2018-08-01 01:53:04 <deltam> With our IRC ad service you can reach a global audience of entrepreneurs and fentanyl addicts with extraordinary engagement rates! https://williampitcock.com/
2018-08-01 16:24:16 <ncopa> algitbot: hi
2018-08-01 16:24:32 <_ikke_> Well, let me rephrase, I don't have the alpine developer role :)
2018-08-01 16:25:31 <ncopa> trfl: do you have reported it on bugs.a.o?
2018-08-01 16:26:26 <trfl> no, I will do that and link some references that seem related :>
2018-08-01 16:26:47 <ncopa> would be great
2018-08-01 16:27:06 <ncopa> Shiz: have you looked anything on julia package?
2018-08-01 17:37:24 <_ikke_> Does this help PureTryOut[m] ?
2018-08-01 17:38:26 <PureTryOut[m]> it does!
2018-08-01 17:38:28 <PureTryOut[m]> but it won t help other Matrix users sadly
2018-08-01 17:38:32 <PureTryOut[m]> I guess you now require voice or something?
2018-08-01 17:38:46 <_ikke_> PureTryOut[m]: there was temporar a +m mode, but that has been lifted
2018-08-01 17:39:13 <_ikke_> how about now?
2018-08-01 17:40:19 <PureTryOut[m]> on the Matrix side a power of 1 is still required, blocking most Matrix users from talking
2018-08-01 17:41:01 <_ikke_> You don't have +v anymore, but can still talk
2018-08-01 17:42:54 <_ikke_> seems like something is out of sync with them matrix bridge, can that be?
2018-08-01 17:43:48 <PureTryOut[m]> not normally
2018-08-01 17:43:55 <PureTryOut[m]> what did you do to give me a power of 1 then?
2018-08-01 17:44:11 <_ikke_> mode +v
2018-08-01 17:44:23 <_ikke_> but I removed it again as well
2018-08-01 18:01:01 <Shiz> ncopa: not yet
2018-08-01 18:01:15 <Shiz> kaniini: still the huawei routers? :p
2018-08-01 19:25:28 <kaniini> Shiz a smorgasboard of shitty IOT devices
2018-08-02 10:03:55 <andypost> ncopa, as I see php7-stats fail on s390x http://build.alpinelinux.org/buildlogs/build-edge-s390x/testing/php7-stats/php7-stats-2.0.3-r4.log
2018-08-02 10:04:40 <andypost> does it makes sense to rm tests for the arch? I bet package is incompatible with it
2018-08-02 10:05:00 <ncopa> i'd rather disable the arch
2018-08-02 10:23:33 <andypost> ncopa, here's is it https://github.com/alpinelinux/aports/pull/4913 + 4912 makes sense to backport to 3.8
2018-08-02 10:27:47 <ncopa> feel free to merge https://github.com/alpinelinux/aports/pull/4913
2018-08-02 10:28:28 <ncopa> so you want backport c24613c34e96cc9e0912ae5ebe11a1690cd189f6?
2018-08-02 10:29:18 <ncopa> sorry
2018-08-02 10:30:12 <andypost> ncopa, http://dup.pw/aports/128fba54 this one
2018-08-02 10:30:27 <ncopa> ok
2018-08-02 10:31:18 <andypost> ncopa, http://dup.pw/aports/fdfd7f76 merged
2018-08-02 10:31:57 <ncopa> thanks
2018-08-02 10:32:33 <ncopa> andypost: thank you for taking care of the php stuff. it is appreciated
2018-08-02 10:33:02 <clandmeter> yes ❤️
2018-08-02 10:35:23 <andypost> np) license fixes coming
2018-08-02 11:31:57 <liwakura> apk --no-network seems to be largely ignored
2018-08-02 11:32:17 <liwakura> giving the apkindex cache an very old mtime causes apk reliably to re-fetch it
2018-08-02 11:33:27 <liwakura> i figure thats also the cause for #9126
2018-08-02 11:46:30 <clandmeter> liwakura: i think it would be nice if we had a readme on mkinitfs repo.
2018-08-02 11:51:53 <liwakura> for github?
2018-08-02 11:52:09 <clandmeter> no just in general
2018-08-02 11:52:25 <liwakura> uhm
2018-08-02 11:52:28 <clandmeter> we dont host anything on github officially
2018-08-02 11:52:33 <liwakura> for the other PR, wait with that merge
2018-08-02 11:52:45 <clandmeter> which one?
2018-08-02 11:53:26 <liwakura> 34, [WIP] initramfs-init: Always configure network when ip= is given
2018-08-02 11:53:39 <clandmeter> np
2018-08-02 11:54:23 <clandmeter> i just merge the man pages. it should not break anything.
2018-08-02 11:54:50 <liwakura> not sure what to put into the readme, since github is the only case where i can imagine people looking at it
2018-08-02 11:54:59 <liwakura> maybe compilation instructions?
2018-08-02 11:55:06 <clandmeter> exactly
2018-08-02 11:55:18 <clandmeter> short note what it is
2018-08-02 11:55:22 <clandmeter> and to use it.
2018-08-02 11:55:42 <clandmeter> cause i just tried to build it and i need to figure out which deps it needs.
2018-08-02 11:55:53 <clandmeter> i tried to open the readme...
2018-08-02 11:56:11 <clandmeter> imho all our project should have a readme.
2018-08-02 11:56:30 <liwakura> apk is also lacking a manpage, but i want to focus myself at mkinitfs first
2018-08-02 11:56:41 <clandmeter> there is a PR for that.
2018-08-02 11:56:46 <clandmeter> or was it abuild
2018-08-02 11:57:02 <clandmeter> i think its abuild.
2018-08-02 13:22:06 <bernhardgruen> I am finally back after some burnout prevention timeout where I just read from time to time. I saw some messages with nice ideas how to improve PRs. I hope this helps to reduce the burden.
2018-08-02 13:26:01 <_ikke_> bernhardgruen: Welcome back!
2018-08-02 13:26:29 <_ikke_> Good that you noticed it on time and took preventative measures
2018-08-02 13:28:05 <bernhardgruen> Thank you _ikke_.
2018-08-02 15:08:22 <bernhardgruen> I have a question about -openrc subpackages. Should every aport (which has an openrc file) have such a subpackage? Or is there no need to use that type of subpackage?
2018-08-02 15:09:22 <_ikke_> bernhardgruen: I think they are moving to that, so that you can switch out init systems easier later
2018-08-02 15:10:27 <bernhardgruen> _ikke_, ok, I will test it for some of my packages and update them accordingly.
2018-08-02 15:29:09 <Shiz> bernhardgruen: you just need to add it to subpackages= btw
2018-08-02 15:29:15 <Shiz> and abuild will warn if you don't
2018-08-02 15:29:19 <Shiz> it'll move hte files itself
2018-08-02 15:31:30 <bernhardgruen> Shiz, thank you. I will test it for Docker. Do I (as an user) still just need to install with 'apk add docker' or do I have to install openrc additionally? That would make things less user friendly IMHO.
