist
2018-07-01 15:43:53 <azarus> Alpine Linux and OpenBSD are the 2 operating systems I use most
2018-07-01 15:44:13 <skarnet> azarus: land of the elitists ;)
2018-07-01 15:44:24 <skarnet> (not a criticism btw)
2018-07-01 15:44:41 <azarus> I have a long history of distro hopping ;P
2018-07-01 15:47:12 <mps> skarnet: one can simply by macbook or iphone to live in the land of elitists, and not have hassle with BSD :)
2018-07-01 15:47:46 <azarus> apple products and BSD have relatively little in common
2018-07-01 15:49:46 <mps> azarus: I know that, my son works with that crap from apple because it have to do job on that platform but he is not happy with it
2018-07-01 15:50:15 <azarus> his choice, I suppose
2018-07-01 15:50:43 <mps> no, he have to do job for earning money for living
2018-07-01 15:51:34 <mps> he works on Linux for everything else
2018-07-01 15:51:42 <azarus> oh, cool.
2018-07-01 15:54:07 <mps> before few days he explained to me what the crap is that Xcode development 'environment' and how big it is, and for doing simple task as making ios application
2018-07-01 15:54:53 <mps> he had to install about 4GB (IIRC size) to that simple task
2018-07-01 15:55:10 <mps> s/to that/to do that/
2018-07-01 15:56:42 <mps> ok, tnx to all for chat. going to lunch.
2018-07-02 03:00:26 <unkn-error> Hi
2018-07-02 03:00:28 <unkn-error> Do you guys use Alpine Linux on bare-bone metal hardware like on a Laptop, PC, Server, Pi etc. or just in Docker?
2018-07-02 03:10:51 <Xe> unkn-error: i use it on a cluster of 6 servers on my home network
2018-07-02 03:10:58 <Xe> metal servers
2018-07-02 03:30:50 <unkn-error> If Alpine will be used as a server with GUI on a Raspberry Pi, and exposed to internet a "sys" mode traditional install will be okay (secure)? Or it will be better to be installed in "data" or "diskless" mode? From a security point of view
2018-07-02 03:32:13 <unkn-error> Xe, I guess that "sys" mode install would be more convenient, no?
2018-07-02 05:51:16 <azarus> ... Anybody else getting bad signatures on installation of 3.8.0? https://i.imgur.com/i8z3FaR.png
2018-07-02 05:51:41 <azarus> (tried with multiple mirrors)
2018-07-02 07:16:59 <nmeum> ncopa[m]: do you still use xfce? If so: could you review the various xfce PRs on github (#4601 - #4613)?
2018-07-02 07:19:00 <nmeum> azarus: I use nl.alpinelinux.org and I didn't get any bad signature errors so far
2018-07-02 07:23:19 <clandmeter> azarus, i just tried an install from dl-cdn and its fine.
2018-07-02 07:23:51 <clandmeter> make sure you didnt break your /etc/apk/keys
2018-07-02 07:25:12 <azarus> strange, don't see why that would break
2018-07-02 07:52:27 <clandmeter> azarus, ah you are using qemu.
2018-07-02 07:52:31 <clandmeter> thats weird.
2018-07-02 07:52:35 <azarus> yeah
2018-07-02 07:52:37 <clandmeter> which installation media did you use?
2018-07-02 07:52:51 <azarus> virtual
2018-07-02 07:53:10 <clandmeter> have you tried netboot?
2018-07-02 07:53:17 <azarus> nope
2018-07-02 07:53:51 <clandmeter> i was using  qemu-system-x86_64 -m 512M -enable-kvm -kernel /usr/share/alpine-ipxe/ipxe.lkrn -curses
2018-07-02 07:59:07 <clandmeter> using the virt iso also works fine for me.
2018-07-02 07:59:16 <clandmeter> azarus, x86_84 i guess?
2018-07-02 07:59:25 <azarus> trying different things, not only x86_64
2018-07-02 07:59:42 <clandmeter> sounds like a network issue?
2018-07-02 07:59:47 <azarus> maybe
2018-07-02 09:23:12 <awilfox> hi, so the 3.8 mini root filesystem thingies?  when you extract one like 'tar -C /mountpoint -xzf alpine-minirootfs-3.8.0-whatever.tar.gz', it sets /mountpoint as chmod 700 (rwx------), which means when you try to use it as a user (on the target system or such) you can't
2018-07-02 09:23:17 <awilfox> you don't have permission to read/write /
2018-07-02 09:23:31 <awilfox> this was an issue in 3.7 too, but I don't think I ever said anything
2018-07-02 09:23:50 <awilfox> I only ever tried armhf so I thought it was just my armhf tar being dumb
2018-07-02 09:23:59 <awilfox> but it's just happened to me on 3.8 with ppc64le
2018-07-02 09:51:00 <azarus> what happend with https://github.com/alpinelinux/aports/pull/4003 ? Has been open since april, and there was nothing yet :/
2018-07-02 17:33:21 <gnu3ra> ~.
2018-07-03 06:55:18 <fcolista> this python requirements binding to specific versions are a nightmare
2018-07-03 06:55:49 <fcolista> with edge we will never be able to make it work. I think we should go ahead and using pip
2018-07-03 09:16:41 <azarus> 
ACTION is compiling testing/firefox-61.0

2018-07-03 09:16:49 <azarus> needs a newer nss tough...
2018-07-03 09:56:21 <azarus> compiled! woot
2018-07-03 09:56:49 <azarus> but I'm not sure if touch nss is wise...
2018-07-03 09:57:19 <azarus> touching*
2018-07-03 09:57:30 <azarus> eh. I'll send it anyway
2018-07-03 09:58:39 <azarus> nss includes a soname bump.
2018-07-03 09:58:58 <skarnet> why does firefox need nss...
2018-07-03 09:59:48 <azarus> skarnet: it's firefox's TLS implementation?
2018-07-03 09:59:58 <skarnet> ... and?
2018-07-03 10:00:08 <azarus> and firefox-61.0 needs a newer nss than we have
2018-07-03 10:00:19 <azarus> (note, not firefox-esr)
2018-07-03 10:00:19 <skarnet> that, like, doesn't answer my question at all
2018-07-03 10:00:37 <azarus> Firefox uses NSS as it's TLS implementation.
2018-07-03 10:00:41 <mepholic> skarnet: lmao
2018-07-03 10:00:44 <mepholic> not that nss
2018-07-03 10:00:51 <skarnet> oh.
2018-07-03 10:00:53 <azarus> ... what nss is skarnet talking about?
2018-07-03 10:00:54 <skarnet> what is nss then?
2018-07-03 10:01:01 <azarus> network securtiy services by mozilla
2018-07-03 10:01:02 <mepholic> nsswitch
2018-07-03 10:01:04 <mepholic> lol
2018-07-03 10:01:06 <azarus> aha
2018-07-03 10:01:12 <azarus> skarnet: https://pkgs.alpinelinux.org/package/edge/main/x86/nss
2018-07-03 10:01:35 <skarnet> naming is hard
2018-07-03 10:01:38 <mepholic> both maintained by mozilla i think
2018-07-03 10:01:40 <skarnet> so hard that mozilla fucks it up
2018-07-03 10:01:40 <mepholic> lol
2018-07-03 10:01:41 <azarus> yup
2018-07-03 10:01:58 <skarnet> oh wait, mozilla fucking something up isn't proof that something is hard
2018-07-03 10:02:27 <azarus> so, if an upgrade introduces a soname bump, all packages that depend on it need to have their pkgrel's bumped?
2018-07-03 10:02:54 <skarnet> I guess, since it's an ABI change
2018-07-03 10:03:20 <azarus> OK
2018-07-03 10:03:28 <skarnet> anyway, sorry for the noise, my mind is currently busy with the _other_ NSS, which is why I got confused.
2018-07-03 10:04:04 <azarus> skarnet: no problem :)
2018-07-03 10:11:05 <azarus> wow. 17 pkgrel bumps... i feel like I'd hurt the builders, as those are some pretty intense packages to compile
2018-07-03 10:20:10 <mps> anyone can tell why I'm get '/bin/bash: ../depcomp: No such file or directory' with 'abuild build' and I have automake installed
2018-07-03 10:21:01 <mps> I'm trying to make weex aport
2018-07-03 10:21:59 <mps> "Non-interactive FTP and FTPS client for updating web pages" http://weex.sourceforge.net/
2018-07-03 10:26:05 <azarus> hmm, i'd like to emulate what travis does to test if every changed aport builds. currently stuck with this: https://ptpb.pw/S0c1
2018-07-03 10:26:41 <azarus> I changed the build-pkgs script so I can give the commit range via $1, but it doesn't actually detect which packages have changed
2018-07-03 10:29:37 <mps> ok, solved by adding --disable-dependency-tracking to configure in build APKBUILD file
2018-07-03 10:31:42 <mps> but, have another question. How can be --build triplet changed to ARCH-unknown-none as the abuild variable
2018-07-03 10:32:26 <mps> where ARCH is the build host architecture
2018-07-03 11:13:57 <mps> solved somehow APKBUILD for weex by adding 'export CHOST="${CARCH}-unknown-none"' and 'export CBUILD=$CHOST' to build function
2018-07-03 11:14:48 <mps> but not sure if that is 'proper' way/method to do such hack in AL APKBUILD aports
2018-07-03 11:16:08 <mps> anyone experienced in APKBULD is willing to comment about that, I could post APKBUILD file to tpaste
2018-07-03 11:23:05 <clandmeter> post it and see
2018-07-03 11:26:09 <mps> clandmeter: here it is http://tpaste.us/oQLY
2018-07-03 11:31:22 <azarus> interesting, can anyone build pidgin-sipe?
2018-07-03 11:31:40 <azarus> it outputs http://ix.io/1fRE for me
2018-07-03 11:34:24 <azarus> ah my fault, disregard
2018-07-03 11:39:24 <clandmeter> mps autocrap is outdated
2018-07-03 11:39:56 <clandmeter> add update_config_sub in prepare
2018-07-03 11:45:28 <azarus> okay -- published my upgrade of testing/firefox to the list.
2018-07-03 11:49:49 <mps> clandmeter: don't understand. should I add prepare function?
2018-07-03 11:59:22 <mps> clandmeter: I see now. Added prepare sub/function with 'default_prepare' and 'update_config_sub' and it works
2018-07-03 11:59:34 <mps> clandmeter: tnx for hint
2018-07-03 13:05:57 <azarus> aw
2018-07-03 13:07:21 <azarus> do you guys have build environments for different arches?
2018-07-03 13:19:54 <clandmeter> env as in?
2018-07-03 13:20:22 <azarus> to build/test packages in
2018-07-03 13:20:38 <azarus> for example, what kinds of machines are used to build aarch64?
2018-07-03 13:23:02 <clandmeter> we build native.
2018-07-03 13:23:12 <clandmeter> we have a thunderx for aarch64
2018-07-03 13:23:43 <clandmeter> we have 2 of them, one sponsored by gigabyte.com and one by packet.net
2018-07-03 13:24:30 <azarus> oh cool, but is there a way to reasonably emulate them?
2018-07-03 13:24:45 <azarus> i fumbled for hours with qemu-system-aarch64, to no avail
2018-07-03 13:25:10 <clandmeter> you can build with qemu, not sure about the performance
2018-07-03 13:25:31 <azarus> doesnt really matter; just want to know if I could enable some packages for other arches
2018-07-03 13:25:32 <clandmeter> or cross build
2018-07-03 13:25:50 <azarus> does abuild have some kind of mechanism for cross compiling?
2018-07-03 13:25:54 <clandmeter> you mean you want to test of your packages build on other archs?
2018-07-03 13:26:11 <azarus> yes, or existing ones that have been disabled for odd reasons
2018-07-03 13:26:39 <azarus> that aren't the case anymore.
2018-07-03 13:27:07 <clandmeter> we have limited cross building support to bootstrap aports.
2018-07-03 13:27:15 <clandmeter> its in the scripts directory in aporst.
2018-07-03 13:27:15 <mps> qemu is slow on non-native cpu
2018-07-03 13:27:26 <azarus> mps: i know, but at least it's something
2018-07-03 13:27:32 <azarus> clandmeter: OK, thanks
2018-07-03 13:28:25 <mps> I had armhf build env in qemu on x86_64 but it was too slow so I moved to small arm board
2018-07-03 13:28:36 <azarus> is that faster?
2018-07-03 13:28:47 <clandmeter> a rpi is probably faster yes
2018-07-03 13:28:48 <mps> of course, significantly
2018-07-03 13:29:06 <azarus> ah alright, then I'll get my raspberry pi 2 running alpine again
2018-07-03 13:29:15 <azarus> just to test for armhf
2018-07-03 13:29:18 <mps> faster than qemu on i7 with 6GB ram
2018-07-03 13:29:26 <clandmeter> or rpi3+
2018-07-03 13:29:47 <azarus> for aarch64?
2018-07-03 13:30:09 <azarus> then I could run armhf and aarch64, neat
2018-07-03 13:30:13 <clandmeter> rpi3 and on can support both arm arch
2018-07-03 13:30:27 <clandmeter> this is how we do armhf
2018-07-03 13:30:37 <clandmeter> of course not on rpi :)
2018-07-03 13:31:00 <azarus> OK, will look into getting an RPi3 for testing aarch64 aports :)
2018-07-03 13:31:05 <clandmeter> we have xgene board that can do arm32 on aarch64
2018-07-03 13:31:54 <clandmeter> azarus, https://www.theregister.co.uk/2018/07/02/tiny_computer_receives_tiny_linux_distribution/
2018-07-03 13:32:34 <azarus> Cool!
2018-07-03 13:34:31 <clandmeter> we use 96 cores to build on aarch64 ;-)
2018-07-03 13:34:36 <clandmeter> 
ACTION hides from skarnet 

2018-07-03 13:35:08 <mps> nice article about Alpine :)
2018-07-03 13:35:14 <azarus> 
ACTION would love to have an aarch64/armhf server...

2018-07-03 13:35:17 <azarus> a proper one!
2018-07-03 13:36:03 <mps> I build on arm64 some packages
2018-07-03 13:36:27 <clandmeter> azarus, did you look at scaleway?
2018-07-03 13:36:50 <mps> six cpus and 4GB ram, works quite fine
2018-07-03 13:36:55 <azarus> ooh, cools
2018-07-03 13:36:58 <azarus> cool*
2018-07-03 13:37:26 <clandmeter> ARM64-2GB	4 ARMv8	2 GB	50 GB	200 Mbits/s	€2.99/mo
2018-07-03 13:37:37 <mps> and it is my 'notebook' for work when I'm not at the main machine
2018-07-03 13:37:50 <azarus> that's actually a reasonable price
2018-07-03 13:38:15 <skarnet> clandmeter: sorry, I can't hear you over the speed of a Talos II
2018-07-03 13:38:17 <azarus> 2GB might be a bit... eh for building, but hey, one can always use swap
2018-07-03 13:38:31 <clandmeter> but remember, those are thunderx based. they do not support arm32.
2018-07-03 13:38:31 <azarus> pft, power9 isn't supported by alpine, is it
2018-07-03 13:38:45 <azarus> or is that ppc64le?
2018-07-03 13:39:15 <skarnet> ppc64, either endianness (but making it work BE *is* some work)
2018-07-03 13:39:38 <azarus> skarnet: do you have a Talos II?
2018-07-03 13:39:56 <skarnet> I don't. I wish I had though :D
2018-07-03 13:40:08 <skarnet> had one*
2018-07-03 13:40:24 <skarnet> but I've had the opportunity to see one in action. It's... fast.
2018-07-03 13:40:41 <azarus> I would also love to have one, just don't have the cash for it
2018-07-03 13:41:03 <skarnet> I may be able to afford one next yar
2018-07-03 13:41:06 <skarnet> next year, even
2018-07-03 13:41:27 <azarus> to have something fast that's non-x86-64 would be nice
2018-07-03 13:41:32 <clandmeter> skarnet, what would be your use case?
2018-07-03 13:42:06 <skarnet> clandmeter: building a distro sounds like a good use case, doesn't it? :P
2018-07-03 13:42:35 <clandmeter> youre building a disto now?
2018-07-03 13:43:02 <skarnet> didn't I just say I didn't have a Talos II?
2018-07-03 13:43:40 <clandmeter> so? that means its out of the question?
2018-07-03 13:44:04 <clandmeter> it all depends on one machine?
2018-07-03 13:44:12 <skarnet> but if I had one, I would *certainly* find a use for it. For instance, having a shitload of VMs on it with every OS I want to support, so I can test my software portability without relying on polarhome.com or other accounts.
2018-07-03 13:45:00 <skarnet> and yes, I won't be the person building a distro any time soon - not if I can avoid it
2018-07-03 13:45:23 <clandmeter> you have my blessing :p
2018-07-03 13:45:30 <skarnet> (I'll probably do that before learning C++ though. There's stuff I don't want to do and stuff I REALLY don't want to do.)
2018-07-03 13:46:11 <azarus> learning C++ doesn't cost 5000$ USE tough
2018-07-03 13:46:13 <azarus> USD*
2018-07-03 13:46:23 <azarus> unlike the TALOS II :p
2018-07-03 13:46:30 <skarnet> it costs valuable time
2018-07-03 13:46:49 <azarus> so does talking to me, then ;)
2018-07-03 13:47:10 <skarnet> not to flatter you, but talking to you is more pleasant than learning C++
2018-07-03 13:47:34 <clandmeter> thats my problem as well, hardware enough but all out of time...
2018-07-03 13:49:06 <clandmeter> i would swap some threadrippers and thunderx boxes for a single core time machine.
2018-07-03 14:01:04 <azarus> "thunderx boxes" -- sounds like really fancy Xboxes
2018-07-03 16:56:46 <HRio> clandmeter: /etc/machine-id possibly feels a bit systemd specific, but it seems lvm2 and other SW now checks for this file... should we perhap generate one in firstboot as well?
2018-07-03 18:09:26 <HRio> something like this in firstboot "tr -d - < /proc/sys/kernel/random/uuid > /etc/machine-id" perhaps?
2018-07-03 18:13:25 <clandmeter> checks as in mandatory now?
2018-07-03 18:22:00 <HRio> No but I guess all alpine systems will have the same id, do not know if thats good or bad when moving disks between hosts
2018-07-03 18:23:32 <HRio> seems glib uses that file as well if dbus is not on the system
2018-07-03 18:30:49 <HRio> zfs uses glib, so there might be some expectations of unique ids there as well
2018-07-03 18:41:01 <clandmeter> did you check what other non systemd systems do? is there some general adoption of it?
2018-07-03 18:44:47 <HRio> no was just a bit worried when I read this https://www.freedesktop.org/software/systemd/man/machine-id.html that we might have code in alpine expecting that its populated properly.