2018-08-02 15:31:43 <Shiz> nope
2018-08-02 15:31:46 <Shiz> there's an install_if rule :)
2018-08-02 15:31:54 <Shiz> so it'll auto-install -openrc subpackages if you have openrc installed
2018-08-02 15:33:18 <bernhardgruen> Nice.
2018-08-02 15:44:00 <bernhardgruen> Just updated my PR for docker. Works locally like a charm as long as I install the packages via repository (and not via 'apk add docker-18.08.0-r0.apk')
2018-08-02 17:25:50 <andypost> kaniini, can you help with https://bugs.alpinelinux.org/issues/9119 jirutka said it depends on provider_priority but I can't find docs about it
2018-08-02 17:47:03 <clandmeter> andypost, when two packages provides the same the priority determines which one is installed
2018-08-03 11:52:16 <danieli> we don't have python 3.7 yet, do we?
2018-08-03 12:46:43 <liwakura> no
2018-08-03 12:46:56 <liwakura> https://pkgs.alpinelinux.org/packages?name=python3&branch=edge
2018-08-03 16:26:56 <clandmeter> to everyone: we are going to take offline patchwork and bugs to do maintenance. hope to get it back asap.
2018-08-03 17:43:56 <andypost> any idea how properly give a name to subpackage "-ui" or "-lib" it it contains html+scripts https://github.com/alpinelinux/aports/pull/4923#discussion_r207615553
2018-08-03 17:54:44 <clandmeter> andypost, assets?
2018-08-03 17:55:59 <andypost> clandmeter, nice idea! https://pkgs.alpinelinux.org/packages?name=*assets&branch=edge
2018-08-03 17:59:21 <andypost> clandmeter, is there kind of suggests/recommends for packages?
2018-08-03 17:59:47 <clandmeter> im in the middle of a migration. please hold the line.
2018-08-03 18:00:12 <clandmeter> i dont think so
2018-08-03 18:00:17 <clandmeter> what ever makes sense
2018-08-03 18:00:26 <clandmeter> if you think you did something smart, doc it on wiki.
2018-08-03 18:00:45 <clandmeter> if someobdy doesnt agree let them updte it :)
2018-08-03 18:21:11 <andypost> )) I just found no examples where after install of package user could get a message about - you may need to install something else to get better)
2018-08-03 18:22:02 <andypost> example is "optdepends" http://limyreth.sinners.be/files/1/php-xhprof/PKGBUILD
2018-08-03 20:06:59 <clandmeter> andypost, post install messages is not really implemented
2018-08-03 20:07:10 <clandmeter> its possible with post-install but its not prefered.
2018-08-03 20:07:31 <andypost> guess not usable actually
2018-08-03 20:07:45 <clandmeter> we disucssued it a few times
2018-08-03 20:08:14 <clandmeter> the idea is to put a message file in /usr/share/pkgname/xxx iirc
2018-08-03 20:08:32 <clandmeter> and after apk has done all its processing it would give you the messages
2018-08-03 20:08:43 <clandmeter> not after each pkg which is messy
2018-08-03 20:09:06 <clandmeter> but it will also need logic to when it should be displayed
2018-08-03 20:09:17 <clandmeter> add upgrade or delete...
2018-08-03 20:09:41 <clandmeter> it needs to be version aware which is technically possible.
2018-08-03 20:09:52 <andypost> oh sure.. more messy if more then 1 package will display
2018-08-03 20:10:18 <clandmeter> even more messy if you run from ram
2018-08-03 20:10:31 <clandmeter> everu reboot would be a wall painting of messages
2018-08-03 20:10:54 <andypost> then I'm done with packaging)
2018-08-03 20:11:04 <clandmeter> but i think in the end nobody cared enough to really fix it.
2018-08-03 20:11:51 <clandmeter> andypost, did you understnad the priority thing?
2018-08-03 20:12:19 <andypost> not yet, but gonna start it right now
2018-08-03 20:12:26 <andypost> clandmeter, can you briefly review https://github.com/alpinelinux/aports/pull/4923/files#diff-8b2a586fb2ec85d46f431ead467f5c9aR49
2018-08-03 20:12:35 <clandmeter> take a look how we do it with lua
2018-08-03 20:12:44 <clandmeter> we have 3 versions
2018-08-03 20:12:55 <clandmeter> all 3 provide lua
2018-08-03 20:13:04 <clandmeter> but the highest priority will win
2018-08-03 20:13:08 <andypost> I seen that but that 3 different versions
2018-08-03 20:13:24 <andypost> but in my case it is subpackages
2018-08-03 20:13:38 <clandmeter> subpkg?
2018-08-03 20:14:19 <clandmeter> andypost, what is there to see?
2018-08-03 20:14:35 <andypost> yes, if I give more priority to php5-cli then php5 is not used but all other subpaks depends on php5 (which is empty)
2018-08-03 20:15:26 <clandmeter> i need to know what you try to achieve.
2018-08-03 20:15:35 <andypost> probably better to move files to php5 with priority and keep php5-cli empty
2018-08-03 20:16:15 <andypost> https://bugs.alpinelinux.org/issues/9119#note-3
2018-08-03 20:17:44 <andypost> I need to keep php5-cli for BC but require php5 to install its files
2018-08-03 20:22:44 <clandmeter> ok i need to look closer to it.
2018-08-03 20:22:49 <clandmeter> but not now
2018-08-03 20:22:57 <clandmeter> im soaking from this crazy weather
2018-08-04 03:12:33 <rmbeer> hello...
2018-08-04 03:13:05 <rmbeer> mariadb have the log off by default... it's a problem, why mariadb break in 10min and i unknown how to solve this error....
2018-08-04 13:42:23 <andypost> can someone close the issue https://bugs.alpinelinux.org/issues/9179
2018-08-05 19:21:18 <PureTryOut[m]> ok so to resolve some failing dependencies gnome-shell, gnome-session and mutter have to be rebuild. mutter was easy enough, but gnome-shell is 2 versions behind mutter. getting it to the same version is not as easy though
2018-08-05 19:21:49 <_ikke_> What are the problems you are facing?
2018-08-05 19:21:57 <PureTryOut[m]> first of all it requires some new (not yet packaged) dependencies like ibus. I just made an APKBUILD for it though. now it also requires libsystemd, which is obviously a problem...
2018-08-05 19:22:13 <_ikke_> yes, that is a problem
2018-08-05 19:22:20 <_ikke_> and it's not something you can disable during configuration?
2018-08-05 19:22:24 <maxice8> should be
2018-08-05 19:22:35 <PureTryOut[m]> did gnome-shell 3.26 just not need libsystemd? I see no patch to get it out
2018-08-05 19:22:36 <PureTryOut[m]> not entirely sure about that yet
2018-08-05 19:22:52 <_ikke_> ./configure --help?
2018-08-05 19:23:08 <maxice8> gnome-shell has -Dsystem=false with meson
2018-08-05 19:23:18 <PureTryOut[m]> actually, ibus is not an option to disable, but systemd seems to be
2018-08-05 19:23:34 <_ikke_> not sure if ibus is a problem
2018-08-05 19:23:54 <PureTryOut[m]> no they switched from configure to Meson
2018-08-05 19:23:55 <PureTryOut[m]> not so much, the APKBUILD was easy enough
2018-08-05 19:24:05 <PureTryOut[m]> I don't really want to maintain it though...
2018-08-05 19:25:09 <_ikke_> right
2018-08-05 19:25:39 <PureTryOut[m]> also they now have an optional dependency on network manager, do we want it?\
2018-08-05 19:25:54 <_ikke_> not by default I guess
2018-08-05 19:26:24 <PureTryOut[m]> ok so build time dependency yes, runtime dependency no then
2018-08-05 19:26:42 <_ikke_> right
2018-08-05 19:44:03 <PureTryOut[m]> hmm to build with networkmanager support it requires some changes. I'm just disabling it for now to at least fix the install error
2018-08-05 21:28:04 <rmbeer> how to deactivate --syslog from mariadb?... what file i must modify from the system?...
2018-08-05 23:17:14 <PureTryOut[m]> awilfox: I see you updated the patch to make netgroup support optional in polkit. are you planning on updating polkit to a newer version anytime soon? https://bugs.freedesktop.org/show_bug.cgi?id=50145
2018-08-05 23:18:36 <awilfox> apk policy polkit
2018-08-05 23:18:38 <awilfox> polkit policy:
2018-08-05 23:18:40 <awilfox>   0.115-r0:
2018-08-05 23:18:42 <awilfox>     lib/apk/db/installed
2018-08-05 23:18:44 <awilfox>     /srv/packages/user
2018-08-05 23:21:09 <awilfox> see also https://github.com/sabotage-linux/gettext-tiny/issues/27
2018-08-05 23:31:45 <awilfox> PureTryOut: we also just released 1.0-beta1-snapshot which features plasma-addons finally: https://wiki.raptorcs.com/wiki/File:Screenshot_20180803_060755.png
2018-08-06 06:35:43 <azarus> Guys, have you noticed how Gimp 2.10 doesn't have the missing icon problem anymore? :D
2018-08-06 06:36:04 <azarus> Makes using it so much more pleasant than 2.8.
2018-08-06 07:35:28 <trfl> noticed a cosmetic bug in the package browser, if a package depends on multiple libraries from another package, that dependant package is repeated in the list https://pkgs.alpinelinux.org/package/edge/testing/x86_64/mgba
2018-08-06 07:57:57 <clandmeter> trfl, ive seen it. can you create an issue for it on bugs.a.o
2018-08-06 07:58:11 <clandmeter> project alpine-infra
2018-08-06 10:29:33 <mercury^> Hi. I issued some pull requests to aports about three weeks ago. There are no unanswered comments or requests, so I am wondering if anything else is expected of me before they are merged or rejected.
2018-08-06 10:30:11 <_ikke_> mercury^: Probably not, there is just a bit of a backlog
2018-08-06 10:30:25 <_ikke_> We're trying to do better in this regards, we just need to find a good way
2018-08-06 10:36:29 <mercury^> Another thing: when I configured my x86_64 kernel starting from the vanilla configuration, I found that quite a few obscure options are enabled, while support for some (moderately outdated) mainstream hardware as well as some features that I wanted was missing. Is there a rationale for the configuration somewhere?
2018-08-06 10:37:07 <clandmeter> the only rationale would be git log.
2018-08-06 17:28:44 <wlp0s29f7u3> Shiz: your -Z patch for syslogd seems to be breaking on 32-bit
2018-08-06 17:29:46 <wlp0s29f7u3> currently investigating
2018-08-06 17:30:07 <wlp0s29f7u3> there is some -1 coming into 'now'
2018-08-06 17:44:41 <liwakura> ah, #8028
2018-08-06 20:56:03 <liwakura> clandmeter: and boot.alpinelinux.org is 404'ing, the whole images/ folder
2018-08-06 21:50:53 <clandmeter> liwakura, please add an issue
2018-08-07 07:25:27 <danieli> clandmeter: did boot.a.o get killed? ^
2018-08-07 07:37:45 <clandmeter> nee
2018-08-07 07:37:53 <clandmeter> its working just not those dirs
2018-08-07 07:43:00 <clandmeter> i need to fix it, but im busy fixing other things.
2018-08-07 07:53:19 <danieli> gotcha
2018-08-07 07:58:06 <clandmeter> ncopa, when do you get a chance to setup edge snapshotting?
2018-08-07 08:03:15 <andypost> anyone have s390x hardware to run tests on php7-xhprof package?
2018-08-07 08:03:53 <clandmeter> heh
2018-08-07 08:04:06 <clandmeter> if you could even get it, you cannot afford it :)
2018-08-07 08:04:19 <clandmeter> we have a builder though
2018-08-07 08:04:29 <danieli> which is only for core devs, as far as i understand
2018-08-07 08:04:46 <danieli> so that helps with just about nothing
2018-08-07 08:05:00 <andypost> clandmeter, do you mean to push package with patch? but then I can't get remains in src dir
2018-08-07 08:05:20 <clandmeter> im not sure what the policy is regarding containers. i dont  manage this arch.
2018-08-07 08:05:57 <andypost> clandmeter, I'm sure it needs real HW cos about profiling and HW-timers
2018-08-07 08:06:02 <andypost> tl'dr https://github.com/longxinH/xhprof/issues/15#issuecomment-410904643
2018-08-07 08:27:10 <clandmeter> andypost, there are only two persons with access to that arch, ncopa is one of them.
2018-08-07 08:27:55 <andypost> clandmeter, then I think better to commit patch & wait builder run
2018-08-07 10:25:24 <andypost> ncopa, follow-up to fail2ban with fix https://github.com/alpinelinux/aports/pull/4967
2018-08-07 10:39:27 <ncopa> andypost: thanks!
2018-08-07 11:35:07 <andypost> clandmeter, any idea how to find why package was not uploaded to repo?
2018-08-07 11:35:22 <clandmeter> which pkg?
2018-08-07 11:35:33 <clandmeter> which arch?
2018-08-07 11:35:36 <andypost> r1 build passed on s390x but no log http://build.alpinelinux.org/buildlogs/build-edge-s390x/testing/php7-xhprof/
2018-08-07 11:36:06 <andypost> actually 3 arch only uploaded https://pkgs.alpinelinux.org/packages?name=php7-xhprof*&branch=edge
2018-08-07 11:36:15 <andypost> fo others not logs
2018-08-07 11:36:40 <andypost> http://build.alpinelinux.org/buildlogs/build-edge-x86.log I see only debug mention
2018-08-07 11:37:06 <clandmeter> how do you know r1 is build?
2018-08-07 11:37:28 <andypost> I seen it at build-a-o
2018-08-07 11:37:40 <clandmeter> andypost, i think the builder is stuck on community
2018-08-07 11:37:55 <andypost> but this is testing
2018-08-07 11:38:30 <clandmeter> yes
2018-08-07 11:38:54 <clandmeter> im not sure how the logic is exactly, could be its holds the upload until finished building.