2018-07-03 18:48:45 <HRio> i.e. that we might run in to strange issues, due to that we do not populate it
2018-07-03 18:50:57 <HRio> whats dists are fully systemd free at the moment (except us and devuan)? arch has systemd right?
2018-07-03 18:57:57 <HRio> tested devuan 2.0.0 live-cd and they have dbus and the machine-id in /var/lib/dbus/, but thats a desktop live CD.
2018-07-03 19:05:53 <mps> HRio: Debian could be freed from systemd, http://angband.pl have description how to do that
2018-07-03 19:06:35 <HRio> yes but I guess /etc/machine-id is generated before that process can begin
2018-07-03 19:09:58 <HRio> I guess we will have to dig trough lvm and zfs sources to see if we can *still* live without it...
2018-07-03 19:11:10 <HRio> anyhow AFK time for me...
2018-07-03 19:38:28 <HRio> centos includes /etc/machine-id already in initramfs
2018-07-03 19:39:09 <HRio> so its used very early on...
2018-07-03 22:14:12 <vkrishn> hi, tried couple of times, seems checksm  for alpine-extended-3.8.0-x86.iso does not match
2018-07-03 22:14:17 <vkrishn> on nl.a.o
2018-07-03 22:15:34 <azarus> is patchwork.alpinelinux.org down for anyone else?
2018-07-03 22:16:46 <skarnet> https://downforeveryoneorjustme.com/patchwork.alpinelinux.org
2018-07-03 22:16:51 <skarnet> says it's down
2018-07-03 22:16:56 <azarus> checked that too, wanted to make sure
2018-07-03 22:16:59 <azarus> oh well
2018-07-03 22:17:17 <skarnet> sorry, couldn't be sure you knew of that ^^
2018-07-03 22:17:26 <azarus> no problem
2018-07-03 22:17:29 <azarus> anyhow, time to sleep
2018-07-03 22:26:53 <azarus> pkgs.alpinelinux.org is down as well
2018-07-03 22:27:20 <azarus> heck, alpinelinux.org too
2018-07-04 00:36:31 <Galatea> hi, hi have an issue with grub, i cant update it
2018-07-04 00:36:36 <Galatea> I*
2018-07-04 02:15:29 <Galatea> this is what i get doing a "grub-mkconfig"
2018-07-04 02:15:33 <Galatea> if [ x$feature_default_font_path = xy ] ; then    font=unicode else You have a memory leak (not released memory pool):  [0x7f3747015ee0] dtree Internal error: Unreleased memory pool(s) found. You have a memory leak (not released memory pool):  [0x7f3747015ee0] dtree Internal error: Unreleased memory pool(s) found. insmod part_gpt You have a memory leak (not released memory pool):  [0x7f1d4abe2ee0] dtree Internal error: Unrelease
2018-07-04 11:06:51 <shenron> hello im trying to compile a cpp application that requires libexecinfo, ive installed that package but it still keeps throwing this error  undefined reference to `backtrace_symbols', google shows some discussion the libexecinfo still requires some glibc header files, so is it possible to compile this with libexecinfo or do I have to use libunwind for musl systems?
2018-07-04 12:23:44 <clandmeter> shenron, you have to pass the lib to the linker.
2018-07-04 12:38:25 <mepholic> hey all, I've noticed something strange.
2018-07-04 12:38:32 <mepholic> why do apkbuilds with postinstalls and preinstalls not have any entries in sha512sums
2018-07-04 12:38:40 <mepholic> that seems like a mistake, and dangerous
2018-07-04 12:38:49 <mepholic> I know if you put them in sources=, abuild complains
2018-07-04 12:38:56 <mepholic> and only sources= have sha512sums= entries
2018-07-04 12:39:04 <mepholic> it seems like at the very minimum, abuild should detect the presence of those files and add/check their checksums
2018-07-04 12:39:17 <mepholic> awilfox pointed out to me that there's a commit for this: https://git.alpinelinux.org/cgit/abuild/commit/abuild.in?id=df8e468f0e527ca9158d3c27fc80f91262834b7b
2018-07-04 12:39:29 <mepholic> i guess back in 2009, when install scripts were a subpackage?
2018-07-04 12:40:44 <mepholic> idk, it seems to me that the shasums should be checked regardless, even if it's just to prevent cases where they were somehow garbled on download
2018-07-04 12:41:12 <mepholic> if there's any reasoning as to why they're not, i'd love to hear it
2018-07-05 10:15:26 <nmeum> the TLS certificate for nl.alpinelinux.org is broken. it has the wrong common name. nl3.alpinelinux.org instead of nl.alpinelinux.org
2018-07-05 10:15:34 <nmeum> clandmeter: ^ is that a known issue?
2018-07-05 10:31:50 <clandmeter> Yes i need to update it
2018-07-05 10:32:44 <clandmeter> nmeum, does apk error now?
2018-07-05 11:57:55 <nmeum> clandmeter: yes, apk prints an error message
2018-07-05 11:58:46 <mepholic> so nobody has any comments or concerns about pre and postinstall scripts that contain arbitrary commands not having shasums?
2018-07-05 11:59:44 <nmeum> why do you believe that it is problem that those scripts are not checksumed? they are not fetched from any external site, unlike tarballs for instance
2018-07-05 12:02:50 <mepholic> well, other sources that are contained in the git repo are also checksummed
2018-07-05 12:03:04 <mepholic> so i guess another question would be, why are those checksummed?
2018-07-05 12:03:22 <mepholic> it's just a strange inconsistancy
2018-07-05 12:08:19 <nmeum> indeed but it shouldn't have any security implications so there is no need to be concerned about it
2018-07-05 12:14:07 <clandmeter> nmeum: i was under the impression ppl used http for apk.
2018-07-05 12:14:38 <mepholic> clandmeter: why?
2018-07-05 12:14:48 <mepholic> https is just one layer of defense in depth
2018-07-05 12:14:55 <mepholic> signed packages being another
2018-07-05 12:15:24 <mepholic> and potentially shasummed pre and post install scripts being another
2018-07-05 12:15:31 <clandmeter> because you are wasting cpu
2018-07-05 12:15:41 <mepholic> I know that if the git repo is owned, it's effectively game over
2018-07-05 12:16:19 <mepholic> but it also makes it harder for someone to add "rm -rf /" or whatever to every preinstall script, or whatever
2018-07-05 12:16:39 <mepholic> like what happened to gentoo recently on their github page
2018-07-05 12:16:48 <mepholic> granted, I know they'd just shove that in the apkbuild itself
2018-07-05 12:17:36 <mepholic> clandmeter: what platforms do you expect people to run where "wasted cpu" is a problem for an additional layer of security?
2018-07-05 12:17:50 <mepholic> especially on a thing that happens only every once in a while
2018-07-05 12:18:06 <mepholic> or are you more concerned about "wasted cpu" on the mirrors themselves?
2018-07-05 12:18:25 <clandmeter> im no concerned at all
2018-07-05 12:18:39 <clandmeter> im just saying its  a waste of cpu
2018-07-05 12:18:52 <mepholic> yeah, so are a lot of other things
2018-07-05 12:19:28 <mepholic> you could argue that gpg signing packages is a waste of cpu
2018-07-05 12:19:44 <skarnet> arguing on irc is a waste of cpu
2018-07-05 12:19:55 <nmeum> clandmeter: I do get where you are coming from but why even have ssl if the certificate is invalid?
2018-07-05 12:20:01 <skarnet> (and human cpu at that!)
2018-07-05 12:21:08 <clandmeter> nmeum: we provide https because ppl request it. even if me and others think its a waste of cpu.
2018-07-05 12:21:28 <clandmeter> it gives a sense of false security
2018-07-05 12:21:59 <clandmeter> this has been discussed many times here.
2018-07-05 12:23:58 <mepholic> and as has been mentioned in the past, GPG signatures are the end-to-end guarantee that should be relied upon. at the same time though, if GPG keys are poorly handled, GPG can ALSO provide a false sense of security. there's a reason why infosec professionals advocate for defense in depth
2018-07-05 12:33:32 <clandmeter> nmeum: should be fixed now.
2018-07-05 12:38:54 <nmeum> clandmeter: indeed, thanks!
2018-07-05 15:44:16 <mountain> Hi guys im trying to compile something on alpine and get this error Error relocating /usr/local/lib/somelib.so: __cxa_init_primary_exception: symbol not found. I've had a look around and it suggest passing to the link the stdc++ library which i do with -lstdc++ but it still throws this error, has anyone got a way to fix this up?
2018-07-05 16:35:04 <Shiz> mountain: use g++ as your linker/compiler for C++ code, not gcc
2018-07-05 16:35:13 <Shiz> for somelib.so
2018-07-05 21:32:23 <ebike> hey guys, any news on visual studio code port?
2018-07-05 22:00:27 <HRio> tmh1999[m]: hi did you have a chance to look at the failing s390x build of strace?
2018-07-06 09:05:07 <mountain> @shiz i am using g++, i compiled g++ 7.30 from source on alpine
2018-07-06 09:32:34 <Shiz> hmm
2018-07-06 09:33:00 <Shiz> and somelib.so is compiled from scratch?
2018-07-06 09:33:06 <Shiz> (e.g. not from another binary release)
2018-07-06 09:33:34 <Shiz> __cxa_init_primary_exception is a libsupc++ symbol, which g++ should automatically link against
2018-07-06 09:33:40 <Shiz> I guess you can try adding -lsupc++ yourself
2018-07-06 12:29:09 <mountain> @shiz I find the linking is behaving completely differently on alpine, the codebase was original built on debian/glibc but we are porting it to alpine for an embedded system.
2018-07-06 12:29:22 <mountain> my g++ was compiled like this Configured with: /root/objdir/../gcc-7.3.0//configure --prefix=/root/GCC-7.3.0 --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --build=x86_64-alpine-linux-musl --host=x86_64-alpine-linux-musl --target=x86_64-alpine-linux-musl --with-pkgversion='CryptOS 7.3.0' --enable-checking=release --disable-multilib --disable-werror --enable-__cxa_atexit --enable-default-pie --enable-cloog-bac
2018-07-06 12:30:24 <mountain> yes somelib.so is being compiled from scratch and compiles with no errors, im just getting this error when trying to start the daemon
2018-07-06 12:30:47 <mountain> i've linked -lsupc++ as you suggested but it still errors out
2018-07-06 12:37:24 <mountain> doing  ar -t /usr/lib/libsupc\+\+.a shows the members it contains
2018-07-06 14:43:28 <Shiz> hmm
2018-07-06 15:04:10 <mountain> the g++ spec got truncated i put them on pastebin https://pastebin.com/zseGCrfX
2018-07-06 15:04:43 <mountain> i tried to build it as similar to the debian version but targeting musl
2018-07-06 15:43:44 <mepholic> i found a nasty bug in abuild checksum
2018-07-06 15:44:05 <mepholic> if the checksum line exists in the middle of the APKBUILD, and you run abuild checksum
2018-07-06 15:44:17 <mepholic> it can kill some relevent lines
2018-07-06 15:44:40 <mepholic> in my case, it demolished the build() {
2018-07-06 15:44:42 <mepholic> line
2018-07-06 15:45:18 <mepholic> ... and the whole prepare() section
2018-07-06 15:45:20 <mepholic> O_O
2018-07-06 15:47:00 <mepholic> the checksum line can be put in the middle of the APKBUILD by a 'newapkbuild'
2018-07-06 16:01:49 <clandmeter> yes its been reported
2018-07-06 16:02:16 <clandmeter> checksums should always be at the bottom
2018-07-06 16:02:28 <clandmeter> some bug in recent commit
2018-07-06 16:05:18 <clandmeter> https://github.com/alpinelinux/abuild/commit/75b8cacaf0468470c502314560a436b80b7fe9af
2018-07-06 16:11:39 <mepholic> ah, oops
2018-07-06 16:14:19 <clandmeter> mepholic: if you got time please create an issue for it. check if its not already added.
2018-07-06 16:17:36 <mepholic> man this is a whole bunch of weirdness
2018-07-06 16:17:50 <mepholic> maybe the proper solution was to move the unpack down
2018-07-06 16:22:57 <mepholic> clandmeter: done
2018-07-06 16:23:02 <clandmeter> grazie!
2018-07-06 18:45:30 <sphrak> hey, do I have to do anything in particular to build a -dev package?
2018-07-06 18:54:57 <_ikke_> make sure it's added as a subpackage
2018-07-06 18:57:14 <sphrak> @_ikke_ ahh it was already done by previous packager so its all fine :)
2018-07-06 18:58:49 <sphrak> _ikke_: **
2018-07-07 16:03:11 <sphrak> is it the maintainer of a package or somebody else that reviews the aport PRs on github?
2018-07-08 23:47:13 <ebike> Hi there, anyone had any experience getting suspend to work on laptop lid close?
2018-07-08 23:47:33 <[[sroracle]]> you need acpid for that i think
2018-07-08 23:47:33 <ebike> I can suspend ok with "pm-suspend" but hook doesn't work
2018-07-08 23:47:45 <[[sroracle]]> or at least that's what i've used in the past
2018-07-08 23:47:45 <ebike> I have acpid service running
2018-07-08 23:48:15 <[[sroracle]]> did you create the action and event for laptop lid
2018-07-08 23:48:32 <ebike> please elaborate? i was using the wiki guid
2018-07-08 23:48:41 <ebike> giude
2018-07-08 23:49:28 <[[sroracle]]> i put something like this in /etc/acpi/events: http://ix.io/1gxf
2018-07-08 23:49:33 <ebike> was foolowing https://wiki.alpinelinux.org/wiki/Suspend_on_LID_close
2018-07-08 23:49:42 <[[sroracle]]> then you write the corresponding shell script that it refers to
2018-07-08 23:49:52 <[[sroracle]]> (and make sure it's executable)
2018-07-08 23:50:45 <ebike> so just doing a hook with /etc/acpi/LID/00000080 does not work?
2018-07-08 23:50:56 <[[sroracle]]> i'm not sure, i've never done it that way
2018-07-08 23:51:18 <ebike> silly me, that will teach me for following the wiki
2018-07-08 23:51:45 <[[sroracle]]> there's more information on the arch wiki as usual https://wiki.archlinux.org/index.php/acpid
2018-07-08 23:52:24 <ebike> thanks, will give it a try
2018-07-09 10:01:57 <ny483> Hello, I am new here and fairly new to Alpine Linux. I am trying to solve an issue with creating my own repository. This repository should provide apk files to both aarch64 and armhf architectures. I have created a package with its APKBUILD arch set to noarch (because that specific package should be arch independent) and I would like to add this noarch package into my repository, however my repository only contains folders for aa
2018-07-09 10:04:39 <ny483> When I build the package using CARCH=$arch abuild -d (that package has no dependencies so -d should be ok)  in a for loop for aarch64 and armhf, I am getting two apk files. I then move those apks into aarch64/ and armhf/ folders, and create the index
2018-07-09 10:05:48 <ny483> When I try to install that package, I get an error cherry-rasclient-1.0.0-r0: package mentioned in index not found, and looking at the nginx logs I can see it is looking for that package in noarch/ folder, how can I solve this?
2018-07-09 10:06:55 <ny483> How can I get this to work correctly?
2018-07-09 10:35:47 <ny483> indexing the repository folders with --rewrite-arch seems to index it the way I want is there any other way?
2018-07-09 14:12:05 <kevdog8989> hi people is there away to build an alpine iso as full install not as a live CD?
2018-07-09 14:38:59 <Ganwell> kevdog8989: what is difference between live cd and full install to you? more (all) packages cached?
2018-07-09 14:40:04 <Ganwell> kevdog8989: so you want Extended: Most common used packages included. Suitable for routers and servers. Runs from RAM.
2018-07-09 14:43:17 <kevdog8989> well different between a live and full install is that the full install is written to disk and persists between reboots, but it is for an embbeded application where we dont want user to have to ssh in and run alpine-setup
2018-07-09 14:43:58 <kevdog8989> so when we burn the iso to a medium it boots into the full install not a live system
2018-07-09 15:11:25 <_ikke_> kevdog8989: you run alpine-setup and then choose disk when it asks you
2018-07-09 15:13:34 <kevdog8989> yeh i know that but I want to ship[ it in an embedded system where a user wont be running the setup them selves
2018-07-09 15:17:02 <_ikke_> so you want an unattended installer, is that it?
2018-07-09 15:28:13 <kevdog8989> yes
2018-07-09 15:29:17 <_ikke_> You can probably make something that installs to disk after loggin in (with an answer file for alpine-setup_
2018-07-09 15:55:18 <kevdog8989> yes i was trying to avoid that i dont see why you can just do a full install
2018-07-09 15:55:48 <kevdog8989> generate an iso for a full install
2018-07-09 15:56:23 <kevdog8989> maybe a modified mkinitfs?
2018-07-09 16:16:43 <Ganwell> kevdog8989: I guess you can take the minimal install root, chroot into it, issue the commands to install everything and then create iso image from that folder. https://git.alpinelinux.org/cgit/alpine-iso/tree/Makefile#n324
2018-07-09 16:33:40 <Guest4485> hi if i want to build a package for multiple architectures how do I do that as at the moment abuild just produces an x86_64 apk automatically, my arch is set to all so should it not build apks for each arch?
2018-07-09 17:40:52 <nmeum> Guest4485: abuild doesn't automatically build/install a cross compiler toolchain thus it only builds packages for the host architecture
2018-07-09 17:41:15 <nmeum> if you want to cross compile packages take a look at scripts/bootstrap.sh in the aports repository
2018-07-09 18:10:37 <vkrishn> jirutka: does tarantool still has problem compiling with other archs ? newer version is in 1.9.x
2018-07-09 18:12:39 <vkrishn> also it did not get updated for 2 al cycle, likely candidate for purge
2018-07-09 18:18:32 <azarus> How should one treat custom licenses?
2018-07-09 18:19:08 <azarus> (ones not listed by the SPDX)
2018-07-09 18:34:56 <nmeum> azarus: good question, the previous approach was setting license to custom and installing the license file to /usr/share/licenses/$pkgname/LICENSE
2018-07-09 18:35:03 <nmeum> many packages still do this, e.g. unrar
2018-07-09 18:35:21 <azarus> nmeum: and is that wrong to do nowadays?
2018-07-09 18:36:03 <nmeum> I don't think that anybody came up with a better solution so far
2018-07-09 18:36:18 <azarus> ah, okay then. Thank you! :)
2018-07-09 18:43:15 <ebike> Anyone know if the port of  "Visual Studio Code" is actively being worked on?
2018-07-09 19:04:45 <azarus> ebike: not that I know of or heard of anyone working on it
2018-07-09 19:05:55 <azarus> just submitted my veracrypt port: https://patchwork.alpinelinux.org/patch/4049/
2018-07-09 19:07:10 <azarus> ebike: do you want to work on it?