2018-08-07 11:39:08 <clandmeter> ncopa, can you clarify?
2018-08-07 11:41:48 <ncopa> hi
2018-08-07 11:42:39 <ncopa> andypost: if a repo is not built then it is not uploaded
2018-08-07 11:42:45 <ncopa> i mean
2018-08-07 11:42:53 <ncopa> if a build in a repo fails, then it does not upload anything
2018-08-07 11:43:00 <ncopa> for that repo
2018-08-07 11:43:17 <ncopa> http://build.alpinelinux.org/buildlogs/build-edge-x86.log says that 'hub' fails to build
2018-08-07 11:43:50 <clandmeter> i think he means s390x
2018-08-07 11:43:55 <andypost> ncopa, so it means while any package broken others will not build
2018-08-07 11:44:13 <ncopa> in that repo. correct
2018-08-07 11:44:18 <clandmeter> ncopa, what is a repo?
2018-08-07 11:44:22 <clandmeter> main?
2018-08-07 11:44:25 <ncopa> yes
2018-08-07 11:44:39 <ncopa> so it build everytihng in main, if it fails it aborts
2018-08-07 11:44:51 <ncopa> and nothing is uploaded
2018-08-07 11:44:53 <clandmeter> right
2018-08-07 11:45:04 <ncopa> if main succeeds, it will upload main and go on to community
2018-08-07 11:45:09 <ncopa> if it fails, it aborts
2018-08-07 11:45:09 <clandmeter> but if something fails in community, does things in main get uploaded?
2018-08-07 11:45:15 <ncopa> yes
2018-08-07 11:45:18 <clandmeter> ok
2018-08-07 11:45:30 <clandmeter> but andypost says his aport has already been build.
2018-08-07 11:45:41 <ncopa> its in testing, right?
2018-08-07 11:45:41 <clandmeter> and current it holds on community
2018-08-07 11:45:48 <clandmeter> yes
2018-08-07 11:45:49 <clandmeter> testing
2018-08-07 11:45:57 <ncopa> which arch?
2018-08-07 11:46:07 <ncopa> aarch64 is failing in main
2018-08-07 11:46:18 <ncopa> s390x is failing in testing
2018-08-07 11:46:27 <clandmeter> r1 build passed on s390x but no log http://build.alpinelinux.org/buildlogs/build-edge-s390x/testing/php7-xhprof/
2018-08-07 11:46:32 <ncopa> x86 is failing in testing
2018-08-07 11:47:07 <clandmeter> ah right
2018-08-07 11:47:07 <andypost> ncopa, and I see no log for s390x like for x86
2018-08-07 11:47:12 <clandmeter> i think i looked wrong.
2018-08-07 11:47:24 <clandmeter> thats normal
2018-08-07 11:47:28 <clandmeter> its an external builder
2018-08-07 11:47:32 <clandmeter> it does not do nfs
2018-08-07 11:47:36 <clandmeter> so it does post upload logs
2018-08-07 11:47:37 <ncopa> if build server decides to build the failing package first, then it will abort before your package is built
2018-08-07 11:48:23 <ncopa> i dont know how to fix the aarch64 server
2018-08-07 11:48:26 <clandmeter> so we need to fix watchman to get s390x building again
2018-08-07 11:48:45 <ncopa> its linux-headers that has conflicting identifiers with libc headers
2018-08-07 11:49:00 <clandmeter> that sounds familiar
2018-08-07 11:49:04 <ncopa> i can have a look at hub on x86
2018-08-07 11:49:32 <andypost> clandmeter, I guess watchman needs to rm broken test for s390x
2018-08-07 11:49:49 <clandmeter> andypost, yes, or disable it completely.
2018-08-07 11:50:10 <clandmeter> we cannot test s390x
2018-08-07 11:50:23 <clandmeter> so if the tests reports its broken, we should trust the test.
2018-08-07 11:50:50 <clandmeter> report the error upstream and disable the aport (or fix if you know how to do it).
2018-08-07 11:51:00 <clandmeter> disable for that arch.
2018-08-07 11:53:58 <ncopa> looks like nbtscan disapeared
2018-08-07 11:54:01 <ncopa> from upstream
2018-08-07 12:42:21 <andypost> ncopa, better to disable watchman for s390x cos never version fails on x86-64
2018-08-07 12:42:35 <andypost> *s/never/newer
2018-08-07 12:43:19 <andypost> https://github.com/alpinelinux/aports/pull/4968
2018-08-07 12:52:25 <ncopa> andypost: maybe report it upstream?
2018-08-07 12:52:25 <ncopa> it seems to fail on 32bit x86 too
2018-08-07 12:55:12 <andypost> ncopa, not sure it makes sense - last release is year ago and therea re tons of commits after that
2018-08-07 12:55:54 <Shiz> kaniini: freenode decided to use your antissh? :P
2018-08-07 12:56:03 <kaniini> yes
2018-08-07 12:56:20 <kaniini> in part
2018-08-07 20:54:53 <ayee> "ERROR: http://dl-cdn.alpinelinux.org/alpine/v3.4/main: temporary error (try again later)"
2018-08-07 20:55:24 <ayee> My Dockerfile building is failing, this is a container from 9 months ago or so. Did something change? or am I pebkac'ing?
2018-08-07 20:55:56 <_ikke_> Repo seems to work for me
2018-08-07 20:56:09 <_ikke_> do you have internet access?
2018-08-07 20:56:38 <_ikke_> usually it means it cannot resolve the host or not reach the repo
2018-08-07 20:59:01 <unmy> isn't dl-cdn some kind of round robin domain?
2018-08-07 20:59:15 <_ikke_> it's a reverse proxy
2018-08-07 20:59:43 <_ikke_> distributed
2018-08-07 20:59:45 <unmy> because not all mirrors got older version of alpine like 3.4
2018-08-07 20:59:59 <_ikke_> It's backed by dl-4
2018-08-07 21:00:45 <_ikke_> https://mirrors.alpinelinux.org/
2018-08-07 21:00:50 <unmy> ah ok, dl-3 dont have 3.4
2018-08-07 21:01:31 <unmy> forgot that alpine got such a sweet mirror status page :)
2018-08-07 21:14:49 <ayee> hmm, I restarted docker and it works. Thanks for the help! I think it was pebkac. My corporate network is annoying. :(
2018-08-08 06:31:37 <ncopa> i think those 'temporary error (try again later)' are due to fastly intentionally breaks PMTU
2018-08-08 06:42:11 <detha> what is fastly up to with that?
2018-08-08 07:16:52 <ncopa> detha: their response was: its-a-feature
2018-08-08 07:17:20 <ncopa> i believe they drop icmp packets as part of their DoS protection
2018-08-08 07:17:28 <ncopa> many does that
2018-08-08 07:17:40 <ncopa> the workaround is to configure clamp-mss
2018-08-08 07:18:27 <ncopa> this video explains the problem https://blog.ipspace.net/2013/01/tcp-mss-clamping-what-is-it-and-why-do.html
2018-08-08 07:19:02 <ncopa> interestingly they dont say that filtering ICMP (breaking PMTU) is a bad thing
2018-08-08 07:19:56 <awilfox> https://bugs.kde.org/show_bug.cgi?id=354585 similar issue happened at kde's cdn
2018-08-08 07:20:03 <awilfox> they managed to make them fix it after enough complaints
2018-08-08 07:20:34 <detha> ncopa: sounds off. They may limit ICMP (and if they didn't limit ICMP to control planes it would be bad), but I can't imagine them filtering segmentation required
2018-08-08 07:21:13 <ncopa> detha: im pretty sure they do
2018-08-08 07:21:16 <awilfox> see the KDE bug
2018-08-08 07:21:25 <awilfox> it's very common to block ICMP because people think it is "hackable" or w/e
2018-08-08 07:21:31 <awilfox> the truth is IPv6 does not work with any ICMP blocked
2018-08-08 07:21:44 <awilfox> and any provider who gives IPv6 with ICMP filtered or blocked in any way is not giving real IPv6
2018-08-08 07:22:04 <awilfox> this is a part of the actual protocol spec
2018-08-08 07:22:57 <ncopa> who is hosting cdn.kde?
2018-08-08 07:23:01 <ncopa> they host it themselves?
2018-08-08 07:23:03 <awilfox> incapsula
2018-08-08 07:23:14 <detha> There are some fastly folks on freenode, so one can ask
2018-08-08 07:23:22 <awilfox> it's german provider, very near the KDE e.V. office
2018-08-08 07:24:02 <ncopa> when i reported the problem to fastly, with a reproducer, they basically said: "you can reduce your mtu manually" and "we dont do anything wrong on our side"
2018-08-08 07:24:13 <ncopa> i talked with them on irc but they asked me to send an email
2018-08-08 07:24:54 <awilfox> ncopa: also I just a few hours ago found out via #musl that freenode has loaded the necessary module in to their network so you can make unregistered users able to join but not talk
2018-08-08 07:24:56 <detha> My guess would be that it is dropped somewhere between client and fastly, not that uncommon alas
2018-08-08 07:25:18 <awilfox> ncopa: `/mode -r` and then `/mode +q $~a`
2018-08-08 07:25:23 <ncopa> awilfox: nice
2018-08-08 07:25:53 <ncopa> should work now?
2018-08-08 07:25:55 <awilfox> yay :)
2018-08-08 07:26:03 <ncopa> awilfox: thank you!
2018-08-08 07:27:54 <awilfox> 02:28 <-- freenode PRIVMSG #alpine-devel :hi
2018-08-08 07:27:56 <awilfox> 02:28 --> freenode :niven.freenode.net 404 awilcox #alpine-devel :Cannot send to channel
2018-08-08 07:27:58 <awilfox> yep
2018-08-08 07:28:00 <awilfox> it works :D
2018-08-08 07:28:26 <clandmeter> i think that was loaded some time ago
2018-08-08 07:28:33 <clandmeter> we already discussed it here :)
2018-08-08 07:28:42 <awilfox> ah, I just saw it on #musl this afternoon
2018-08-08 07:28:43 <clandmeter> so yes please -r everywehre
2018-08-08 07:29:07 <ncopa> clandmeter: have we set all our #alpine-* channels like that?
2018-08-08 07:29:36 <clandmeter> no
2018-08-08 07:29:49 <clandmeter> i got disconnected last night (i guess you too?)
2018-08-08 07:29:56 <clandmeter> i couldnt join any channel.
2018-08-08 07:30:10 <awilfox> yes you both left at same time:
2018-08-08 07:30:12 <awilfox> 21:41 <-- clandmeter (~clandmete@alpine/clandmeter) has quit (Quit: Ping timeout (120 seconds))
2018-08-08 07:30:14 <awilfox> 21:41 <-- ncopa (~ncopa2@alpine/ncopa) has quit (Quit: Ping timeout (120 seconds))
2018-08-08 07:30:18 <clandmeter> same app :)
2018-08-08 07:30:24 <awilfox> ah yeah
2018-08-08 07:30:26 <clandmeter> we are the same person actually
2018-08-08 07:30:30 <clandmeter> but dont tell
2018-08-08 07:30:31 <awilfox> XD
2018-08-08 07:30:31 <ncopa> :D
2018-08-08 07:30:42 <clandmeter> i have two keyboards ;-)
2018-08-08 07:30:51 <awilfox> hehe
2018-08-08 07:31:09 <ncopa> speaking of keyboard i have a kernel patch for my keyboard i need to upstream...
2018-08-08 07:32:29 <awilfox> oh yeah, that bluetooth patch.  I think we integrated that one in to easy-kernel
2018-08-08 07:32:49 <awilfox> looked helpful
2018-08-08 07:32:59 <ncopa> i think i just need to rebase it and send it to proper maintainer
2018-08-08 10:55:41 <andypost> ncopa, please help to test the patch on s390x https://github.com/alpinelinux/aports/pull/4975
2018-08-08 10:58:25 <ncopa> XHProf: Sampling Mode Test
2018-08-08 10:58:25 <ncopa> Author: kannan [tests/xhprof_008.phpt]
2018-08-08 10:58:35 <ncopa> those failed
2018-08-08 10:58:43 <ncopa> i'll continue after lunch
2018-08-08 10:59:28 <andypost> ncopa, thanx a lot
2018-08-08 11:26:10 <PureTryOut[m]> has anyone of you ever encountered "locale::facet::_S_create_c_locale name not valid" with some program and if so, how did you guys fix it? exporting LC_ALL=C doesn't fix it for me here
2018-08-08 13:30:35 <ncopa> andypost #9205 we need move it to community before we can backport it to 3.8
2018-08-09 19:47:22 <liwakura> clandmeter: readme for mkinitfs is PR'ed
2018-08-09 21:53:17 <muhaha>  Can anyone help? I am trying to build sscep https://pastebin.com/CDhviw0R  -> gcc: error: ../openssl/libcrypto.a: No such file or directory
2018-08-10 07:19:38 <muhaha> Can anyone help? I am trying to build sscep https://pastebin.com/CDhviw0R..  -> gcc: error: ../openssl/libcrypto.a: No such file or directory
2018-08-10 07:19:58 <muhaha> seems there is problem with libraries between 1.0 and 1.1
2018-08-10 09:02:15 <liwakura> muhaha: libcrypto.a is usually part of libressl-dev
2018-08-10 09:02:54 <liwakura> ah, i see, openssl-dev should also do the job..
2018-08-10 09:03:18 <muhaha> liwakura: but it does not work either
2018-08-10 09:59:50 <ncopa> muhaha: its normally not a good idea to mix openssl with libressl, so I'd recommend using libressl-dev
2018-08-10 10:00:00 <ncopa> what is the exact error message you get with libressl?
2018-08-10 10:05:05 <muhaha> @ncopa: Problem is with https://github.com/certnanny/sscep/blob/master/Linux/Makefile , there is $(OPENSSL)/libcrypto.a  , make OPENSSL=/usr/lib works OK
2018-08-10 10:05:45 <ncopa> i filed an issue for building with libressl: https://github.com/certnanny/sscep/issues/84
2018-08-10 10:07:25 <muhaha> try it with openssl-dev
2018-08-10 10:07:54 <muhaha> there is also problem with gcc flags..