2018-07-09 19:07:23 <azarus> (i could try my hand at it if you want)
2018-07-09 19:18:37 <ebike> I was hoping someone would take it on, it is the only app missing from Alpine that stops me using it for work ..
2018-07-09 19:19:57 <TBB> it's available in at least one container format, so you can use it on Alpine that way
2018-07-09 19:20:15 <ebike> example?
2018-07-09 19:21:22 <TBB> snap
2018-07-09 19:21:37 <ebike> huh?
2018-07-09 19:22:11 <TBB> the ubuntu application container format
2018-07-09 19:22:18 <TBB> vscode is distributed as a snap
2018-07-09 19:22:44 <ebike> oh, sorry, I don't know what snap is
2018-07-09 19:23:04 <ebike> i am used to docker
2018-07-09 19:23:34 <TBB> same kind of thing
2018-07-09 19:23:54 <ebike> ok, so where do I find the snap of VSC?
2018-07-09 19:24:38 <TBB> it's in https://snapcracft.io/vscode
2018-07-09 19:24:54 <ebike> Cheers :)
2018-07-09 19:24:56 <TBB> now, I don't know how to install that one, I have no experience of running snaps
2018-07-09 19:25:14 <azarus> ... wasn't even sure snaps run on alpine
2018-07-09 19:25:23 <TBB> my guess is, after you've installed snappy on Alpine you issue a command or two to install it
2018-07-09 19:26:15 <ebike> will try it out
2018-07-09 19:26:44 <ebike> @TBB THat link is dead
2018-07-09 19:29:40 <TBB> dead how?
2018-07-09 19:30:12 <ebike> server not found, also, snappy is listed as a "Fast compression and decompression library"
2018-07-09 19:30:58 <TBB> oh!
2018-07-09 19:31:14 <TBB> the link is active, but I may have messed up with the Alpine package!
2018-07-09 19:31:37 <ebike> I thing we need "snapd"?
2018-07-09 19:33:50 <TBB> that would be it; but that's probably not ported so far. I did notice a mention of ... ah, this: https://flathub.org/apps/details/com.visualstudio.code
2018-07-09 19:34:19 <TBB> that's an unofficial flatpak package of it; same concept, a bit different format, and flatpak is in Alpine repos
2018-07-09 19:35:20 <ebike> cool, I have tried ftatpak before
2018-07-09 19:40:03 <TBB> I have this suspicion, unfounded at the moment, that snapd just might require systemd
2018-07-09 19:40:25 <TBB> which is a bit sad considering the ecosystem
2018-07-09 19:41:01 <ebike> that would suck ..
2018-07-09 19:42:45 <TBB> okay, now it's a confirmed suspicion :/ ah well, flatpak it is then
2018-07-09 19:43:26 <ebike> yep, thanks for the help ... got to do some real work now .. bye!
2018-07-09 20:23:17 <azarus> anyone up for checking out my new patch? https://patchwork.alpinelinux.org/patch/4049/
2018-07-09 20:23:49 <azarus> (feel like as of late, not a single response to anything I sent to alpine...)
2018-07-10 02:47:02 <ebike> hey guys, anyone tried "Meld" lately? It installs but does not run
2018-07-10 02:48:36 <ebike> .. meld not found ..
2018-07-10 02:49:18 <ebike> it is there in /usr/bin/meld .. just does not run
2018-07-10 03:23:53 <tdtrask> fcolista, ncopa: please look at https://bugs.alpinelinux.org/issues/9057. It appears the upgrade of mini_httpd from 1.27 to 1.29 has broken ACF logon. I tested alpine 3.8 with mini_httpd pinned to 1.27, and everything worked fine. Looks like the logon actually occurs, but the ACF redirect is not sent by mini_httpd to the browser. Thanks.
2018-07-10 03:27:04 <ebike> anyone tried installing electron with npm? I get ..
2018-07-10 03:27:08 <ebike> npm ERR! code ELIFECYCLE npm ERR! errno 1 npm ERR! electron@2.0.4 postinstall: `node install.js` npm ERR! Exit status 1
2018-07-10 03:28:51 <ebike> npm ERR! electron@2.0.4 postinstall: `node install.js`
2018-07-10 03:37:12 <ebike> nevermind, I see I had to install it as root user, not just using sudo ..
2018-07-10 03:47:20 <ebike> anyone around?
2018-07-10 03:48:41 <ebike> just installed meld and electron (via npm) and neither app is found when i TYPE IT'S NAME OR FULL  PATH ..
2018-07-10 03:49:05 <ebike> oops sorry hit caps
2018-07-10 03:49:46 <ebike> other apps in /usr/bin are fine
2018-07-10 04:24:58 <ebike_> anyone help with installing electron via npm?
2018-07-10 09:22:34 <azarus> hrm, me and Marian Buschsieweke are frequently posting to alpine-aports@, yet never getting responses... is everyone else busy with other things?
2018-07-10 09:24:15 <skarnet> alpine-aports@ is mainly for patch submissions. You want alpine-devel@, I think.
2018-07-10 09:24:29 <azarus> and I do submit patches frequently on alpine-aports
2018-07-10 09:24:58 <skarnet> it's normal if those take time.
2018-07-10 09:36:02 <clandmeter> azarus, there is more activity in gh prs.
2018-07-10 09:36:35 <clandmeter> although recently all has low activity due to new release and summertime.
2018-07-10 09:36:42 <azarus> skarnet: sure, i understand, but this is kinda eh :/
2018-07-10 09:36:52 <azarus> clandmeter: yup
2018-07-10 09:37:20 <clandmeter> we need more capable developers who can handle patches.
2018-07-10 09:42:16 <azarus> agreed, waiting months for any sort of feedback is kinda frustrating
2018-07-10 09:44:03 <clandmeter> azarus, why do you choose  to send patches to the ml?
2018-07-10 09:44:37 <azarus> clandmeter: it's more convenient than gh, i guess
2018-07-10 09:44:48 <azarus> (and also slightly idealistic reasons tbh)
2018-07-10 09:45:19 <clandmeter> i can understand the second part
2018-07-10 09:45:54 <clandmeter> gh is more convenient for both end i think.
2018-07-10 09:46:11 <azarus> for the user, that's quite subjective
2018-07-10 09:46:35 <azarus> git-send-email is easier for me than bumbling around a web ui
2018-07-10 09:46:46 <azarus> but then again, for others that's different
2018-07-10 09:46:59 <clandmeter> you dont need to use the webif
2018-07-10 09:47:09 <awilfox> there's a cli script for github somewhere
2018-07-10 09:47:12 <azarus> i know, there's "hub" but ergh
2018-07-10 09:47:26 <awilfox> azarus: it's ok, we're just old :)  (I prefer git send-email too)
2018-07-10 09:47:32 <azarus> is it bad to no longer prefer github?
2018-07-10 09:48:09 <clandmeter> you have a choice so you should choose what you prefer.
2018-07-10 09:48:21 <clandmeter> with the cost of less activity.
2018-07-10 09:49:12 <azarus> I'm OK with that, but it's frustrating if I post something to the ml, later someone posts the same thing on gh, and the thing gets merged from gh
2018-07-10 09:49:28 <clandmeter> right
2018-07-10 09:50:33 <azarus> i hope I don't sound too demanding or something
2018-07-10 09:50:41 <clandmeter> not at all.
2018-07-10 09:50:52 <clandmeter> i understand your feelings.
2018-07-10 09:51:06 <azarus> thank you for understanding
2018-07-10 09:51:56 <clandmeter> azarus, if we would use something like gitea, would you mind switching?
2018-07-10 09:52:11 <azarus> clandmeter: if it gets activity, sure, why not
2018-07-10 09:52:25 <clandmeter> its web based similar like gh
2018-07-10 09:52:36 <azarus> I know, used it myself a bit
2018-07-10 09:52:56 <clandmeter> the ideal solution would be to use a single point of entry.
2018-07-10 09:52:58 <azarus> tbh, having a CI is nice, and lacking for the ml
2018-07-10 09:54:18 <clandmeter> I dont mind gh but i know others do. if we can find an alternative somewhere between ml and gh which everyone could accept that would be nice.
2018-07-10 09:55:29 <azarus> I'll do my best to be accepting and not act as a roadblock
2018-07-10 09:55:47 <clandmeter> ill go over a few of the pw patches
2018-07-10 09:55:52 <clandmeter> just to make you happy
2018-07-10 09:55:54 <clandmeter> :)
2018-07-10 09:56:01 <azarus> really, I don't want to appear needy :x
2018-07-10 09:56:03 <azarus> But thanks :D
2018-07-10 09:56:12 <clandmeter> im not saying ill merge them ;-)
2018-07-10 09:56:17 <azarus> hehe ;)
2018-07-10 09:57:16 <azarus> clandmeter: (i have 20 patches pending, pick and choose what you like best :P )
2018-07-10 09:57:29 <clandmeter> yes its a mess
2018-07-10 09:57:37 <clandmeter> i already regret looking at it ;-)
2018-07-10 09:58:09 <azarus> ctrl+w for a quick way out in most browsers
2018-07-10 10:00:12 <clandmeter> or just ctrl+t until i run out of memory
2018-07-10 10:00:44 <azarus> OOM killer, yay
2018-07-10 10:18:20 <mps> to add little to discussion, I posted my first aport to ML few weeks ago and waiting for any comment or critic answer
2018-07-10 10:18:48 <azarus> mps: that's https://patchwork.alpinelinux.org/patch/4015/ , right?
2018-07-10 10:19:08 <mps> in meantime I'm preparing few new packages and some patches
2018-07-10 10:19:12 <mps> azarus: yes
2018-07-10 10:19:44 <azarus> same ;)
2018-07-10 10:20:10 <mps> azarus: what? you posted this package also?
2018-07-10 10:20:18 <azarus> no, same situation for me
2018-07-10 10:20:26 <mps> aha, ok
2018-07-10 10:21:23 <mps> azarus: also I thought to ask questions which you just asked, but trying to be patiend
2018-07-10 10:21:36 <mps> s/patiend/patient/
2018-07-10 10:21:53 <azarus> mps: yup, I was trying to be patient too, but I suppose my patience was running a bit low
2018-07-10 10:22:27 <Gottox> s/patient/patience/
2018-07-10 10:22:32 <mps> also, noticed your (and Marian) posts because I subscribed to alpine-aports
2018-07-10 10:22:55 <mps> Gottox: I'm not native speaker, sorry
2018-07-10 10:23:09 <azarus> Gottox: nah, patience is the noun, patient the adjective
2018-07-10 10:23:44 <Gottox> ah, my mistake
2018-07-10 10:25:23 <mps> anyhow, but I posted just one aport to see is it ok and to get some advice from developers because don't want to 'polute' list with low quality aports
2018-07-10 10:26:29 <mps> and, to add note, I'm ready to maintain package which I post
2018-07-10 13:26:28 <nsh_> hi guys i've just noticed a bunch of my builds failing due to a repository untrusted signature
2018-07-10 13:26:30 <nsh_> ERROR: http://dl-cdn.alpinelinux.org/alpine/edge/main: UNTRUSTED signature
2018-07-10 13:26:39 <nsh_> fetch http://dl-cdn.alpinelinux.org/alpine/edge/main/armhf/APKINDEX.tar.gz
2018-07-10 13:26:57 <nsh_> is there something with wrong with the repo? doesnt seem to matter what repo mirror im using it fails
2018-07-10 13:27:40 <nsh_> im using alpine-devel@lists.alpinelinux.org-4a6a0840.rsa.pub alpine-devel@lists.alpinelinux.org-5243ef4b.rsa.pub alpine-devel@lists.alpinelinux.org-5261cecb.rsa.pub keys
2018-07-10 13:29:09 <nsh_> i can still build x86_64 sources though :S
2018-07-10 13:57:41 <clandmeter> nsh_, http://tpaste.us/YvDz
2018-07-10 14:00:39 <nsh_> mine is fine for x86_64 repos not armhf or aarch64  why would that be?
2018-07-10 14:09:20 <nsh_> what would the cause be if the repo is not faulty that the APKINDEX signature is not trusted when I have the keys? looking for places to go debug
2018-07-10 14:10:16 <clandmeter> nsh_, can you download those index files and check its contents?
2018-07-10 14:10:52 <clandmeter> nsh_, btw, this is our dev channel. alpine-linux is for support.
2018-07-10 14:13:24 <nsh_> contents .SIGN.RSA.alpine-devel@lists.alpinelinux.org-524d27bb.rsa.pub APKINDEX DESCRIPTION
2018-07-10 14:13:46 <nsh_> where is this key alpine-devel@lists.alpinelinux.org-524d27bb.rsa.pub? as it is not included in alpine
2018-07-10 14:16:12 <clandmeter> alpine-keys
2018-07-10 14:23:06 <nsh_> hmm some DH removed this from our container   Asante sana @clandmeter
2018-07-10 18:53:35 <ebike> Hi there, anyone got "electron" to work with muslc? I installed with npm but it won't run because it relies in libc
2018-07-10 18:54:52 <Xe> ebike: can you just compile electron from source?
2018-07-10 18:55:27 <ebike> I hav'nt tried compiling on Alpine yet, will give it a go ..
2018-07-10 18:55:31 <Xe> in general you might want to set up some kind of glibc container that has permission to your X socket
2018-07-10 18:55:43 <Xe> especially if this ends up being your daily driver machine
2018-07-10 18:56:11 <ebike> have no idea what you mean, but I will research
2018-07-10 18:56:48 <Xe> effectively an environment you can put all of the copies of google chrome
2018-07-10 18:57:23 <ebike> ok, what if I just install the flatpak version of electron, will that work?
2018-07-10 19:14:47 <stout762> Has anyone successfully booted v 3.8.0 on a RPI 3 model B+?
2018-07-10 19:22:28 <clandmeter> stout762: yes sir
2018-07-10 19:24:40 <stout762> clandmeter. Awsome, I'm trying but am clearly doing something stupid, I'm sure that I missed a step somewhere.
2018-07-10 19:25:24 <stout762> I burned and booted the latest raspbian lite to verify board, sd card, and console settings. That all went as expected.
2018-07-10 19:33:36 <stout762> clandmeter: I extracted the v 3.8.0 to a fat32 formatted sd card, set enable_uart=1, and tried to boot the device. I get no output on the console with either the armhf or aarch64 versions.
2018-07-10 19:34:31 <clandmeter> probably a partition issue
2018-07-10 19:34:49 <clandmeter> if the partition is ok it should load the kernel.
2018-07-10 19:34:56 <clandmeter> well first the bootloader
2018-07-10 19:36:16 <stout762> That's what I figured. This is what I have done. Using fdisk delete existing partitions, create new primary partition, set type to c W95 FAT32 (LBA), set boot flag.
2018-07-10 19:36:47 <stout762> I then format for FAT32: sudo mkfs.fat -F 32 /dev/sdh1
2018-07-10 19:37:34 <stout762> I then extract the tar ball: sudo tar xvf alpine-rpi-3.8.0-aarch64.tar.gz -C /mnt/usb --no-same-owner
2018-07-10 19:38:35 <stout762> Here is my fdisk table output: /dev/sdh1  *     2048 31116287 31114240 14.9G  c W95 FAT32 (LBA)
2018-07-10 19:50:39 <clandmeter> stout762: does your led blinkl?
2018-07-10 19:50:49 <clandmeter> do you have boot logo disabled?
2018-07-10 19:55:29 <stout762> I do not have a display attached, only the console. When I apply power: red power led comes on, then some activity of the green status led, then finally the orange network light comes on and occassionally blinks
2018-07-10 19:56:00 <stout762> The green activity led stays off after the initial activity.
2018-07-10 19:56:40 <stout762> and nothing is written to the console
2018-07-10 19:57:02 <clandmeter> i didnt try serial port.
2018-07-10 19:57:44 <clandmeter> did you add the correct console option to bootline?
2018-07-10 19:59:31 <stout762> I create the usercfg.txt file and added enable_uart=1 to it
2018-07-10 20:00:10 <stout762> I also have the ethernet cable connected and do not see the device requesting an IP address
2018-07-10 20:02:12 <clandmeter> stout762: networking is not started by default
2018-07-10 20:02:55 <stout762> clandmeter: ok
2018-07-10 20:03:47 <stout762> clandmeter: should I dig up an HDMI display and try that? In case the console is working?
2018-07-10 20:04:38 <clandmeter> thats one think you can try
2018-07-10 20:05:27 <clandmeter> stout762: you can add alpine_repo=http://dl-cdn.alpinelinux.org/alpine/v3.8/main to cmdline
2018-07-10 20:05:32 <clandmeter> that should init dhcp
2018-07-10 20:09:33 <stout762> here is my cmdline.txt: modules=loop,squashfs,sd-mod,usb-storage quiet dwc_otg.lpm_enable=0 console=tty1
2018-07-10 20:09:50 <stout762> Do I just change it to: modules=loop,squashfs,sd-mod,usb-storage alpine_repo=http://dl-cdn.alpinelinux.org/alpine/v3.8/main quiet dwc_otg.lpm_enable=0 console=tty1
2018-07-10 20:09:57 <clandmeter> you need to modify it to support serial console.
2018-07-10 20:10:14 <clandmeter> yes just append al;pine_repo
2018-07-10 20:10:15 <stout762> Ok, how do I do that?
2018-07-10 20:10:39 <clandmeter> it depends on which rpi/arch
2018-07-10 20:10:44 <clandmeter> check rpi documentation.
2018-07-10 20:12:25 <clandmeter> on newer rpi's bluetooth uses the serial port thats why i disabled enable_uart
2018-07-10 20:14:33 <stout762> clandmeter: head -> desk! I added console=serial0,115200 to cmdline.txt and I am now logged in via the console. Thank you so much for your help.
2018-07-11 06:28:35 <StarWarsFan> hi@all
2018-07-11 06:28:47 <StarWarsFan> just a question to the developers:
2018-07-11 06:29:13 <StarWarsFan> how can i define specific dependencies on APKBUILD for separate alpine releases?
2018-07-11 06:29:47 <StarWarsFan> like for the build on 3.7 the dependency is i. e. libressl2.6-libcrypto
2018-07-11 06:29:58 <StarWarsFan> but for 3.8 it is libressl2.7-libcrypto
2018-07-11 07:02:11 <awilfox> 'libressl-dev' is probably what you want
2018-07-11 07:15:40 <StarWarsFan> hm...
2018-07-11 07:23:37 <StarWarsFan> but what is wrong if i got this during build of apkindex:
2018-07-11 07:23:52 <StarWarsFan> WARNING: No provider for dependency 'so:libcrypto.so.1.0.0'
2018-07-11 07:30:11 <awilfox> libcrypto.so.1.0.0 would be openssl
2018-07-11 07:30:13 <awilfox> try openssl-dev
2018-07-11 08:44:03 <StarWarsFan> that lead into other probs as other depenendencies require libressl
2018-07-11 09:11:12 <StarWarsFan> besides this problem:
2018-07-11 09:11:33 <StarWarsFan> why doeas index build report this:
2018-07-11 09:12:34 <StarWarsFan> ...
2018-07-11 09:12:34 <StarWarsFan> WARNING: No provider for dependency 'so:libcrypto.so.1.0.0'
2018-07-11 09:12:34 <StarWarsFan> WARNING: No provider for dependency 'dmidecode'
2018-07-11 09:12:34 <StarWarsFan> WARNING: No provider for dependency 'wget'
2018-07-11 09:12:34 <StarWarsFan> WARNING: Too many unsatisfiable dependencies, not reporting the rest.
2018-07-11 09:12:37 <StarWarsFan> WARNING: Total of 108 unsatisfiable package names. Your repository may be broken.
2018-07-11 09:13:03 <awilfox> ok, so that means your custom repository doesn't have packages that provide those dependencies
2018-07-11 09:13:07 <awilfox> that's fine, because aports/main does
2018-07-11 09:13:12 <awilfox> you don't need to provide those
2018-07-11 09:13:39 <StarWarsFan> ah ok, so i don't need to worry about that?
2018-07-11 09:14:04 <awilfox> correct
2018-07-11 09:14:11 <StarWarsFan> nice, thx
2018-07-11 09:14:37 <StarWarsFan> but how can this happen:
2018-07-11 09:14:39 <StarWarsFan> ERROR: cuimenu-2.0.3-r11: package mentioned in index not found (try 'apk update')
2018-07-11 09:14:53 <StarWarsFan> the package is there...
2018-07-11 09:15:40 <StarWarsFan> you can see on https://web.nettworks.org/e-ng/packages/edge/testing/x86_64/
2018-07-11 09:23:50 <awilfox> that's odd
2018-07-11 09:24:18 <azarus> StarWarsFan: is the apkindex up to date?
2018-07-11 09:24:29 <StarWarsFan> yes
2018-07-11 09:24:32 <azarus> hmm
2018-07-11 09:24:41 <StarWarsFan> removed it and created again
2018-07-11 09:47:38 <StarWarsFan> ah, got it!
2018-07-11 09:47:44 <StarWarsFan> strange thing...
2018-07-11 09:47:54 <StarWarsFan> my last change was to fix the following warning:
2018-07-11 09:48:07 <StarWarsFan> WARNING: cuimenu*: No arch specific binaries found so arch should probably be set to "noarch"
2018-07-11 09:48:34 <StarWarsFan> so i increased pkgrel and set arch="noarch"
2018-07-11 09:48:50 <StarWarsFan> no the installation fails with the error above
2018-07-11 09:49:04 <StarWarsFan> so i reverted the arch-change
2018-07-11 09:49:07 <StarWarsFan> et voila
2018-07-11 09:49:20 <StarWarsFan> interesting huh!?
2018-07-11 22:31:39 <mickibm> for a patch in mkinitfs, which would be the reliable way to get the ARCH? i need to submit a patch and am looking at line 627 in initramfs-init.in
2018-07-11 22:32:04 <mickibm> cat /etc/apk/arch or just always do the $(apk --print-arch)
2018-07-11 22:32:21 <mickibm> for a patch for ppc64le...
2018-07-11 22:33:04 <mickibm> cause i think having an ARCH variable would be helpful if im gonna be checking the arch anyways
2018-07-11 22:47:55 <mickibm> im just adding a archfile at the top... will submit PR shortly
2018-07-11 23:17:27 <mickibm> for the curious: https://github.com/alpinelinux/mkinitfs/pull/32
2018-07-12 10:20:05 <ncopa> hi, im back from PTO
2018-07-12 10:20:09 <ncopa> how are things going?
2018-07-12 10:22:01 <ncopa> rdutra: looks like you have upgraded xfce to 4.13, which is a gtk3 based upgrade
2018-07-12 10:22:28 <ncopa> and is a development branch
2018-07-12 10:22:40 <skarnet> wb, hope you enjoyed your vacation :)
2018-07-12 10:23:16 <rdutra> ncopa: wb, yeah, I merged some patches to upgrade xfce to 4.13 and upgraded the other packages that were missing
2018-07-12 10:23:30 <ncopa> have you checked with upstream xfce devs that they will be able to release xfce 4.14 before our next stable release? i dont think we should ship alpine 3.9 with xfce 4.13
2018-07-12 10:24:13 <rdutra> ncopa: no, I didn't know it is a development branch...
2018-07-12 10:24:15 <ncopa> skarnet: thanks! it was very nice to have a couple of weeks off
2018-07-12 10:25:27 <ncopa> like gnome and mate and cinnamon(?) the odd version numbers are development (4.9, 4.11, 4,13) and even version numbers are stable (4.10, 4.12, 4.14)
2018-07-12 10:26:43 <ncopa> my edge desktop does not upgrade for some reason: http://tpaste.us/Jgm7
2018-07-12 10:26:46 <rdutra> ncopa: hmm, I thought it was just for Gnome..my mistake
2018-07-12 10:27:32 <ncopa> it may not matter, if they manage to get 4.14 out before October/November this year
2018-07-12 10:27:44 <ncopa> but thye have traditionally been very slow
2018-07-12 10:29:30 <rdutra> ncopa: I also upgraded Mate packages to last released versions (1.21) :-/ and fixed a bug that was opened for some time regarding one of mate packages
2018-07-12 10:29:46 <ncopa> and i had the impression that gtk3 xfce is still work in progress
2018-07-12 10:29:46 <rdutra> they seems to be working fine
2018-07-12 10:30:22 <ncopa> i think mate follows gnome release cycle which is every 6 month
2018-07-12 10:31:02 <ncopa> $ apk info -R xfce4-appfinder | tpaste
2018-07-12 10:31:02 <ncopa> http://tpaste.us/8xD6
2018-07-12 10:32:00 <rdutra> ncopa: libxfconf was upgraded, so I rebuild xfce packages that depends on it
2018-07-12 10:32:16 <ncopa> looks like xfce4-appfinder 4.12 depends on so:libxfce4ui-1.so.0 so:libxfconf-0.so.2 and xfce4-appfinder 4.13 depends on so:libxfce4ui-2.so.0 so:libxfconf-0.so.3
2018-07-12 10:32:29 <ncopa> i suspect something was not rebuilt, since i cannot upgrade
2018-07-12 10:44:36 <ncopa> clandmeter: community/nbd does not build:
2018-07-12 10:44:38 <ncopa> nbd-client.c:51:29: fatal error: netlink/netlink.h: No such file or directory
2018-07-12 10:44:38 <ncopa>  #include <netlink/netlink.h>
2018-07-12 10:44:47 <clandmeter> check commits
2018-07-12 10:45:26 <clandmeter> alpine-commits :)
2018-07-12 10:53:43 <ncopa> anyone have a clue how to figure out that is holding back the `apk upgrade -U -a`? http://tpaste.us/Jgm7
2018-07-12 11:09:42 <ncopa> proj4 is due to mapnik needs to be rebuilt
2018-07-12 11:11:12 <ncopa> there is something related xfce4 that needs to be rebuilt
2018-07-12 11:24:29 <ncopa> rdutra: can you please ask the xfce devs if it is realistic that xfce 4.14 comes out before our alpine 3.9 release? (Oct-Nov something)
2018-07-12 11:25:29 <ncopa> we need to either complete the xfce4 upgrade and rebuild some packages (exo libgarcon libxfce4util upgrades seems to require rebuild of various packages)
2018-07-12 11:25:36 <ncopa> or revert to 4.12
2018-07-12 11:26:49 <rdutra> ncopa: ok, I can do that
2018-07-12 11:29:53 <ncopa> it looks like they dont recommend packagers to upgrade to 4.13 yet: https://mail.xfce.org/pipermail/xfce4-dev/2018-July/032139.html
2018-07-12 11:36:01 <tya99> ncopa: u there?
2018-07-12 11:36:11 <ncopa> tya99: im here. how are you doin?
2018-07-12 11:36:26 <tya99> i'm about to log a bug i was wondering if you could help me find out what the originating factor of this bug is
2018-07-12 11:36:28 <tya99> in short
2018-07-12 11:36:36 <tya99> if you run setup-xorg-base in a virtual machine
2018-07-12 11:36:51 <tya99> it installs eudev and runs setup-udev which then on reboot means you cannot find your network interface
2018-07-12 11:37:04 <tya99> configured by setup-alpine ie in /etc/networking/interfaces
2018-07-12 11:37:19 <tya99> im logging a bug now but it would be nice to find out what eudev does when it runs the first time that causes this
2018-07-12 11:37:30 <tya99> setup-udev removes mdev and starts udev
2018-07-12 11:37:58 <tya99> i have reproduced it about 3 times so i know the steps for certain
2018-07-12 11:38:13 <ncopa> ok. what vm is it?
2018-07-12 11:40:01 <tya99> virtualbox
2018-07-12 11:40:23 <ncopa> alpine-virt or alpine-vanilla?
2018-07-12 11:40:32 <ncopa> i mean alpine-standard
2018-07-12 11:41:18 <tya99> ah alpine-extended
2018-07-12 11:41:53 <tya99> that one :P alpine-extended-3.8.0-x86_64.iso
2018-07-12 11:42:04 <tya99> i just opened a bug
2018-07-12 11:42:10 <ncopa> i saw :)
2018-07-12 11:42:23 <ncopa> very good. and good that you explain how to reproduce it
2018-07-12 11:42:38 <tya99> cuz i wont it fixed ;D
2018-07-12 11:42:51 <tya99> if i didn't say how to make it happen again good luck finding the details of it
2018-07-12 11:43:13 <tya99> but yeah i don't understand what happens when udev runs for the first time after mdev is removed to know any more about why that happens
2018-07-12 11:43:41 <ncopa> can you edit the issue and add the architecture (x86_64) and that use use alpine-extended?
2018-07-12 11:46:25 <tya99> ncopa: i dont think i can edit that i can only add further comments
2018-07-12 11:48:28 <tya99> clicking that edit ting just brings up a comments textbox
2018-07-12 11:54:09 <ncopa> add a comment
2018-07-12 11:54:11 <ncopa> its ok
2018-07-12 11:54:18 <tya99> i have done that
2018-07-12 11:56:39 <tya99>  if you want me to try anything else i can
2018-07-12 12:04:29 <ncopa> i cannot reproduce :-(
2018-07-12 12:04:50 <ncopa> tya99: can you check that udev-trigger is started in sysint runlevel?
2018-07-12 12:05:04 <tya99> how do i do that
2018-07-12 12:05:14 <tya99> sorry if that sounds stupid :P
2018-07-12 12:05:14 <ncopa> rc-update | grep udev
2018-07-12 12:05:23 <ncopa> shoul give you 3 lines
2018-07-12 12:05:25 <tya99> ah k
2018-07-12 12:05:31 <ncopa> udev in sysinit
2018-07-12 12:05:35 <ncopa> udev-postmount in default
2018-07-12 12:05:44 <ncopa> and udev-trigger in sysinit
2018-07-12 12:06:31 <tya99> virtualmachine:/home/user# setup-udev
2018-07-12 12:06:33 <tya99>  * Caching service dependencies ... [ ok ]
2018-07-12 12:06:35 <tya99>  * Starting udev ... [ ok ]
2018-07-12 12:06:37 <tya99>  * Generating a rule to create a /dev/root symlink ... [ ok ]
2018-07-12 12:06:39 <tya99>  * Populating /dev with existing devices through uevents ... [ ok ]
2018-07-12 12:06:41 <tya99>  * Waiting for uevents to be processed ... [ ok ]
2018-07-12 12:07:02 <tya99> ncopa: yes it does show them
2018-07-12 12:07:34 <ncopa> do you have linux-firmware installed?
2018-07-12 12:07:39 <ncopa> apk version linux-firmware
2018-07-12 12:08:54 <tya99> no it appears i do not
2018-07-12 12:09:44 <tya99> i shall install it and try again (rolled back snapshot)
2018-07-12 12:10:27 <ncopa> do you install to disk or run "diskless"?
2018-07-12 12:10:34 <tya99> diskless
2018-07-12 12:10:47 <ncopa> how do you save your config?
2018-07-12 12:10:51 <tya99> yes
2018-07-12 12:10:58 <ncopa> where do you save it?
2018-07-12 12:11:13 <tya99> in /media/sda1
2018-07-12 12:11:22 <ncopa> which is a virtual disk?
2018-07-12 12:11:27 <tya99> yes
2018-07-12 12:11:42 <ncopa> ok, that kind of info may be relevant
2018-07-12 12:11:46 <ncopa> to reproduce
2018-07-12 12:11:57 <tya99> i have tried a number of different configurations
2018-07-12 12:12:01 <tya99> and have been able to reproduce it
2018-07-12 12:12:10 <ncopa> how much memory have you given your vm?
2018-07-12 12:12:14 <tya99> such as booting from sr0, and saving config in sda1 and cache in sdb1
2018-07-12 12:12:36 <tya99> 1024MB
2018-07-12 12:12:51 <tya99> i will try with linux-firmware installed
2018-07-12 12:12:56 <ncopa> no
2018-07-12 12:13:00 <ncopa> it will not make any difference
2018-07-12 12:13:06 <tya99> oh :P
2018-07-12 12:13:11 <ncopa> and i dont think you will be able to install it
2018-07-12 12:13:19 <tya99> yeah probably not
2018-07-12 12:16:23 <tya99> i have updated the bug with those details
2018-07-12 12:16:53 <tya99> i will try it again
2018-07-12 12:17:24 <tya99> ncopa: i installed it by simple formatting /dev/sda1 with fdisk
2018-07-12 12:17:35 <tya99> and then running
2018-07-12 12:17:37 <tya99>     setup-bootable /media/cdrom /dev/sda1
2018-07-12 12:17:47 <ncopa> ok
2018-07-12 12:17:50 <tya99> then once booted i ran setup-alpine
2018-07-12 12:18:15 <tya99> the reason was because i only planned to keep the apkprov on sda1 and the cache on sdb1
2018-07-12 12:18:39 <tya99> that way to upgrade in the future, all i'd have to do is insert a new CD image into the VM
2018-07-12 12:18:57 <tya99> (and obviously fix any errors that occured during the upgrade like config location changes etc)
2018-07-12 12:20:04 <tya99> the fact is everything is working up until the point i install udev and run setup-udev
2018-07-12 12:20:11 <tya99> (i rolled back the VM's snapshot)
2018-07-12 12:22:16 <tya99> ie sda1 contains https://dpaste.de/OpVn
2018-07-12 12:22:52 <ncopa> do you run dd if=/usr/share/syslinux/mbr.bin of=/dev/sda ?
2018-07-12 12:22:59 <ncopa> my disk image is not bootable...
2018-07-12 12:23:08 <ncopa> i get Missing operating system
2018-07-12 12:23:08 <tya99> no i made it bootable with fdisk
2018-07-12 12:23:17 <ncopa> oh ok...
2018-07-12 12:23:21 <ncopa> that is what i forgot
2018-07-12 12:23:21 <tya99> ie https://wiki.alpinelinux.org/wiki/Create_a_Bootable_USB#Format_USB_stick
2018-07-12 12:23:26 <tya99> :)
2018-07-12 12:23:50 <tya99> then i ran
2018-07-12 12:23:52 <tya99>     setup-bootable /media/cdrom /dev/sda1
2018-07-12 12:24:02 <tya99> after booting i ran setup-alpine
2018-07-12 12:24:07 <tya99> then lbu commit
2018-07-12 12:24:11 <tya99> then rebooted all is good
2018-07-12 12:24:19 <tya99> then soon as i apk add eudev and run setup-udev
2018-07-12 12:24:30 <tya99> not so good (course one must lbu commit again or that change will be lost
2018-07-12 12:27:19 <ncopa> ok, i think i ca reproduce now
2018-07-12 12:29:01 <ncopa> and i have a fix for it
2018-07-12 12:36:34 <ncopa> tya99: does this fix it for you? echo "rc_need=udev-settle" > /etc/conf.d/networking && lbu ci
2018-07-12 12:50:40 <tya99> one sec will try
2018-07-12 12:54:11 <ncopa> rdutra: i think we go for xfce 4.13 now
2018-07-12 12:54:17 <tya99> ncopa: yes it does fix it
2018-07-12 12:54:18 <ncopa> xubuntu seems to do the same
2018-07-12 12:54:49 <tya99> ncopa: u are my hero!
2018-07-12 13:08:10 <rdutra> ncopa: :) good
2018-07-12 15:09:13 <tya99> hmm if i wanted to install virtualbox-guest-modules-vanilla ncopa  and i was getting this error on diskless https://dpaste.de/w9Rr
2018-07-12 15:09:43 <tya99> course that is in .modloop which is ro
2018-07-12 15:10:13 <tya99> i think that might only work in a sys install
2018-07-12 15:57:15 <tya99> i might have to do a sys install i don't think diskless was really designed to work in a virtual machine
2018-07-13 06:27:48 <clandmeter> wow, nice bug report of mr anonymous...
2018-07-13 06:47:00 <mickibm> ha
2018-07-13 09:05:02 <nshCore> can anyone tell give a bit more information how abuild traces dependencies to declare in an APKINDEX, im getting a bunch of errors when trying to build a rpi profile but the same code building x86_64 archs is fine. it complains about ERROR: unsatisfiable constraints:   so:libc.musl-x86_64.so.1 (missing):     required by: alpine-base-layout
2018-07-13 09:06:16 <nshCore> i see the dep defined as this D: so:libc.musl-x86_64.so.1 in the APKINDEX why not explicitly state it needs the musl package?
2018-07-13 09:08:09 <clandmeter> nshCore, what is the cmd you run to gen the img?
2018-07-13 09:08:53 <nshCore> 		sh -c "./scripts/mkimage.sh \ 			--tag edge \ 			--outdir /home/builder/iso \ 			--repository http://dl-cdn.alpinelinux.org/alpine/edge/main \ 			--extra-repository http://dl-cdn.alpinelinux.org/alpine/edge/community \ 			--extra-repository http://dl-cdn.alpinelinux.org/alpine/edge/testing \ 			--arch armhf \ 			--profile rpi"
2018-07-13 09:12:08 <clandmeter> nshCore, seems about right
2018-07-13 09:13:31 <clandmeter> nshCore, ah
2018-07-13 09:13:42 <clandmeter> nshCore, run the script from scripts dir
2018-07-13 09:14:58 <nshCore> hmm not sure it should make a difference like i said the same command changing the arch to x86_64 and profile to standard works fine running from the same location
2018-07-13 09:15:22 <nshCore> its actually just all mounted inside a alpine:edge container and built from there
2018-07-13 09:21:40 <nshCore> it doesnt seem to make sense that alpine-baselayout or any package can define a so as a dep like D: so:libc.musl-x86_64.so.1 and resolve it how would it know which pkg its actually in? and in the case of musl where the name of the actual .SO changes depending on the arch your building for
2018-07-13 09:23:17 <clandmeter> nshCore, please try from script dir, if thats doesnt work let me know.