2018-08-10 10:07:59 <muhaha> -lz can fix it
2018-08-10 10:12:15 <ncopa> makefile is broken
2018-08-10 10:12:24 <ncopa> you should report it upstream
2018-08-10 10:12:34 <ncopa> i suppose building the dynamic should work
2018-08-10 10:13:17 <ncopa> yup
2018-08-10 10:16:29 <ncopa> this works: http://tpaste.us/ZPbe
2018-08-10 10:16:41 <ncopa> if you need static build you need fix the Makefile
2018-08-10 10:20:28 <ncopa> http://tpaste.us/lW9w
2018-08-10 10:25:26 <muhaha> Superb! Thank You
2018-08-10 10:41:45 <ncopa> yw
2018-08-10 12:05:01 <andypost> ncopa, is it backportable to 3.8? https://github.com/alpinelinux/aports/pull/4981
2018-08-10 12:06:21 <ncopa> andypost: i read "package was broken ..." and stopped :) yes. if it is a bugfix, then it can be backported
2018-08-10 12:06:45 <ncopa> question is if fix is correct and does not introduce new breakages
2018-08-10 12:07:07 <andypost> I did test and now it works)
2018-08-10 12:07:25 <andypost> I was used httpd in foreround
2018-08-10 12:08:04 <ncopa> i dont have much experience with apache. what does "Require all granted" do?
2018-08-10 12:08:12 <ncopa> I hope it does not give access to everything everywhere
2018-08-10 12:13:26 <ncopa> andypost: im not happy with default config
2018-08-10 12:13:36 <ncopa> default apache config
2018-08-10 12:13:42 <ncopa> it will allow access by default
2018-08-10 12:14:11 <ncopa> i'd like to make it only accept localhost by default
2018-08-10 12:14:20 <ncopa> similar to this: https://src.fedoraproject.org/cgit/rpms/phpMyAdmin.git/tree/phpMyAdmin.htaccess
2018-08-10 12:16:01 <ncopa> ok, lets do like this: we backport https://github.com/alpinelinux/aports/pull/4981, afterwards we tightne up the default apache config
2018-08-10 12:17:30 <andypost> ncopa, good point, will add limit to localhost & then it needs to update post_install message
2018-08-10 12:18:16 <ncopa> andypost: it looks like it has been open for long time. we cannot limit to localhost in the backport
2018-08-10 12:18:24 <ncopa> as it will break things for users
2018-08-10 12:18:43 <ncopa> so i think we can merge your fix as is
2018-08-10 12:18:55 <andypost> hm, yep it will break it for existing users
2018-08-10 12:18:55 <ncopa> and then do the localhost config thing as separate issue
2018-08-10 12:19:03 <ncopa> and do that only for edge
2018-08-10 12:19:50 <andypost> ncopa, so Require ip 127.0.0.1 & ::1 ?
2018-08-10 12:20:17 <andypost> I will check second rule will work when no ipv6
2018-08-10 12:22:14 <ncopa> cant we just copy https://src.fedoraproject.org/cgit/rpms/phpMyAdmin.git/tree/phpMyAdmin.htaccess ?
2018-08-10 12:22:23 <ncopa> we can probably remove the apache 2.2 stuff
2018-08-10 12:22:30 <ncopa> and fix the path names
2018-08-10 12:24:02 <andypost> ncopa, not sure - line 47 and later about apache22
2018-08-10 12:24:23 <andypost> will check soon
2018-08-10 12:25:04 <ncopa> line 47 and later seem to restrict all access to subdirectories
2018-08-10 12:25:07 <ncopa> i think that is good
2018-08-10 12:25:27 <andypost> yep, but it is in apache2.2 format
2018-08-10 12:25:40 <ncopa> it does not work on apache 2.4?
2018-08-10 12:25:59 <_ikke_> The format changed in between
2018-08-10 12:26:03 <ncopa> ok
2018-08-10 12:26:29 <ncopa> i suppose we should ship it as the new format then
2018-08-10 12:27:15 <andypost> yep, guide is https://httpd.apache.org/docs/2.4/upgrading.html
2018-08-10 12:27:46 <andypost> that;s why package was broken with current apache
2018-08-10 12:30:24 <ncopa> ok
2018-08-10 12:40:13 <andypost> ncopa, the list of dirs a bit different now according guide https://docs.phpmyadmin.net/en/latest/setup.html#securing-your-phpmyadmin-installation
2018-08-10 12:40:50 <ncopa> ok
2018-08-10 13:34:09 <andypost> ncopa, updated https://github.com/alpinelinux/aports/pull/4981/files#diff-4dec9d775e8a554711fb05330d8a609bR3
2018-08-10 13:34:20 <andypost> this dirs at least must be hidden
2018-08-10 13:35:20 <ncopa> ok
2018-08-10 13:35:23 <ncopa> looks good to me
2018-08-10 13:35:28 <ncopa> have you tested that it works?
2018-08-10 13:57:27 <andypost> ncopa, yep, just finished tests
2018-08-10 14:14:29 <andypost> what to do with package when checksum is wrong? http://build.alpinelinux.org/buildlogs/build-edge-s390x/testing/cpd/cpd-0.5.1-r0.log
2018-08-10 14:14:45 <andypost> looks its release was reuploaded
2018-08-10 14:16:51 <_ikke_> andypost: where is the checksum worng?
2018-08-10 14:17:08 <_ikke_> Probably update the APKBUILD with the new sum
2018-08-10 14:17:51 <andypost> _ikke_, but it means that tarbal with this version changed over year
2018-08-10 14:36:43 <andypost> _ikke_, does it makes sense to increment pkgrel for that? https://github.com/alpinelinux/aports/pull/4982
2018-08-10 14:45:54 <_ikke_> Should people who already have the current release get the new release?
2018-08-10 14:46:48 <_ikke_> Or are people who already have the current version installed, ok
2018-08-10 14:46:48 <andypost> not sure, because version is the same but archive somehow changed
2018-08-10 14:47:48 <andypost> and because we have no old archive (only checksum) it is not clear and could be dangerous
2018-08-11 11:14:04 <andypost> fcolista, why python3-dev needs py3-setuptools for aarch64? looks the same failure http://build.alpinelinux.org/buildlogs/build-edge-aarch64/testing/py3-pysocks/py3-pysocks-1.6.8-r0.log
2018-08-11 11:14:53 <fcolista> andypost, the "setup" function used is in setuptools package
2018-08-11 11:16:52 <andypost> fcolista, but I just used to test py3-pysocks build on scaleway arm8 and it passed fine
2018-08-11 11:17:19 <andypost> maybe something wrong with builder?