2018-07-13 09:28:52 <nshCore> no same thing
2018-07-13 09:40:25 <clandmeter> nshCore, can you run the script directly? ./mkimage.sh
2018-07-13 09:40:34 <clandmeter> if i run it with sh -c i also get errors
2018-07-13 09:47:05 <nshCore> I have to do sh -c"command" as its running a docker container everything inside the " " is executed  in the context of alpine
2018-07-13 09:55:39 <clandmeter> sorry I have no experience with docker.
2018-07-13 10:00:31 <nshCore> it shouldnt matter if its docker/lxc or chroot its executing in the alpine env at that point it doesnt care who/what/why created the env its already inside it, Im not sure why a armhf arch build is looking for an x86_64 dep, when the APKINDEX of the edge repo states the  dep as D:/bin/sh so:libc.musl-armhf.so.1
2018-07-13 10:01:40 <nshCore> its getting confused over what arch is building which why it would make more sense to explicitly state it wants musl dep rather than a .so so it can go grab it from the correct arch repo
2018-07-13 11:43:01 <azarus> (shameless plug) this patch fixes javascript on webkit, and thus fixes issue #8492: https://patchwork.alpinelinux.org/patch/4056/
2018-07-13 13:30:36 <ncopa> does anybody know what the proper way to package go projects?
2018-07-13 13:30:45 <ncopa> I have a tiny single aports-ghpr.go
2018-07-13 13:30:58 <ncopa> https://git.alpinelinux.org/cgit/user/ncopa/aports-ghpr/tree/aports-ghpr.go
2018-07-13 13:31:23 <ncopa> and a makefile that will run: go build -v aports-ghpr.go
2018-07-13 13:31:26 <ncopa> https://git.alpinelinux.org/cgit/user/ncopa/aports-ghpr/tree/Makefile
2018-07-13 13:32:09 <ncopa> but I get the following error: aports-ghpr.go:10:2: cannot find package "github.com/google/go-github/github" in any of:
2018-07-13 13:32:22 <ncopa> i think `go get` fixes it
2018-07-13 13:32:32 <ncopa> but i dont think that is the proper way to do it
2018-07-13 14:02:04 <azarus> thanks ncopa for merging my webkit patch
2018-07-13 14:02:11 <azarus> :)
2018-07-13 14:02:12 <ncopa> np
2018-07-13 14:02:23 <azarus> (well "my" -- really, gentoo devs and TBK)
2018-07-13 14:03:07 <TBK[m]> and openembedded :P
2018-07-13 14:03:26 <azarus> oh, dammit, should've bumped pkgrel
2018-07-13 14:04:03 <azarus> and the pkgrel of all packages linking against webkit
2018-07-13 14:04:06 <TBK[m]> I will push a PR, the APKBUILD needs to be modanized
2018-07-13 14:04:13 <azarus> ah yup that too
2018-07-13 14:04:26 <azarus> thanks!
2018-07-13 14:06:09 <azarus> ugh, i always mess this up
2018-07-13 14:07:01 <TBK[m]> create a patchworks script that asks you a bunch of checklist questions :P
2018-07-13 14:07:58 <azarus> i should! been sending too many *wrong* things
2018-07-13 14:08:21 <azarus> "Are you ***sure*** you want to send this?? [y/n] "
2018-07-13 15:13:12 <clandmeter> ncopa, i think currently the suggested tool is go dep.
2018-07-13 16:05:00 <sphrak> andypost: hello, Im a bit confused by PR #4686, is it merged.. or not merged? because github kinda says both
2018-07-13 16:05:26 <sphrak> no not the bug
2018-07-13 16:06:51 <andypost> sphrak, merged https://git.alpinelinux.org/cgit/aports/commit/?id=3dfcbcd5e145ccd6a578273820fc51cd97fa4ba2
2018-07-13 16:06:52 <sphrak> nvm, I can see it in the package list now.-..
2018-07-13 16:07:13 <sphrak> andypost: ah yes, thanks you:)
2018-07-13 16:09:00 <andypost> sphrak, next time please provide a link to changelog or release notes
2018-07-13 16:13:44 <TBK[m]> andypost:
2018-07-13 16:14:04 <TBK[m]> andypost: was it reported through pkgs.al.org?
2018-07-13 16:23:19 <rdutra> ncopa: regarding your package (https://github.com/alpinelinux/aports/commit/d102e6e0f490f27a3d320f62f1d04e84b5facbc0)
2018-07-13 16:23:19 <rdutra> I'm also curious about the best way to solve dependencies in go packages
2018-07-13 16:27:42 <TBK[m]> Glide is added as a makedepends so why not use it "glide init && glide get github.com/google/go-github/github"
2018-07-13 16:27:44 <sphrak> andypost: will do, this was my first contribution
2018-07-13 17:00:39 <andypost> TBK[m], what do you mean about report?
2018-07-13 17:02:08 <TBK[m]> andypost: I misunderstod. I thought you updated the pkg based on a pkg.al.org flag report.
2018-07-13 17:02:46 <andypost> TBK[m], ah no, it was a PR
2018-07-13 19:11:27 <azarus> TBK[m]: please see my new patches to rebuild against webkit2gtk-2.20.3-r1
2018-07-13 19:11:40 <azarus> (beginning with https://patchwork.alpinelinux.org/patch/4060/ )
2018-07-14 14:46:17 <agowa338> Is the mailinglist broken? When trying to subscribe, the server returns "451 4.7.1 Service unavailable"
2018-07-14 15:45:42 <agowa338> and does the same when trying to send a message to the list...
2018-07-15 17:29:46 <vkrishn> seems a nice simple lib, https://equi4.com/metakit/ , would be nice to have in AL if possible
2018-07-15 17:30:27 <_ikke_> vkrishn: You could try to make an apkbuild for it
2018-07-16 10:44:46 <ncopa> morning
2018-07-16 10:44:53 <ncopa> we need rust on all our arches
2018-07-16 10:44:58 <ncopa> https://mail.gnome.org/archives/desktop-devel-list/2017-January/msg00001.html
2018-07-16 10:46:08 <ncopa> sigh... befre weekend the PR queue was ~230, its again up to 260
2018-07-16 10:46:47 <_ikke_> :-(
2018-07-16 10:47:19 <_ikke_> 30 PRs in 2 days
2018-07-16 10:48:55 <grJotQbLTxV5> it's good that people are actively contributing. it's bad that the ones able to approve the contributions are overwhelmed with the amount.
2018-07-16 11:01:39 <rdutra> ncopa: I think mksully22 opened a PR to enable rust on ppc64le. Not sure if jirutka took a look at it
2018-07-16 11:09:26 <rdutra> ncopa: this one https://github.com/alpinelinux/aports/pull/4549. but I'm not an expert in rust
2018-07-16 11:32:19 <Ganwell> ncopa: I've been testing packages for arm with a mixture of cross-compiling and qemu-arm-static. It's quite slow and error-prone. Is there a better way? (for trying to build rust on arm)
2018-07-16 11:45:30 <clandmeter> Ganwell, there probably is.
2018-07-16 11:45:35 <clandmeter> im thinking how.
2018-07-17 13:38:16 <fhunleth> Hi! I submitted my first package yesterday and was wondering what the frequency of building non-x86_64 packages was. If there's a build or test error on, say, ppc64le, will I get a notification?
2018-07-17 13:39:42 <rnalrd> hi fhunleth
2018-07-17 13:39:51 <rnalrd> we have buildlogs here: http://build.alpinelinux.org/buildlogs/
2018-07-17 13:40:40 <rnalrd> also on #alpine-commits there's the build output result after each commit to master tree
2018-07-17 13:44:27 <fhunleth> rnalrd: thanks! I looked up the aarch64 build for fwup and it looks like it completed successfully. It doesn't show up at https://pkgs.alpinelinux.org/packages?name=fwup&branch=edge yet. Not that it needs to, but is there something else that has to happen? The x86_64 package showed up surprisingly quickly yesterday.
2018-07-17 13:49:16 <fhunleth> Also, some users of the package told me that they'd highly prefer if the package were in "community" or "main". That makes sense, but I was wondering what the etiquette was for doing this. I just submitted the package yesterday. Should I wait a week (or month) for any dust to settle before sending a PR up to move it?
2018-07-17 14:11:07 <rnalrd> fhunleth, usually a package goes into main or community after a period of testing, it should be known that it's working well
2018-07-17 14:11:39 <rnalrd> in main goes only "core" package or package with good support from upstream and have stable supported releases
2018-07-17 14:11:47 <rnalrd> everything else goes into community
2018-07-17 14:12:06 <rnalrd> fhunleth, what's the package?
2018-07-17 14:12:37 <fhunleth> fwup
2018-07-17 14:13:06 <rnalrd> it's all there
2018-07-17 14:13:13 <rnalrd> for all arches !s390x
2018-07-17 14:17:21 <fhunleth> Great! A few people are using it right now, so I'll check in with them periodically to see how it goes. Thanks again for the answers.
2018-07-18 09:54:30 <trfl> I am building a program which produces one single binary frontend for its one single .so library, however this library is 1.4 GB when built with -g
2018-07-18 09:55:07 <trfl> this does get stripped properly so only the -dbg apk becomes unreasonably large at 600MB, is this really desirable?
2018-07-18 10:21:07 <trfl> s/building/packaging/
2018-07-18 11:35:03 <ncopa> would be nice to avoid a 600MB -dbg package if possible
2018-07-18 11:40:17 <clandmeter> ncopa, www still not updating...
2018-07-18 11:52:22 <ncopa> ok, i'm on it
2018-07-18 11:52:37 <ncopa> where did alpine-www move? to nl3.a.o?
2018-07-18 11:52:52 <ncopa> what is the new lxc name?
2018-07-18 11:54:44 <ncopa> seems like www.a.o is pointing to old 88.159.20.184
2018-07-18 11:55:01 <ncopa> so it is still at gigabyte
2018-07-18 11:55:18 <ncopa> but the hostname it used, nl.alpinelinux.org is pointing to new location
2018-07-18 11:55:29 <ncopa> that is why it is not updating
2018-07-18 11:55:34 <ncopa> it tries to upload to nl.a.o
2018-07-18 11:56:02 <clandmeter> www has moved
2018-07-18 11:56:06 <ncopa> to where?
2018-07-18 11:56:22 <clandmeter> to 8.x
2018-07-18 11:56:50 <clandmeter> could be i forgot to update the dns for www
2018-07-18 11:57:12 <clandmeter> but it will use the proxy to redirect via vpn
2018-07-18 11:57:35 <clandmeter> 172.16.8.8 is new www
2018-07-18 11:57:58 <clandmeter> would be nice if we dont have to forward ports
2018-07-18 11:58:03 <ncopa> which is not routeable from the bld1*
2018-07-18 11:58:16 <clandmeter> yes i cannot help as i dont have access :)
2018-07-18 11:58:23 <clandmeter> what is the config on that wwwbuild?
2018-07-18 11:58:30 <clandmeter> i can move it to www
2018-07-18 11:58:49 <clandmeter> just mqtt-exec to exec makefile
2018-07-18 12:01:10 <ncopa> i copied wwwbuild.tar.gz to 172.16.8.1:
2018-07-18 12:01:19 <clandmeter> cool thx
2018-07-18 12:01:19 <ncopa> which is the container as is
2018-07-18 12:02:11 <clandmeter> let me check the dns.
2018-07-18 12:02:32 <clandmeter> ncopa, we need latest mqtt-exec with auth patch in all branches.
2018-07-18 12:02:34 <clandmeter> can you do that?
2018-07-18 12:03:12 <clandmeter> i wanted to set it up yesterday, but for older builders it wouldnt be possible.
2018-07-18 12:03:25 <clandmeter> and the auth patch isnt even in edge.
2018-07-18 12:03:41 <ncopa> ok, i will look at that
2018-07-18 12:04:16 <clandmeter> when its in all brances me and ikke will add auth to mqtt for builders.
2018-07-18 12:04:40 <clandmeter> except s390x
2018-07-18 12:04:43 <clandmeter> you will have to do that.
2018-07-18 12:05:31 <ncopa> ok
2018-07-18 12:06:54 <clandmeter> and its probably the right time to disable regular mqtt in favor of mqtt(tls) and mqtt-ws(tls)
2018-07-18 12:10:24 <ncopa> i pushed password auth to edge
2018-07-18 12:10:33 <ncopa> can you please test that it works before i backport it?
2018-07-18 12:16:53 <ncopa> ok, all is set, i just need confirmation that it works before i push to all supported branches
2018-07-18 12:23:13 <clandmeter> i think we already tested it
2018-07-18 12:23:18 <clandmeter> i use it in backup
2018-07-18 14:08:41 <ncopa> i have backported mqtt-exec to all stable branches
2018-07-18 14:09:04 <ncopa> i think weneed refactor the buildrepo hooks
2018-07-18 14:09:45 <ncopa> the lua-mosquitto-publish does not yet support tls or password
2018-07-18 14:34:14 <ncopa> clandmeter: so what do you need on the builders to make them be able to push messages?
2018-07-18 14:34:19 <ncopa> certs?
2018-07-18 14:34:22 <ncopa> ca cert?
2018-07-18 14:34:48 <ncopa> i guess that with certs we dont need to manage passwords or users?
2018-07-18 14:40:05 <clandmeter> We do if we want acl
2018-07-18 14:41:01 <clandmeter> We use acme cert
2018-07-18 14:42:26 <clandmeter> Else we need to use our own ca which is less flexible
2018-07-18 14:42:48 <ncopa> so no client certs i assume
2018-07-18 14:43:28 <clandmeter> Correct
2018-07-18 14:44:34 <clandmeter> You can use dn with client certs. But anyone who wants to use it will need one
2018-07-18 14:45:11 <clandmeter> Or use separate certs for ws and non ws
2018-07-18 14:45:39 <ncopa> is there some way i can test if tls is working?
2018-07-18 14:45:46 <danieli> connect to 8883 iirc
2018-07-18 14:46:02 <danieli> for gui, try mqtt-fx, java client
2018-07-18 14:46:06 <danieli> it's pretty neat
2018-07-18 14:46:20 <clandmeter> Sure
2018-07-18 14:46:41 <clandmeter> Use Port 8883 and set cert
2018-07-18 14:46:53 <pl> Hi. When is a package moved from testing to main or community, provided that no bugs were reported for it?
2018-07-18 14:47:07 <ncopa> pl, when someone asks for it
2018-07-18 14:47:19 <ncopa> there is no automation in it
2018-07-18 14:47:36 <pl> ncopa, thanks :) Should I just follow up to the packaging request bug?
2018-07-18 14:47:56 <ncopa> we initially created the testing repo so we could create packages that we dont user ourselves, so builder can build it and users test them
2018-07-18 14:48:03 <ncopa> before we move them to main or community
2018-07-18 14:48:24 <ncopa> so testing repo is so users can help us test
2018-07-18 14:48:37 <ncopa> unfortunately, we dont get feedback on testing packages often
2018-07-18 14:49:01 <clandmeter> -p --certfile /etc/ssl/certs.pem from my head
2018-07-18 14:49:20 <clandmeter> -p 8883...
2018-07-18 14:50:00 <ncopa> should it be ca file?
2018-07-18 14:50:22 <clandmeter> The one ships with libssl
2018-07-18 14:50:30 <clandmeter> Sorry on mobile
2018-07-18 14:52:01 <ncopa> looks like it also needs a key file
2018-07-18 14:52:17 <ncopa> Error: Both certfile and keyfile must be provided if one of them is.
2018-07-18 14:52:28 <clandmeter> No
2018-07-18 14:52:49 <clandmeter> On client the certs. Pem is all
2018-07-18 14:53:14 <ncopa> mosquitto_sub is not happy
2018-07-18 14:53:31 <ncopa> $ mosquitto_sub -v -h msg.alpinelinux.org -p 8883 -t '#' --cert /etc/ssl/cert.pem
2018-07-18 14:53:31 <ncopa> Error: Both certfile and keyfile must be provided if one of them is.
2018-07-18 14:53:31 <clandmeter> You are holding it wrong
2018-07-18 14:53:54 <clandmeter> Not cert opt
2018-07-18 14:54:09 <clandmeter> There is another option
2018-07-18 14:54:40 <ncopa> --cafile or --capath
2018-07-18 14:54:59 <pl> It seems like I don't have the permissions to close a bug I opened. It has been solved but not closed by anybody, and it is not stale. How show I proceed in these cases?
2018-07-18 14:55:18 <ncopa> which bug is it? i can close it for you
2018-07-18 14:55:29 <clandmeter> --cafile file
2018-07-18 14:55:41 <pl> ncopa, https://bugs.alpinelinux.org/issues/8698
2018-07-18 14:55:44 <clandmeter> Else you need ca-certz installed
2018-07-18 14:55:45 <ncopa> clandmeter: --cafile /etc/ssl/cert.pem worked
2018-07-18 14:55:57 <clandmeter> Yes
2018-07-18 14:56:22 <clandmeter> Now I go back to letour  :)
2018-07-18 14:56:23 <ncopa> pl: so it is working?
2018-07-18 14:56:41 <ncopa> pl: I can close it? there is no comment indicating that issue is fixed
2018-07-18 14:58:55 <pl> ncopa, well, the requested package is now available
2018-07-18 14:59:42 <pl> ncopa, thanks :)
2018-07-18 15:00:39 <ncopa> yw. thanks for following up
2018-07-18 15:03:30 <nullato> Hi. Where can I read how Alpine Linux is organized? Is there an "AlpineLinux developer" status, like Debian has?
2018-07-18 15:16:12 <_ikke_> What kind of information does that status give?
2018-07-18 15:16:30 <_ikke_> We're still busy improving documentation / information, but it takes time
2018-07-18 15:50:25 <clandmeter> My status is afk ;)
2018-07-18 15:54:35 <_ikke_> lol
2018-07-18 18:43:57 <mickibm> question: why do we care about the MACHINE_UUID in mkinitfs? i ask because this doesnt exist in ppc64le
2018-07-18 18:47:48 <mickibm> ncopa: ^ soooo what about an arch variable in mkinitfs?
2018-07-19 09:22:04 <trfl> the cython package is only available for python2, however can also be built for python3. What is the preferred way to provide a python package for both py2 and py3? renaming the current cython package to cython2, creating cython3 and creating a cython package which installs both?
2018-07-19 09:48:20 <Ganwell> trfl: (IMO) if python2 already exists: a package that builds both, but has sub-packages. and yes rename the folder to cython. See py-cffi.
2018-07-19 09:48:55 <Ganwell> trfl: if python2 does not exist, do only python3.
2018-07-19 09:49:10 <rdutra> there are some examples at: https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Python
2018-07-19 10:01:10 <HRio> fabled: seems there are some issues with "main/tevent: upgrade to 0.9.37" http://tpaste.us/5Yly
2018-07-19 10:10:37 <fabled> HRio, hum, is that some public compilation failing, or something you compile privately?
2018-07-19 10:10:45 <fabled> looks trivial to fix
2018-07-19 10:12:18 <fabled> it just needs sys/types.h include. i can add a patch
2018-07-19 10:12:22 <fabled> thanks for the report
2018-07-19 10:18:06 <zacksiri> https://www.irccloud.com/pastebin/b2iiAnem/
2018-07-19 10:18:16 <HRio> fabled: thanks
2018-07-19 10:18:19 <zacksiri> anyone have any idea how to resolve?
2018-07-19 10:19:07 <HRio> fabled: I am working on https://github.com/alpinelinux/aports/pull/4283
2018-07-19 11:21:12 <HRio> fabled: thanks, sssd builds fine again
2018-07-19 15:53:20 <Tsutsukakushi> ?close
2018-07-19 19:26:11 <absurdist> are there modules for QCA9984?
2018-07-19 19:34:46 <danieli> it's in ath10k iirc
2018-07-19 19:34:56 <danieli> which is in linux-firmware-ath10k
2018-07-19 19:35:05 <danieli> @ absurdist
2018-07-19 20:26:08 <HRio> Hi trying a rtc DS3231 connected to a rpi3 anybody else that have tested this? seems to work after adding dtoverlay=i2c-rtc,ds3231 to usercfg.txt, and busybox syslog starts with the right time after boot, but then something in alpine init chanegs back to 1970 :-0
2018-07-19 20:27:43 <HRio> http://tpaste.us/v7XY
2018-07-19 20:28:52 <clandmeter> swclock?
2018-07-19 20:31:39 <clandmeter> HRio: i modified the init to check for rtc and use swclock if no rtc is found.
2018-07-19 20:31:44 <clandmeter> can that be related?
2018-07-19 20:38:55 <HRio> clandmeter: default v3.8 install, only hwclock in /etc/runlevels, rtc is /dev/rtc0
2018-07-19 20:39:26 <HRio> need to be AFK for a bit now...
2018-07-19 21:07:36 <HRio> clandmeter: the check you added is not in v3.8 right? only in edge?
2018-07-19 21:25:01 <HRio> the hwclock init.d "script" does the right thing, its the one that actually sets the system time from the rtc. Confirmed by disabling that. So its something after syslog, the creates the mess
2018-07-19 21:26:46 <absurdist> ok danieli
2018-07-19 21:27:19 <absurdist> still looking for a zram-init script
2018-07-19 22:12:28 <HRio> clandmeter: after some more investigations it seems dropping -Z to busybox syslogd, does not result in any "Jan  1" dates in /var/log/messages (http://tpaste.us/Q4lb)
2018-07-19 22:13:58 <HRio> Shiz: seems -Z was added by you, any pointers for what to look for?
2018-07-19 22:15:50 <Shiz> that's very odd
2018-07-19 22:19:22 <HRio> the diskless rpi3 was installed normally, and then I added the rtc today, so there quite a few 1970 files in the apkovl file
2018-07-19 22:19:59 <HRio> no network access during initial setup
2018-07-20 06:03:38 <ncopa> looks like gettext it broken with latest libunistring
2018-07-20 06:03:48 <ncopa> filezilla does not build
2018-07-20 06:04:17 <ncopa> when i try to rebuild gettext it hangs in test-lock
2018-07-20 06:05:20 <ncopa> during build of filezilla it is msgfmt that aborts, so i suspect that it is fortify headers that is catching some overflow
2018-07-20 06:22:51 <ncopa> ok, reverting libunistring to 0.9.7 makes filezilla build again
2018-07-20 06:50:47 <mepholic> ncopa: note: libunistring seems to build fine with gcompat-tiny
2018-07-20 06:51:47 <mepholic> oh, filezilla specifically
2018-07-20 06:51:49 <mepholic> interesting
2018-07-20 06:52:18 <mepholic> lemme pull that in real quick and test it
2018-07-20 06:53:37 <awilfox> gettext-tiny doesn't use libunistring; gettext does
2018-07-20 06:54:03 <awilfox> http://git.savannah.gnu.org/cgit/libunistring.git/tree/ they don't seem to understand you need to put the source in git
2018-07-20 06:54:17 <awilfox> they actually put all the .h files in .gitignore...
2018-07-20 06:54:20 <mepholic> ahhhh
2018-07-20 06:54:44 <awilfox> mepholic: does filezilla build properly with gettext-tiny msgmerge?  if so, that would be pretty hilarious...
2018-07-20 06:54:57 <ncopa> i have some vauge memory of trying to avoid build libunistring from years ago
2018-07-20 06:55:20 <ncopa> lots of gnulib magic
2018-07-20 06:55:21 <awilfox> ncopa: yes, its test suite was extremely broken and so the goal was to just remove it entirely from aports.git
2018-07-20 06:55:29 <awilfox> the test suite seems to be mostly fixed now
2018-07-20 06:55:31 <mepholic> awilfox: there's a large dep tree that we're missing
2018-07-20 06:55:50 <mepholic> i can try to pull it all in, but it's gon be hella sloppy
2018-07-20 06:56:28 <ncopa> libunistring builds, it doesnt run...
2018-07-20 06:58:44 <ncopa> things like this .... http://git.savannah.gnu.org/cgit/libunistring.git/commit/?id=22b50b4fdc0fcf967b1e61ff0d391eb4e7a9824d
2018-07-20 06:58:55 <ncopa> why they even have their own stdbool.h?
2018-07-20 06:59:40 <awilfox> I tried to diff 0.9.7 and 0.9.10 and it made my eyes water
2018-07-20 06:59:52 <awilfox> they've completely reimplemented the entire line break module
2018-07-20 07:00:00 <awilfox> it is an array 35x30 instead of 27x27
2018-07-20 07:00:06 <awilfox> and they added a bunch more types
2018-07-20 07:00:20 <awilfox> so it is probably down to forgetting to check for one in u8_possible_linebreak, since the abort() is being raised there
2018-07-20 07:00:34 <awilfox> I think I enjoyed SILO more >_>
2018-07-20 07:02:32 <ncopa> whatta....
2018-07-20 07:02:48 <ncopa> so i rebuilt libunistring with -dbg subpabkage
2018-07-20 07:03:04 <ncopa> http://tpaste.us/oQ6R
2018-07-20 07:03:08 <ncopa> guess what happens...
2018-07-20 07:03:21 <ncopa> filezilla builds...
2018-07-20 07:03:22 <awilfox> noc rash?
2018-07-20 07:03:24 <awilfox> haha
2018-07-20 07:03:27 <ncopa> yup
2018-07-20 07:03:59 <awilfox> just out of curiosity, if you build again without -dbg, it does still crash right?
2018-07-20 07:04:20 <awilfox> does -dbg do anything besides add -g to compile flags?
2018-07-20 07:04:25 <mepholic> o_O
2018-07-20 07:04:35 <awilfox> (obviously it won't strip binaries, but does it do anything in *build* besides add -g)
2018-07-20 07:05:25 <awilfox> mepholic: I thought filezilla was a qt5 app
2018-07-20 07:05:48 <mepholic> wxgtk
2018-07-20 07:05:52 <awilfox> aah
2018-07-20 07:05:52 <mepholic> it appears
2018-07-20 07:06:03 <mepholic> and I don't realllllly wanna touch that right at this particular moment
2018-07-20 07:06:05 <mepholic> lmao
2018-07-20 07:06:09 <awilfox> yeah
2018-07-20 07:06:22 <awilfox> I'll be playing around with that when I get around to audacity
2018-07-20 07:06:26 <awilfox> and not a moment before
2018-07-20 07:06:32 <mepholic> i already messed with wxgtk a while ago
2018-07-20 07:06:40 <mepholic> and i recall it being very... welll
2018-07-20 07:06:42 <mepholic> "fun"
2018-07-20 07:07:03 <awilfox> it was hard to build on just a glibc x86_64 system
2018-07-20 07:07:15 <awilfox> musl ppc64 is going to be fun :)
2018-07-20 07:07:21 <mepholic> "fun"
2018-07-20 07:07:23 <awilfox> don't even know if it is endian safe
2018-07-20 07:08:03 <ncopa> -dbg shouldnt do anything else than add -g and split out the symbols to separate file
2018-07-20 07:08:24 <ncopa> this gettext-tiny sounds like something i'd want
2018-07-20 07:09:44 <awilfox> we're still working at Adélie on making it work right
2018-07-20 07:10:04 <awilfox> we've built about hmm.. half the distro with it at this point I think?  four bugs fixed, and two unresolved failures so far.
2018-07-20 07:12:05 <ncopa> you build everyting with NLS, right?
2018-07-20 07:12:25 <awilfox> yes
2018-07-20 07:12:36 <awilfox> well, we don't *explicitly disable* NLS is more correct
2018-07-20 07:12:41 <ncopa> half distro. thats pretty impressive
2018-07-20 07:12:45 <awilfox> there are plenty of packages that don't have translations
2018-07-20 07:13:01 <awilfox> yeah.  sabotage is using it everywhere, but they don't ship some of this stuff
2018-07-20 07:13:25 <awilfox> libbytesize needs a patch to work
2018-07-20 07:13:40 <ncopa> ok, i can confirm, build again with out -dbg makes it crash again
2018-07-20 07:13:53 <awilfox> that's in testing/ in alpine it seems; it's a dep of newest udisks2
2018-07-20 07:14:32 <ncopa> i need to go out for a while
2018-07-20 07:14:40 <awilfox> that's really weird
2018-07-20 07:15:00 <mepholic> yeah that's freakin strange
2018-07-20 09:18:54 <vkrishn> does it have the same result, i.e -dbg pkg build, without -O or with -Og ? ncopa ^^^
2018-07-20 09:24:25 <vkrishn> meaning disabling the default 'CFLAGS="-Os -fomit-frame-pointer"'
2018-07-20 11:46:25 <trfl> I have made a set of apkbuilds for vapoursynth, a video processing framework with several plugins; vapoursynth-{ffms2,mvtools,nnedi3}. The main package provides standalone tools for processing video, however more conveniently the 'mpv' media player (currently in community) can link with to perform realtime video filtering. To build vapoursynth, the 'cython' package (currently in main, py2-only) has to be split into py2 and py3 variants.
2018-07-20 11:47:21 <trfl> What would be the best way to submit all this? Perform the cython split first, wait for it to go through, then push each vapoursynth package separately? And regarding the cython package, since this is a python-to-C compiler I wonder if it's right to prefix it with py2- and py3- like we do with python libraries
2018-07-20 11:48:04 <trfl> correction: mpv would link with the vapoursynth main package as a dependency
2018-07-20 11:53:59 <trfl> either way I'm not expecting to land vapoursynth as a dependency of mpv until it leaves testing, so maybe an additional step would be to provide mpv-vs in testing as a means to test this as well
2018-07-20 11:56:37 <danieli> spltting up deps that must be split up first, add other dependencies, add the software itself, then update other packages so they can makedep on that
2018-07-20 11:57:38 <danieli> that's the only flow I see that makes sense
2018-07-20 12:19:41 <trfl> alright, sounds good
2018-07-20 12:43:41 <ncopa> oh well, today is one of those days where nothing seems to work
2018-07-20 12:43:47 <ncopa> gettext test suite hangs
2018-07-20 12:44:07 <danieli> let's see
2018-07-20 12:44:22 <danieli> it hangs locally? ncopa
2018-07-20 12:44:29 <ncopa> yes
2018-07-20 12:44:35 <danieli> what does gdb say?
2018-07-20 12:44:40 <danieli> or just strace -f
2018-07-20 12:45:50 <danieli> kaniini: are you still around? I'd love to pick up your work in implementing SIGs/WGs
2018-07-20 12:45:54 <danieli> it's a pressing issue imo
2018-07-20 12:46:06 <danieli> I mean, first and foremost the work of formalizing it
2018-07-20 12:46:50 <ncopa> https://dpaste.de/vw0z
2018-07-20 12:48:30 <danieli> that looks pretty unhelpful
2018-07-20 12:49:31 <danieli> it's just deferring indefinitely?
2018-07-20 12:49:46 <danieli> got the diff of your apkbuild/patches?
2018-07-20 12:50:25 <ncopa> http://tpaste.us/B6lY
2018-07-20 12:50:33 <ncopa> but it happens even if i just rebuild it
2018-07-20 12:50:43 <danieli> that'd odd, how the heck did it build in the first place?
2018-07-20 12:50:52 <ncopa> exactly
2018-07-20 12:51:02 <danieli> I'm testing locally
2018-07-20 12:51:07 <ncopa> my guess is that glib configure check does something different now with new musl
2018-07-20 12:51:13 <ncopa> i mean gnulib
2018-07-20 12:51:21 <ncopa> gnulib is a big mess
2018-07-20 12:51:29 <danieli> that is true
2018-07-20 12:54:20 <andypost[m]> Can someone cherry-pick https://github.com/alpinelinux/aports/pull/4804 to 3.8?
2018-07-20 12:55:04 <ncopa> ok
2018-07-20 12:55:14 <ncopa> that sounds like more fun than messing with gnulib :)
2018-07-20 12:55:24 <danieli> i just gt a fail for the test suite
2018-07-20 12:55:25 <danieli> 21 failed
2018-07-20 12:55:33 <danieli> i'm running it in a docker container and not bare metal though
2018-07-20 12:56:27 <danieli> http://tpaste.us/bg6M ncopa
2018-07-20 12:56:29 <ncopa> i run this in lxc
2018-07-20 12:56:39 <danieli> my host is glibc arch
2018-07-20 12:57:02 <danieli> i had 21 failed tests, 353 passed, 38 skipped
2018-07-20 12:57:10 <danieli> applied your patch first
2018-07-20 13:00:44 <danieli> ncopa: http://tpaste.us/7NeB
2018-07-20 13:06:27 <ncopa> interesting. it expects it to split the lines, but it doesnt
2018-07-20 13:06:40 <danieli> but the test suite hangs for you?
2018-07-20 13:07:33 <ncopa> gettext-runtime/tests/test-lock hangs
2018-07-20 13:07:58 <ncopa> it is possible that it is caused by updated libunistring
2018-07-20 13:11:12 <danieli> that sounds plausible
2018-07-20 13:12:03 <ncopa> but libunistring shouldnt be involved at all in the test-lock
2018-07-20 13:12:20 <danieli> i wonder why that happens
2018-07-20 13:12:22 <danieli> it didn't for me
2018-07-20 13:12:35 <danieli> did you start with a clean and updated edge build env?
2018-07-20 13:13:09 <ncopa> clean, no, updated edge, yes
2018-07-20 13:15:02 <danieli> my env is fairly newly set up
2018-07-20 13:15:11 <danieli> i restore from a clean snapshot every once in a while
2018-07-20 13:39:09 <ncopa> oh
2018-07-20 13:39:24 <ncopa> maybe i'm just not patient enough
2018-07-20 13:39:29 <danieli> did it work?
2018-07-20 13:39:36 <ncopa> i enabled debug print
2018-07-20 13:39:42 <ncopa> and it worked
2018-07-20 13:39:42 <danieli> my computer has 8 cores @ 5 GHz so it builds fairly quickly
2018-07-20 13:40:06 <ncopa> i use a packet.net machine 16 cores...
2018-07-20 13:40:16 <danieli> jeez
2018-07-20 13:40:19 <danieli> and it's slow?
2018-07-20 13:40:26 <ncopa> hello gnu
2018-07-20 13:40:26 <danieli> patience is a virtue, I suppose
2018-07-20 13:40:28 <danieli> lol
2018-07-20 13:40:30 <ncopa> what do you think
2018-07-20 13:40:43 <danieli> can't imagine it being slow but you never know, hello gnu
2018-07-20 13:40:52 <danieli> can't imagine it being that slow*
2018-07-20 13:41:00 <ncopa> configure script takes an eternity
2018-07-20 13:41:04 <ncopa> 1-2 minutes
2018-07-20 13:41:04 <danieli> yeah, as usual
2018-07-20 13:41:38 <danieli> who do you have to get on your knees for to borrow an lxc container to test builds?
2018-07-20 13:41:47 <danieli> as far as I understand, there's no proper solution in place
2018-07-20 13:42:42 <danieli> I'm diffing tmate with the upstream tmux version / build it's based on so I can merge it into a newer tmux
2018-07-20 13:46:52 <clandmeter> packet.net with 16 cores?
2018-07-20 13:47:05 <clandmeter> you limit it to 16?
2018-07-20 13:51:02 <danieli> afk for a little bit
2018-07-20 13:51:25 <danieli> gonna help my grandpa do a commercial stunt, giving away ice cream to a ton of kids
2018-07-20 14:12:48 <ncopa> sounds fun!
2018-07-20 14:13:21 <ncopa> oh, it has 24 cores
2018-07-20 14:13:30 <ncopa> + hyperthreading
2018-07-20 14:14:20 <ncopa> oh well, the locking issue is still locking
2018-07-20 14:14:35 <ncopa> i think it hangs for real without debug output
2018-07-20 14:14:53 <ncopa> my guess is that the gnulib threading stuff is broken...
2018-07-20 14:15:36 <ncopa> why would gettext use threading in the first place?
2018-07-20 14:21:54 <ncopa> danieli: i think gettext tests fails due to updated libunistring
2018-07-20 14:22:06 <ncopa> they fail here too now
2018-07-20 14:55:01 <ncopa> something weird is definitively going on with libunistring
2018-07-20 14:55:26 <ncopa> when i enable -dbg package, the gettext package builds
2018-07-20 14:55:28 <ncopa> only 1 failure
2018-07-20 14:55:39 <ncopa> so i think libunistring is broken somehow
2018-07-20 15:59:33 <danieli> hm
2018-07-20 16:04:18 <ncopa> this is just weird
2018-07-21 08:49:30 <clandmeter> RIP apu2 :(
2018-07-21 08:52:16 <_ikke_> apu2?
2018-07-21 09:28:44 <clandmeter> Yes my router hw
2018-07-21 09:40:02 <andypost[m]> does it makes sense to backport secfixes list to older releases?
2018-07-21 10:05:27 <terra> ncopa: https://github.com/alpinelinux/aports/blob/master/main/ldb/disable-python-tests.patch
2018-07-21 10:05:41 <terra> ncopa: Is this Alpine specific problem?