2018-08-11 11:17:40 <fcolista> it happens only with aarch64 actually
2018-08-11 11:17:48 <fcolista> you might be right
2018-08-11 11:18:50 <fcolista> <afk>
2018-08-11 11:50:29 <andypost> yep, aarch64 is somehow broken http://build.alpinelinux.org/buildlogs/build-edge-aarch64/testing/py3-pygithub/py3-pygithub-1.40-r0.log
2018-08-12 12:53:20 <Systemnick> Hi all! Note me pls if I can change specific for my needs aports file on-the-fly before (or sth instead of) issuing apk add command
2018-08-12 12:54:16 <Systemnick> So, I need to patch APKBUILD on-the-fly to build specific module from the package
2018-08-12 12:59:28 <_ikke_> Systemnick: You have to: 1. download the APKBUILD and all relevant files. 2 build the package, 3 install the package from your local repository
2018-08-12 13:32:08 <Systemnick> @_ikke_: Have I using newapkbuild for that? Or is it any easy way?
2018-08-12 13:32:38 <_ikke_> Systemnick: no, newapkbuild is only when you want to package something new
2018-08-12 13:32:49 <_ikke_> Systemnick: easiest way is to clone aports
2018-08-12 13:37:19 <Systemnick> @_ikke_: I've clone it, but I have no idea what to do next to build one package
2018-08-12 13:40:12 <Systemnick> @_ikke_: I've tried to add that folder to /etc/apk/repositories and have no results because it's not a APKINDEX
2018-08-12 13:43:22 <Systemnick> @_ikke_: Can you send me a link to docs if you have one?
2018-08-12 13:44:18 <_ikke_> Systemnick: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account
2018-08-12 13:44:46 <_ikke_> You don't need to actually create the APKBUILD, but you can build the existing one
2018-08-12 13:45:09 <_ikke_> https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Build_the_package
2018-08-12 13:45:19 <_ikke_> (you can skip the checksum command if you haven't changed anything
2018-08-12 13:45:22 <_ikke_> )
2018-08-12 13:57:14 <Systemnick> @_ikke_: Heh, so easy! Thank you!
2018-08-12 13:59:01 <_ikke_> :-)
2018-08-12 14:08:37 <liwakura> clandmeter: do you have a moment to talk about mkinitfs?
2018-08-12 14:09:11 <liwakura> i wrote some docker stuff that compiles a mkinitfs checkout into an initramfs and runs it in qemu
2018-08-12 14:09:42 <liwakura> is it generally necessary to do the networking after nlplug-findfs?
2018-08-12 18:12:16 <clandmeter> liwakura: isnt nlplug-findfs purely to find filesystems?
2018-08-12 18:19:47 <liwakura> nlplug-findfs invokes mkdev
2018-08-12 18:19:50 <liwakura> *mdev
2018-08-12 18:20:01 <liwakura> and mdev then does the hotplug magic
2018-08-12 18:20:03 <liwakura> somehow
2018-08-12 18:20:11 <clandmeter> isnt mdev run before?
2018-08-12 18:21:09 <liwakura> not that im aware of
2018-08-12 18:21:32 <liwakura> when you do >grep mdev initramfs-init.in<, you see its mentioned three times
2018-08-12 18:21:39 <liwakura> two as -p option for nlplug-findfs
2018-08-12 18:21:43 <clandmeter> right
2018-08-12 18:21:47 <clandmeter> i just checked
2018-08-12 18:22:14 <liwakura> i updated my PR
2018-08-12 18:22:39 <liwakura> also, i created some docker thing to test mkinitfs things... otherwise im having to do stuff blindly, which isn't too comfy
2018-08-12 18:23:15 <liwakura> like, mkinitfs-revision XYZ -> compile -> build initramfs -> boot in qemu
2018-08-12 18:32:13 <clandmeter> liwakura: from my understand network needs to be done after it loads the modules.
2018-08-12 18:32:49 <clandmeter> and for usb devices you need to add a delay or the network device will not be available yet.
2018-08-12 18:32:57 <clandmeter> for instance for rpi
2018-08-12 18:33:41 <clandmeter> i didnt create nlplug-findfs, so im not sure how it exactly works.
2018-08-12 18:33:47 <liwakura> hah
2018-08-12 18:34:18 <liwakura> mh, yeah, seems like that
2018-08-12 18:34:30 <clandmeter> you can ask ncopa when he is around.
2018-08-12 18:34:43 <clandmeter> he probably know more details about that the hw init part.
2018-08-12 18:34:53 <clandmeter> -that
2018-08-12 18:35:31 <clandmeter> i remember you can run it in debug mode. maybe you can get a trace of the boot and see what actually happens.
2018-08-12 18:35:44 <liwakura> i wrote initramfs stuff for other distros previoiusly
2018-08-12 18:35:51 <liwakura> yeah, im using the debug mode for testing
2018-08-12 18:35:58 <clandmeter> i know nlplug is very verbose
2018-08-12 18:37:02 <liwakura> uhm, using qemu with serial -> stdio and then redirecting into an file is sort of useful
2018-08-12 18:37:13 <clandmeter> yep
2018-08-12 18:38:13 <clandmeter> i think ive done the same with vbox before
2018-08-12 18:39:06 <liwakura> after i found out how qemu is invoked, i stopped using everything else
2018-08-12 18:39:22 <liwakura> qemu with -kernel and -initrd + alpine diskless is an godsend thing
2018-08-12 18:39:26 <clandmeter> qemu is nice
2018-08-12 18:39:35 <clandmeter> but on windows not so
2018-08-12 18:40:01 <clandmeter> i guess you saw i updated netboot?
2018-08-12 18:40:24 <liwakura> yeah
2018-08-12 18:40:43 <liwakura> i saw that issue is fixed, but i was unable to move it from resolved -> closed
2018-08-12 18:40:57 <clandmeter> you cant?
2018-08-12 18:41:05 <clandmeter> i can look into that.
2018-08-12 18:41:52 <liwakura> it might be because im the reporter, but thats weird because "Reporter" is my current role on bugs.a.o
2018-08-12 18:42:14 <clandmeter> _ikke_: you are  the redmine expert here :) ^
2018-08-12 18:42:32 <liwakura> ticket systems with enforced workflows and roles are rather cumbersome, sometimes
2018-08-12 18:42:53 <clandmeter> redmine was nice when we implemented.
2018-08-12 18:43:00 <clandmeter> it hasnt been moving forwards since.
2018-08-12 18:43:19 <clandmeter> we are looking into alternatives.
2018-08-12 18:43:51 <liwakura> mh
2018-08-12 18:44:14 <_ikke_> Yeah, I noticed some 'permissions' were lacking for some roles
2018-08-12 18:44:19 <clandmeter> but thats a whole diff subject :)
2018-08-12 18:46:51 <_ikke_> liwakura: mind sharing your username on redmine?
2018-08-12 18:47:02 <liwakura> nero / Ain Nero
2018-08-12 18:47:19 <liwakura> i'd name myself nero on freenode if that nick wasn't already taken..
2018-08-12 18:47:42 <_ikke_> And what is the particular ticket?
2018-08-12 18:48:10 <liwakura> #9198
2018-08-12 18:48:30 <liwakura> its the only ticket where im unable to modify the properties
2018-08-12 18:49:04 <liwakura> aah wait, i think its because its in the Infra Project...