2018-07-21 10:06:22 <terra> ncopa: I didn't see such thing from other distros
2018-07-21 10:07:08 <terra> this is debian build log: http://downloads.van-belle.nl/samba4/Buildlogs/stretch/ldb_1.3.4+nmu-1~deb9_amd64.build
2018-07-21 10:08:29 <terra> I'm trying to upgrade ldb to 1.3.4, but it fails during testing
2018-07-21 13:00:54 <absurdist> why are packages not optimized with LTO yet?
2018-07-21 13:06:16 <_ikke_> LTO?
2018-07-21 13:06:26 <skarnet> because you haven't done it
2018-07-21 13:07:11 <rdutra> when trying to do an upgrade in my container I got an error: http://tpaste.us/gan6
2018-07-21 13:07:37 <rdutra> there is a -v option to enable verbose but I did not show any useful information
2018-07-21 13:07:40 <rdutra> never got it before..
2018-07-21 13:09:21 <rdutra> any tips?
2018-07-21 13:10:37 <rdutra> nm, there is an `apk fix` command
2018-07-21 13:11:10 <rdutra> it solves the issue
2018-07-21 13:18:07 <_ikke_> Nice to have a fix command that just fixes your issues :D
2018-07-21 13:18:52 <rdutra> really nice :)
2018-07-22 16:01:46 <trfl> I am packaging some software with version numbers like "r20" and "2.23". To retain x.y.z semantic versioning, I have declared _ver with the upstream version and declared pkgver like 0.$_ver.0 or $_ver.0
2018-07-22 16:01:51 <trfl> is this acceptable?
2018-07-22 17:10:31 <trfl> ^ works great on my laptop fwiw
2018-07-22 17:11:05 <trfl> not the exact same usage tho, ``udhcpc -i wlan0''
2018-07-22 23:58:41 <danieli> i forgot to uncomment some lines
2018-07-23 07:39:56 <HRio> jirutka: did you notice the qemu build failure (https://git.alpinelinux.org/cgit/aports/commit/?id=5220530c)?
2018-07-23 07:43:20 <clandmeter> HRio, i think if you want to contact him you need to email him.
2018-07-23 07:45:03 <HRio> clandmeter: a OK.
2018-07-23 07:45:28 <clandmeter> i havent seen him here for almost 2 months.
2018-07-23 07:47:40 <HRio> ok mailed him now
2018-07-23 09:01:21 <HRio> anybody that can look at the qemu build failure? it could be related to the resent update of linux-headers.
2018-07-23 09:01:48 <HRio> qemu has not been built after linux-headers update
2018-07-23 09:09:34 <clandmeter> looks like the dev did not testbuild before pushing.
2018-07-23 09:19:49 <ncopa> morning
2018-07-23 09:19:52 <ncopa> it happens...
2018-07-23 09:20:18 <clandmeter> its sloppy when you set high standards yourself
2018-07-23 09:20:42 <clandmeter> im grumpy so you can safely ignore me.
2018-07-23 09:20:53 <ncopa> yes, agree, we want avoid it
2018-07-23 09:42:27 <HRio> also xen seems to fail
2018-07-23 09:43:26 <clandmeter> probably anything that uses sysinfo
2018-07-23 09:44:10 <HRio> this seems to be elsewhere, so could be unrelated
2018-07-23 09:48:43 <clandmeter> i didnt look close enough
2018-07-23 09:49:39 <clandmeter> ncopa, seems that new apk feature to update stale cache is biting us.
2018-07-23 09:51:29 <ncopa> ok?
2018-07-23 09:51:39 <ncopa> the apk del issue that updates apk cache?
2018-07-23 09:51:48 <clandmeter> i think more
2018-07-23 09:52:04 <clandmeter> somebody complained on tmpfs it tries to update when installing root.
2018-07-23 09:52:20 <clandmeter> https://bugs.alpinelinux.org/issues/9126
2018-07-23 10:30:14 <HRio> xen fails with linux-headers 4.4.6-r2 as well, so there are other breaking changes somewhere :-) http://tpaste.us/ZPez
2018-07-23 11:11:39 <HRio> xen 4.11 (latest) builds but not xen 4.10.1 that we have in @edge at the moment
2018-07-23 11:11:54 <HRio> will prepare a PR for xen 4.11
2018-07-23 12:35:32 <andypost[m]> Should I wait for CVEs filed before backpirting to early releases? https://github.com/alpinelinux/aports/pull/4815
2018-07-23 12:35:53 <andypost[m]> *backporting
2018-07-23 13:00:54 <ncopa> i dont think so. i think we push fix early if possible and deal with the paperwork afterwards if necessary
2018-07-23 13:04:39 <andypost[m]> Will prepare PRs for early branches soon
2018-07-23 13:05:25 <andypost[m]> @ncopa:alpinelinux.org: is 3.5 supported?
2018-07-23 13:09:14 <ncopa> yes
2018-07-23 13:09:40 <ncopa> https://wiki.alpinelinux.org/wiki/Alpine_Linux:Releases
2018-07-23 17:37:01 <HRio> hi anybody that can help with this http://build.alpinelinux.org/buildlogs/build-edge-x86_64/testing/sssd/sssd-1.15.3-r0.log? worked fine in my system and in travis, but not in the proper builder.
2018-07-23 20:01:53 <vkrishn> ncopa, yet another simple way of preservering the pointy releases...
2018-07-23 20:01:54 <vkrishn> 1. rename 'http://nl.alpinelinux.org/alpine/v3.8/main/x86/' to http://nl.alpinelinux.org/alpine/v3.8.0/main/x86/
2018-07-23 20:01:56 <vkrishn> 2. recreate http://nl.alpinelinux.org/alpine/v3.8/main/x86/
2018-07-23 20:01:57 <vkrishn> 3. overlay mount rw=v3.8/main/x86/ ro=v3.8.0/main/x86/
2018-07-23 20:01:58 <vkrishn> and voila! we have almost smooth and transparent rollback mechanism (ahem cough!)
2018-07-23 20:08:47 <vkrishn> if there is a pointy release v3.8.1 and v3.8.0 apks are not preserved, then keeping v3.8.0.iso around also does not make lots of practical sense
2018-07-23 20:15:59 <vkrishn> gtg, I know nobody wants rollback on a rolling release :(
2018-07-23 20:16:16 <danieli> it would break a lot of existing systems
2018-07-23 20:16:30 <danieli> we could move to doing major.minor.patch/bug in the future
2018-07-23 20:20:41 <vkrishn> hmm... not quite, we simply ignore http://nl.alpinelinux.org/alpine/v3.8.0 as historic
2018-07-23 20:21:03 <vkrishn> and keep running http://nl.alpinelinux.org/alpine/v3.8 as usual
2018-07-23 20:21:24 <danieli> we don't control all mirrors
2018-07-23 20:21:40 <vkrishn> to mirrors it would look like whole set of files
2018-07-23 20:22:03 <vkrishn> it cannot see few files, but full set of overlay files
2018-07-23 20:22:35 <danieli> that's a very odd phrasing, mind rephrasing?
2018-07-23 20:25:12 <vkrishn> I mean why should any mirror be a problem ?
2018-07-23 20:27:15 <vkrishn> though it will be a little work making those overlay mounts on nl.a.o
2018-07-23 20:27:57 <vkrishn> by 'few files' I mean only the updated ones
2018-07-23 20:33:05 <vkrishn> more accurate, overlay should be, mounts rw=v3.8.1/main/x86/ ro=v3.8.0/main/x86/ mountpoint=ro=v3.8/main/x86/
2018-07-23 20:33:08 <vkrishn> v3.8.1/main/x86/ will be hidden from mirrors till final release of v3.8.1 and v3.8.2 gets rolling
2018-07-23 20:36:50 <vkrishn> when v3.8.1 is done, v3.8.1/main/x86/ will stay mirror hidden and new mountpoint=v3.8.1/main/x86/ created
2018-07-23 20:36:53 <vkrishn> then, mounts ro=v3.8.1/main/x86/(hidden) ro=v3.8.0/main/x86/ mountpoint=v3.8.1/main/x86/
2018-07-23 20:37:47 <vkrishn> I hope it makes better sense now
2018-07-23 20:42:17 <vkrishn> if still jinxee, let me know, gtg
2018-07-23 20:56:21 <fragstone> good evening
2018-07-24 10:34:31 <vkrishn> Method III
2018-07-24 10:34:31 <vkrishn> if preserving pointy released apk gets interesting,
2018-07-24 10:34:31 <vkrishn> there is also a central apk repo method, similar to debian's, but slightly different,
2018-07-24 10:34:34 <vkrishn> it requires slight modification of apk tools, but lot cleaner, pls let me know
2018-07-24 10:37:04 <vkrishn> METHOD IV
2018-07-24 10:37:05 <vkrishn> is symlink method which also some distros do, and would not require any modification of build environment, or apk tools
2018-07-24 12:49:03 <HRio> whats the status of fuse3 in testing? ready for main? needed to update sshfs.
2018-07-24 13:06:44 <andypost> ncopa, here's patch for 3.5 https://github.com/alpinelinux/aports/pull/4839 other branches could pick commit from 4815
2018-07-24 13:17:08 <ncopa> HRio: do you know if there are any other packages that uses fuse but cannot use fuse3?
2018-07-24 13:17:35 <ncopa> i'd like to check if we can upgrade fuse to new API and rebuild everything against that
2018-07-24 13:18:05 <ncopa> i suspect we cannot do that yet, but it would be nice to know what the packages are that are missing
2018-07-24 13:18:24 <ncopa> i suppose we can simply move fuse3 to main and rebuild sshfs against that
2018-07-24 13:21:03 <andypost> ncopa, travis does not work for other branches as edge?
2018-07-24 13:21:22 <ncopa> does not look like it does
2018-07-24 13:21:36 <ncopa> but its ok, im test building it here
2018-07-24 13:21:53 <ncopa> and i pushed it. thanks!
2018-07-24 13:22:20 <andypost> now I'm working on php7 backports... there's different versions in early releases
2018-07-24 13:26:28 <ncopa> looks like php5 is in community in v3.[6-8] so we dont need to backport it for those
2018-07-24 13:27:04 <andypost> ncopa, yes only 3.5 has it in main, that's why I filed this PR
2018-07-24 13:27:17 <ncopa> good. thanks
2018-07-24 13:27:40 <andypost> for other branches patch from 4815 applies fine and builds
2018-07-24 13:27:47 <andypost> at least for 86_64
2018-07-24 13:59:45 <ncopa> i dont think we need patch the other branches
2018-07-24 13:59:46 <ncopa> or?
2018-07-24 14:01:01 <andypost> ncopa, all version affected and needs upgrade
2018-07-24 14:02:58 <ncopa> the idea with community repo is that we only support the latest stable
2018-07-24 14:03:07 <ncopa> to reduce work on sec patching and things like that
2018-07-24 14:03:37 <ncopa> if we want support it longer than that we move it to main
2018-07-24 14:08:30 <andypost> ncopa, good to know but I always used to provide patches for community before in case of security
2018-07-24 14:12:30 <ncopa> andypost: i think the official approach is "we fix on request"
2018-07-24 15:11:55 <ncopa> andypost: i pushed fix for 3.7-stable
2018-07-24 15:20:02 <ncopa> andypost: and i now also pushed it for 3.6-stable
2018-07-24 15:20:52 <andypost> ncopa, thanx, I'm testing upgrade php7 for 3.7 on arm8 so will send PR when will be sure it works
2018-07-24 19:48:32 <terra> ncopa: I can run Void's chromium-66 binary on Alpine with LD_PRELOAD'ing two libraries from Void
2018-07-24 19:50:42 <terra> https://paste.ubuntu.com/p/qyp4DP48tt/
2018-07-24 21:44:16 <HRio> ncopa: not investigated the fuse3 situation, just noticed that it was still lingering in testing when having a look at a newer sshfs version
2018-07-25 08:39:42 <ncopa> terra: that is interesting
2018-07-25 15:25:19 <nmeum> does anybody know how to use ap from lua-aports? Sadly it doesn't have much documentation. Just running 'ap builddirs busybox' from $HOME gives me `/usr/bin/lua5.2: /usr/bin/ap:47: attempt to index local 'db' (a nil value)`. Is this a bug or am I doing something wrong?
2018-07-25 15:43:01 <nmeum> ah, it needs to be invoked with a specific repo, e.g. `ap builddirs -d ~/src/aports/main vim`
2018-07-26 18:54:32 <kpcyrd> I think the "revert back from libressl" broke both libressl and openssl: on a fresh edge install I can't install openssl since it conflicts with libressl2.7-libcrypto, I can't uninstall libressl since stuff depends on it and I can't use libressl because `"pkg-config" "--libs" "--cflags" "openssl"` doesn't pick it up anymore
2018-07-26 18:59:41 <kpcyrd> well, turned out it's indeed picked up by pkg-config, but I haven't touched alpine in a while and forgot to install -dev
2018-07-27 22:54:51 <keith_> I also proposed a fix for that issue, at https://github.com/alpinelinux/aports/pull/4871
2018-07-27 22:55:05 <keith_> happy to incorporate any feedback
2018-07-28 03:07:35 <deedubs> are there any examples of a meta package? I have a project that builds a number of binaries but I want to dist them all as individual packages
2018-07-28 03:13:13 <danieli> deedubs: https://wiki.alpinelinux.org/wiki/APKBUILD_meta:Meta
2018-07-28 03:13:36 <danieli> that's a pure metapackage
2018-07-28 03:13:49 <danieli> pure as in it doesn't do anything but depend on other packages
2018-07-28 03:19:24 <deedubs> okay so that's not exactly what I want to do ;D but I guess it may not be a common use case
2018-07-28 03:20:58 <deedubs> I'm trying to build kubernetes, which makes my dual xeon build machine increase the temperature in the room by a solid 15C.  I was hoping to build the whole project, then divvy it up into subprojects
2018-07-28 03:25:47 <danieli> does kubernetes not exist on alpine?
2018-07-28 03:26:00 <danieli> that'd be good for docker (Docker, Inc.) too
2018-07-28 03:28:14 <deedubs> It does exist, just not packaged the way I want :) https://github.com/alpinelinux/aports/blob/master/testing/kubernetes/APKBUILD
2018-07-28 03:28:54 <danieli> https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Multiple_Subpackages this a better example?
2018-07-28 03:36:59 <deedubs> Yeah that seems close to the approach I settled on organically https://github.com/massivekube/aports/blob/master/testing/kubernetes/APKBUILD
2018-07-28 03:55:45 <deedubs> I may end up breaking it into a number of standalone packages :/
2018-07-28 04:16:50 <danieli> that'd be nice, feel free to contribute upstream too
2018-07-28 04:16:54 <danieli> upstream to us
2018-07-28 04:22:43 <deedubs> Yep that's the plan :)
2018-07-28 04:55:52 <deedubs> In the meantime https://github.com/alpinelinux/aports/pull/4872
2018-07-28 11:20:03 <ncopa> good news, i got chromium 66 working
2018-07-28 11:20:17 <ncopa> bad news, chromium 68 is out
2018-07-28 17:02:37 <danieli> ncopa: classic, lol
2018-07-28 17:54:47 <clandmeter> Lol
2018-07-29 00:20:52 <Shiz> kaniini: whatd you do to piss off the dweebs
2018-07-29 00:21:29 <kaniini> Shiz GNAA tried to rile up a bunch of "SJW" people to try to get kloeri fired from his job and shit
2018-07-29 00:21:35 <Shiz> yeah i saw that
2018-07-29 00:21:38 <kaniini> Shiz so i negatively impacted their campaign
2018-07-29 00:21:38 <Shiz> as they do
2018-07-29 00:21:46 <kaniini> Shiz so they are retaliating :)
2018-07-29 00:21:49 <Shiz> lol
2018-07-29 00:21:51 <Shiz> whatd you do
2018-07-29 00:21:54 <Shiz> take over their roots?
2018-07-29 00:22:18 <kaniini> no, i just lent my own social capital in the circles they were trying to do fuckery in, to kloeri's defense
2018-07-29 00:22:42 <kaniini> and so their sockpuppet accounts got banned from various mastodon instances as a result
2018-07-29 00:22:45 <kaniini> and now they are mad :)
2018-07-29 00:23:11 <Shiz> lol
2018-07-29 00:23:13 <Shiz> nice
2018-07-29 00:23:31 <kaniini> can i interest you in some global irc advertising
2018-07-29 00:23:46 <kaniini> i'm more concerned about the CVE in pkgconf honestly
2018-07-29 00:23:48 <kaniini> :)
2018-07-29 00:47:15 <Shiz> oh dear
2018-07-29 00:47:18 <Shiz> whats up there
2018-07-29 00:49:46 <kaniini> Shiz: https://git.dereferenced.org/pkgconf/pkgconf/commit/9b7affe0b1e6512c6c73d19e1220c94fdb5c8159
2018-07-29 00:50:38 <Shiz> whoopsie
2018-07-29 01:10:56 <danieli> kaniini: nice one, lol
2018-07-29 01:11:16 <danieli> I also just saw a commit to pkgconf in alpine
2018-07-29 01:25:16 <kaniini> danieli i have a bug bounty
2018-07-29 01:25:19 <kaniini> danieli for pkgconf
2018-07-29 01:26:06 <danieli> ah, okay, gotcha
2018-07-29 01:27:30 <kaniini> danieli not kidding: http://kaniini.dereferenced.org/2016/12/31/pkgconf-1.1.0-release.html
2018-07-29 01:31:39 <danieli> i don't doubt it, but i'll give it a read
2018-07-29 01:32:44 <danieli> hah, cool
2018-07-29 01:35:24 <danieli> kaniini: reminds me, I passed you an email with ncopa as cc, regards WGs/SIGs
2018-07-29 01:35:39 <kaniini> i haven't had much time to do anything formal in regards to that
2018-07-29 01:36:06 <danieli> all right, was just wondering if you had any interesting thoughts on that
2018-07-29 01:36:10 <danieli> I already wrote a draft
2018-07-29 01:36:28 <danieli> it's in document form on google docs, just to get the content right first, really
2018-07-29 01:36:46 <awilfox> danieli: was it similar to http://lists.alpinelinux.org/alpine-devel/5811.html ?
2018-07-29 01:37:27 <kaniini> [the resistance to SIGs is one of the reasons why adelie exists as an independent project instead of just taking alpine and ubuntuing it up]
2018-07-29 01:37:32 <danieli> it's a lot more specific, pretty similar to the SIG centos article on their wiki
2018-07-29 01:37:44 <danieli> took inspiration from a few other projects
2018-07-29 01:38:12 <kaniini> every attempt i have made to streamline operations in alpine has been met with so much resistance that i've pretty much given up
2018-07-29 01:38:16 <danieli> organisationally, i see it as a necessity as the alpine project continues growing
2018-07-29 01:38:39 <kaniini> my goal was to streamline alpine so that we did not have to hardfork adelie
2018-07-29 01:38:44 <danieli> well, then core devs will have to be a BDFL
2018-07-29 01:38:45 <kaniini> but ultimately, we have had to do so anyway
2018-07-29 01:39:06 <kaniini> danieli the problem is, passive resistance
2018-07-29 01:39:13 <danieli> elaborate?
2018-07-29 01:39:32 <kaniini> "oh, that sounds nice, but i don't have any thoughts on how to proceed"
2018-07-29 01:39:46 <danieli> someone really just needs to get the groundwork laid down, or so it seems
2018-07-29 01:40:07 <kaniini> yes, but when there is not really any enthusiasm for such things even amongst the core devs...
2018-07-29 01:40:20 <danieli> they're spread pretty thin just maintaining the project to be honest
2018-07-29 01:40:22 <kaniini> it's not like you can just say "ok we are doing it this way and that's that"
2018-07-29 01:41:07 <danieli> well, of course, but doing the work and then asking for your [core devs'] blessing would work
2018-07-29 01:41:11 <danieli> although it'd be risky in terms of the effort spent
2018-07-29 01:41:26 <kaniini> i'm not convinced core would provide such a blessing honestly
2018-07-29 01:41:34 <kaniini> i do not think core really cares :)
2018-07-29 01:42:29 <danieli> ncopa seems to care, but you tend to get kinda jaded after being neck deep in code and problems all day every day
2018-07-29 01:42:31 <kaniini> case in point (sorry jirutka), jirutka is still a dev *and* still on core, despite from what i can tell, almost everyone else on core wanting him removed from that role
2018-07-29 01:42:59 <kaniini> you would think if ncopa wants jirutka ejected from core that would definitely carry a lot of weight
2018-07-29 01:43:04 <kaniini> but alas, nothing has happened
2018-07-29 01:43:26 <danieli> that's a different situation, one i have no say or weight in
2018-07-29 01:43:44 <kaniini> i understand that, i am just explaining that core is dysfunctional
2018-07-29 01:43:52 <danieli> aren't we all, in one way or another
2018-07-29 01:44:04 <kaniini> sure
2018-07-29 01:44:37 <kaniini> but, core is dysfunctional to the point that it cannot even execute on decisions which clearly have broad consensus amongst the members
2018-07-29 01:44:43 <kaniini> so i do not think it is even functional enough to give blessing to somebody else doing the work
2018-07-29 01:45:08 <danieli> i've already had a chat with ncopa, clandmeter and _ikke_ about it
2018-07-29 01:45:17 <danieli> really just needs finishing, second opinions, and publication
2018-07-29 01:45:20 <kaniini> yes, if you poll people individually
2018-07-29 01:45:29 <kaniini> they will say they support or not support
2018-07-29 01:45:34 <kaniini> but when it comes time to execute
2018-07-29 01:45:37 <kaniini> that's the problem :)
2018-07-29 01:46:09 <danieli> I can be stubborn
2018-07-29 01:46:18 <danieli> if that's enough, we have yet to see
2018-07-29 01:46:37 <kaniini> i can be too, and when i took a 'secretarial' role and micromanaged everything, we did make some progress
2018-07-29 01:46:39 <kaniini> but i do not have time to do that
2018-07-29 01:46:48 <kaniini> i will admit lack of time is a large part of the problem
2018-07-29 01:47:15 <danieli> I can agree there
2018-07-29 01:51:23 <danieli> see PM by the way, not sure if I mentioned, but I pased you a link
2018-07-29 01:51:28 <danieli> passed*
2018-07-29 03:27:48 <deedubs> Many hands make light work
2018-07-29 03:33:01 <deedubs> I'd love to take something off the core teams plate, initial review of PRs maybe?
2018-07-29 03:41:32 <danieli> there's both patchwork and github for that
2018-07-29 03:43:17 <deedubs> yes but is there a proceedure or process inplace
2018-07-29 03:43:35 <deedubs> level 1 reviewers, mergers?
2018-07-29 03:43:42 <deedubs> is more of what I'm getting at
2018-07-29 04:51:24 <kaniini> Shiz so i have been playing with this spam shit
2018-07-29 04:51:33 <kaniini> Shiz it turns out it's all huawei routers
2018-07-29 06:37:52 <_ikke_> kaniini: you mean the irc spam here on freenode?
2018-07-29 06:38:00 <kaniini> yes
2018-07-29 06:40:21 <_ikke_> did you fingerprint the addresses where the spam came from?
2018-07-29 08:09:50 <kaniini> _ikke_ i did worse
2018-07-29 08:09:56 <kaniini> _ikke_ https://github.com/kaniini/antissh
2018-07-29 08:13:15 <kaniini> ncopa i think we should disable tcp forwarding by default.  https://code.foxkit.us/adelie/packages/commit/2552843d964162f7ba105775c05a231d9cdd4ae8 does so in adelie.  thoughts?
2018-07-29 14:15:26 <Xe> kaniini: the way apk handles config file upgrades means that existing user configurations won't be affected, and adding a little bit of a scary warning to that part in the config we provide might be a good idea
2018-07-29 15:29:01 <ncopa> kaniini: i wonder why?
2018-07-29 15:33:10 <ncopa> kaniini: if i understand things correctly, if those hacked huawei routers woudl have had tcp forwarding off by default, they wouldnt have been able to do the irc attack
2018-07-29 15:33:28 <ncopa> kaniini: maybe submit that to openssh upstream?
2018-07-29 15:34:17 <ncopa> sounds like a good idea to disable
2018-07-29 15:37:45 <ncopa> danieli> ncopa seems to care, but you tend to get kinda jaded after being neck deep in code and problems all day every day
2018-07-29 15:37:49 <ncopa> spot on
2018-07-29 15:38:37 <ncopa> i do care, and i am sorry that i have not done more to streamline the process
2018-07-29 16:09:01 <rmbeer> hello...
2018-07-29 16:09:57 <rmbeer> I have problems configuring mariadb, where the executables like mysqladmin, mysql, and mysql_secure_installation are installed?...
2018-07-29 16:10:19 <rmbeer> not is in /usr/bin , and also can't found in other sides....
2018-07-29 16:19:54 <ncopa> danieli: this might be of interest http://lists.alpinelinux.org/alpine-devel/5811.html
2018-07-29 16:32:12 <rmbeer> ok, solve with mariadb-client....
2018-07-29 17:01:28 <danieli> ncopa: that one was linked last night, will give it a re-read
2018-07-29 17:09:26 <danieli> kaniini: wait... is it NOT disabled by default in alpine?
2018-07-29 17:09:34 <danieli> that's bad
2018-07-29 17:09:39 <danieli> well, overall
2018-07-29 18:35:35 <ncopa> i think i finally figured out what causes chromium go "oh snap"
2018-07-29 18:35:52 <ncopa> hopefully i will push update later tonight
2018-07-29 18:51:55 <iRoAk32gijXI> \o/ kicad 5.0.0 works in aports-ugly https://github.com/aports-ugly/aports/commit/3ea5baac9a858a99dbe88e75680c5b8102063b8c
2018-07-29 19:11:40 <danieli> ncopa: what'd you figure out?
2018-07-29 19:11:50 <danieli> man, chromium is a large piece of software
2018-07-29 19:11:58 <danieli> and components of it is everywhere nowadays
2018-07-29 19:12:05 <danieli> spotify, discord, slack, you name it
2018-07-29 19:13:05 <clandmeter> chromium has spotify code in it?
2018-07-29 19:13:18 <danieli> opposite
2018-07-29 19:13:22 <danieli> chromium has chromium code in it
2018-07-29 19:13:41 <danieli> components of it [chromium] is everywhere [in other software]
2018-07-29 19:14:06 <clandmeter> ah ok
2018-07-29 19:14:19 <clandmeter> i thought i missed something.
2018-07-29 19:14:19 <danieli> spotify actually doesn't use electron like many others do, they use CEF directly
2018-07-29 19:15:08 <danieli> electron used to be CEF but it switched to using chromium directly
2018-07-29 19:43:28 <terra> thanks for xorg-server 1.20
2018-07-30 00:55:02 <deedubs> SIGs have worked really well for Kubernetes
2018-07-30 01:09:14 <danieli> they're a necessity for most large open source projects
2018-07-30 01:09:26 <danieli> kinda like namespacing, for people
2018-07-30 01:35:00 <deedubs> SIG-Packaging SIG-Security SIG-Release
2018-07-30 01:43:41 <awilfox> it's working well for us too https://wiki.adelielinux.org/wiki/Project:Project_List
2018-07-30 01:45:14 <danieli> deedubs: in the first iteration, I'm thinking of SIGs for infrastructure, documentation and security
2018-07-30 03:13:33 <deedubs> awesome, It'll be hard to choose which one to participate in :D
2018-07-30 03:14:31 <awilfox> don't choose!  if you feel you have the time and energy, join multiple
2018-07-30 03:14:39 <deedubs> :D
2018-07-30 03:14:53 <awilfox> for example, you can write documentation while fixing infrastructure (if those are two SIGs that are created)
2018-07-30 03:16:54 <deedubs> danieli: That would be great, documentation is one area I think could make a significant impact. For me it would be getting best practices hammered out in package writing, since I do a ton of that
2018-07-30 03:19:28 <danieli> it definitely would, and security is super important
2018-07-30 03:19:42 <danieli> they're both concerns for many alpine users, and some of it is kind of hard to stay on top of
2018-07-30 03:23:45 <deedubs> I'm pumping out packages for aide, lynis, rkhunter as we speak
2018-07-30 03:32:32 <deedubs> I'm still determining how I'm going to scan systems for CVEs
2018-07-30 03:38:54 <danieli> scan systems?
2018-07-30 03:41:51 <deedubs> scan the software on a running system for known CVEs
2018-07-30 03:42:09 <danieli> writing a tool for it?
2018-07-30 03:42:24 <deedubs> I am not
2018-07-30 03:42:27 <danieli> assuming you have access to said system, I'd check versions of installed packages
2018-07-30 03:42:32 <danieli> okay, what's the occasion, if I may ask?
2018-07-30 03:42:34 <deedubs> I am considering doing that
2018-07-30 03:42:57 <deedubs> https://github.com/massivekube
2018-07-30 03:43:29 <deedubs> I'm using alpine as a base for a what is essentially a kubernetes distro
2018-07-30 03:43:37 <danieli> aha, gotcha
2018-07-30 03:43:46 <danieli> sounds pretty interesting
2018-07-30 03:44:34 <deedubs> so I need to be able to run periodic scans of a system, or at least the AMI that created it, to ensure no new known vunerabilities have been reported
2018-07-30 03:45:18 <deedubs> Most of the tools are pretty bloated or written in a interpreted lang which isn't ideal
2018-07-30 03:45:23 <danieli> if I were you, I'd match installed package versions against CVEs assigned to software
2018-07-30 03:45:34 <danieli> writing a tool like that would be really nice, actually
2018-07-30 03:45:37 <danieli> perhaps even `apk audit`
2018-07-30 03:45:59 <danieli> I mean, taking inspiration from `pkg audit [-F]`
2018-07-30 03:46:49 <deedubs> yeah I've been looking at the formatting https://nvd.nist.gov/vuln/data-feeds mulling over writing something in Go to do it
2018-07-30 03:47:31 <deedubs> while hoping someone else has done it aha
2018-07-30 03:54:36 <danieli> doing it in C in apk-tools would be an option
2018-07-30 03:56:14 <deedubs> Alas its been 7 years since I've written C, so it would be awesome to have it builtin, but I won't be the one who can do it
2018-07-30 03:56:27 <danieli> gotcha, it could be a job for a security SIG when it's up
2018-07-30 03:56:45 <danieli> there's still a little red tape and groundwork to go there, but it sounds doable
2018-07-30 03:56:58 <deedubs> Maybe I can POC it out for them
2018-07-30 03:57:21 <danieli> I'll help out
2018-07-30 03:57:29 <danieli> go isn't my strongest suit but it'd be nice to work on it a bit
2018-07-30 03:57:31 <deedubs> Gah a single CVE in the JSON feed is 1955 lines
2018-07-30 03:57:47 <danieli> that's not too bad
2018-07-30 03:59:31 <deedubs> there are 105187 CVEs registered :)
2018-07-30 03:59:39 <danieli> that's what I was thinking
2018-07-30 03:59:46 <danieli> I mean it's not bad in terms of size but it has to be way off
2018-07-30 04:00:22 <[[sroracle]]> you could scrape CVE data from secfixes in the APKBUILDs
2018-07-30 04:00:49 <deedubs> Wouldn't that only give us CVEs for things that were fixed?
2018-07-30 04:00:50 <danieli> well, I'd want to show things that aren't patched upstream (in alpine)
2018-07-30 04:00:58 <danieli> I'd also want*
2018-07-30 04:04:07 <danieli> besides, not all CVE fixes are in the APKBUILDs
2018-07-30 04:04:44 <danieli> ncopa worked on just this, but doing it fully automated won't be feasible
2018-07-30 04:07:10 <deedubs> https://gist.github.com/deedubs/cf53aaa002111c2cc2bd56547fa21c48
2018-07-30 04:07:45 <danieli> https://cve.mitre.org/data/downloads/index.html
2018-07-30 04:08:58 <deedubs> yeah, nist lets you get feeds so you don't need to get the whole db everytime
2018-07-30 04:09:05 <danieli> mm
2018-07-30 04:12:07 <danieli> `Content-Length: 16089806`
2018-07-30 04:12:27 <danieli> 16MB isn't too bad
2018-07-30 04:13:17 <danieli> wc -l of the allitems.csv list is 131463
2018-07-30 04:13:57 <danieli> cleartext CSV is 96MB
2018-07-30 04:19:45 <deedubs> I feel like we could boil those down to CPEs, the severity, description, and a link
2018-07-30 04:21:38 <deedubs> cpe:/o:linux:linux_kernel:2.6.0 HIGH "The get_random_int function in drivers/char/random.c in" CVE-2009-3238
2018-07-30 18:47:51 <deedubs>  https://github.com/massiveco/cve-check/tree/feature/initial
2018-07-30 22:23:21 <terra> nick terra_
2018-07-31 02:07:11 <danieli> ncopa: reminds me, does chrome 68 work? if so, nice work
2018-07-31 02:39:54 <deedubs> danieli: are we going to create #alpine-sig-security ;)
2018-07-31 02:40:00 <danieli> not just yet
2018-07-31 02:40:21 <danieli> there is an existing security effort by the way
2018-07-31 02:43:42 <deedubs> I have no doubt there is :)
2018-07-31 02:44:04 <danieli> it's relatively new as a dedicated effort iirc
2018-07-31 02:44:09 <deedubs> Is that happening on IRC
2018-07-31 02:44:14 <danieli> bugs.a.o
2018-07-31 02:44:15 <deedubs> err I mean mailing list
2018-07-31 02:44:29 <danieli> mailing lists too i suppose
2018-07-31 02:44:38 <danieli> it's old tho
2018-07-31 02:46:07 <deedubs> I'll just keep poking at cve-check
2018-07-31 09:58:11 <ncopa> danieli: chrome 68 works
2018-07-31 10:05:43 <clandmeter> ncopa, not everybody seem to agree with you.
2018-07-31 10:07:08 <Gottox> is it possible to view alpine issues without an account?
2018-07-31 10:07:42 <ncopa> Gottox: should be possible to view them yes
2018-07-31 10:08:01 <ncopa> clandmeter: someone have issues with chromium 68?
2018-07-31 10:08:14 <Gottox> I got a redmine login prompt if I click on the issue.
2018-07-31 10:08:23 <ncopa> clandmeter: ^^^
2018-07-31 10:08:25 <clandmeter> ah crap
2018-07-31 10:08:29 <ncopa> i think it was unintentional :)
2018-07-31 10:08:34 <clandmeter> yes it was
2018-07-31 10:08:35 <clandmeter> haha
2018-07-31 10:09:08 <clandmeter> fixed
2018-07-31 10:09:17 <clandmeter> hmm, need  to dig deeper
2018-07-31 10:09:26 <Gottox> thx :)
2018-07-31 10:09:39 <clandmeter> ncopa, look 1cm above your comment that it works.
2018-07-31 10:11:06 <awilfox> pickfire timing out? :)
2018-07-31 10:12:45 <clandmeter> ok really disabled unauth issue reporting
2018-07-31 14:49:32 <pickfire> awilfox: Yeah, sometimes.
2018-07-31 14:49:41 <pickfire> I am not sure why though.
2018-07-31 21:56:46 <trfl> when splitting a former py2-only package `foo` into py2-foo and py3-foo, should I fix up all packages that depend on it and include their APKBUILDs in the pullrequest as well?
2018-07-31 22:39:18 <trfl> and is it still preferable to include a test suite in check() when it takes over two hours to complete?
2018-07-31 22:45:52 <danieli> is it because the test is super resource intensive?
2018-07-31 22:46:44 <trfl> yes, CPU-bound for the most part, generating and compiling large amounts of C++ code then testing for success
2018-07-31 22:49:27 <trfl> afaict the test framework does not have a built-in way to parallelize them, so we'd have to do that manually by running bits and pieces in parallel
2018-07-31 22:53:21 <danieli> the builder machines are quite powerful, but i don't have access to any of them
2018-07-31 22:53:27 <danieli> ncopa probably knows better regarding that
2018-07-31 22:54:55 <trfl> oh wait there we go, yes it does support running tests in parallel! and I suppose there's no problem building several apk's in parallel, so one expensive apkbuild won't block the rest of the build
2018-07-31 22:55:49 <danieli> i suppose someone with access/first-hand knowledge could have a look at it
2018-07-31 22:55:57 <danieli> if you can parallelize it somehow, that'd be (y)
2018-07-31 23:07:26 <trfl> turns out the paralellization is orchestrated through /dev/shm, which appears to be 0755 root:root in the chroot that I run abuild in (alpine-inside-alpine). This makes the tests fail, and is fixed by setting /dev/shm 0777. Any idea what the /dev/shm permissions can be expected to be on travis and on other build servers?
2018-07-31 23:08:53 <Shiz> trfl: i can check the main build server
2018-07-31 23:08:55 <Shiz> let's see
2018-07-31 23:09:24 <Shiz> on the container *i* have access to it's 0777
2018-07-31 23:09:57 <Shiz> shiz-edge-x86_64:~# ls -ld /dev/shm
2018-07-31 23:09:59 <Shiz> drwxrwxrwt    2 root     root            40 Apr 10 08:08 /dev/shm
2018-07-31 23:10:08 <Shiz> as far as cpu-bound tests go
2018-07-31 23:10:17 <Shiz> I think we have actualy disabled tests before because they took too much time
2018-07-31 23:10:28 <danieli> we have indeed, i remember it happening
2018-07-31 23:14:31 <trfl> sweet! then I'll leave them in for now and have it use $JOBS threads