2018-08-12 18:49:19 <_ikke_> right
2018-08-12 18:49:25 <_ikke_> You are not a reported there
2018-08-12 18:49:45 <clandmeter> :)
2018-08-12 18:49:49 <clandmeter> yep
2018-08-12 18:49:51 <clandmeter> thats it
2018-08-12 18:50:14 <_ikke_> So we can either make him a reporter, or allow non-members to do more things
2018-08-12 18:50:51 <clandmeter> i think for now just change to reporter
2018-08-12 18:51:03 <clandmeter> do we want to allow everbody to close tickets?
2018-08-12 18:51:21 <_ikke_> done
2018-08-12 18:51:58 <_ikke_> If it's their own tickets or they are assigned the ticket?
2018-08-12 18:52:14 <_ikke_> (in redmine you can distinguish between that and other tickets)
2018-08-12 18:52:28 <clandmeter> ah ok
2018-08-12 18:52:34 <clandmeter> yes we should allow that.
2018-08-12 18:56:01 <liwakura> _ikke_: i can close random tickets currently
2018-08-12 18:57:36 <liwakura> also did with tickets that were put on resolved and later forgotten
2018-08-13 07:26:26 <AlecTaylor> hi
2018-08-13 15:31:40 <BitL0G1c1> https://github.com/alpinelinux/aports/pull/4994 - fixes 2 x CVE's in mbedtls
2018-08-13 15:33:05 <BitL0G1c1> would be helpfull to push https://github.com/alpinelinux/aports/pull/4877 afterwards as it depends on mbedtls
2018-08-13 16:07:35 <ncopa> BitL0G1c1:  it would be nice to have sec upgrade in separate commit so it can be fast-tracked and easily backported
2018-08-13 16:07:59 <ncopa> what happens now is that i need comment the environment thing and cannot push the secfix
2018-08-13 16:09:13 <BitL0G1c1> so break the update & profile.d into 2 commits ?
2018-08-13 16:12:34 <ncopa> that would be good yes
2018-08-13 16:12:41 <ncopa> also, is the update ABI compatible?
2018-08-13 16:21:02 <ncopa> hum.... boost 1.67 is broken on non-x86*
2018-08-13 16:21:20 <ncopa> and boost python seems to be broken
2018-08-13 16:21:30 <ncopa> i think its  some missing dependency
2018-08-13 16:27:49 <BitL0G1c1> ncopa - utils moved into /usr/bin
2018-08-13 16:30:10 <ncopa> mksully22: can you help me figure out why 2 tests fails on ppc64le? http://build.alpinelinux.org/buildlogs/build-edge-ppc64le/main/boost/boost-1.67.0-r1.log
2018-08-13 16:36:22 <ncopa> it seems to happen with boost 1.68 too
2018-08-13 19:02:58 <ncopa> would be nice if I could get some help resolving some those: https://bugs.alpinelinux.org/projects/alpine/issues?fixed_version_id=126&set_filter=1&status_id=o
2018-08-13 19:03:33 <_ikke_> Let's see if there is some low-hanging fruit for me
2018-08-13 19:29:11 <mksully22> ncopa: I will look at failing debugger debugger-mi tests for boost on ppc64le
2018-08-14 06:02:43 <nindustries> Hi, any idea when postfix 3.3.1 is getting merged? Running in foreground does not work in 3.3.0
2018-08-14 06:06:03 <nindustries> I see no postfix 3.3.1 in testing
2018-08-14 08:59:04 <nindustries> Hi, any idea when postfix 3.3.1 is getting merged? Running in foreground does not work in 3.3.0
2018-08-14 09:09:20 <PureTryOut[m]> hey could anyone of you review my PR to update gnome-shell? https://github.com/alpinelinux/aports/pull/4952
2018-08-14 12:23:17 <ScrumpyJack> Are we privy to the current vulnerabilities in the wild that have been embargoed? Is AL patched?
2018-08-14 12:24:36 <ScrumpyJack> embargo ends 6pm UTC today
2018-08-14 12:27:06 <_ikke_> Embargo for what?
2018-08-14 12:27:18 <_ikke_> Or is that still embargoed? :P
2018-08-14 12:28:56 <ScrumpyJack> Knowledge of a Linux vulnerability has been embargoed and distributed to some distributions that have signed agreements.
2018-08-14 12:29:31 <ScrumpyJack> I guess AL isn't privy to that.
2018-08-14 12:29:36 <_ikke_> I don't think (but don't know for sure) that AL is included
2018-08-14 12:29:55 <ScrumpyJack> unless we are and none can mention it
2018-08-14 12:29:56 <liwakura> ScrumpyJack: any sources on that?
2018-08-14 12:30:08 <liwakura> dont know of anything
2018-08-14 12:30:15 <ScrumpyJack> no sources. sorry.
2018-08-14 12:30:25 <_ikke_> Well, if any one of us knows, it would be ncopa
2018-08-14 12:30:32 <liwakura> then it better be an privilege escalation or RCE
2018-08-14 12:32:10 <_ikke_> Well, I hope it's none of those, but probably will be
2018-08-14 12:32:34 <liwakura> im quite annoyed by the security theatre for minuscule bugs
2018-08-14 12:32:44 <liwakura> and security "researchers" overhyping it to get new fundings
2018-08-14 12:33:57 <_ikke_> nod
2018-08-14 12:34:26 <liwakura> some some reason, data corruption or just plain misbehavior is not an security incident, even if it can easily kill an company
2018-08-14 12:34:46 <liwakura> because there is no evil attacker involved *shrugs*
2018-08-14 12:36:17 <_ikke_> Companies like Red Hat seem to only backport security issues, which causes them to miss certain bug fixes which are later discovered as security vulns
2018-08-14 12:38:00 <liwakura> re my previous point, "security against attackers" vs "security against company disruption"
2018-08-14 12:39:11 <liwakura> anyways, most of my statements regarding anything related to security are likely just ramblings
2018-08-14 14:35:15 <Shiz> ScrumpyJack: we don't have access to the -distros list, no
2018-08-14 14:36:06 <ScrumpyJack> @Shiz something AL ought to be part of?
2018-08-14 14:37:26 <_ikke_> ScrumpyJack: Is it just a matter of asking nicely if we can get access?
2018-08-14 14:57:23 <ScrumpyJack> @_ikke_ yes https://oss-security.openwall.org/wiki/mailing-lists/distros#membership-criteria
2018-08-14 14:59:39 <clandmeter> yes, but it need to be handled internally which I/we simply didnt have time yet for.
2018-08-14 15:01:29 <ScrumpyJack> clandmeter: you mean no time to find someone internally willing to engage -distros ?
2018-08-14 15:01:46 <ScrumpyJack> we could start a ML discussion
2018-08-14 15:02:14 <_ikke_> 5 and 9 might be an issue for us?
2018-08-14 15:02:19 <clandmeter> its a bit more complicated. Shiz worked on it before.
2018-08-14 15:03:01 <clandmeter> only a few core members can have access, and it needs to be secured.
2018-08-14 15:03:32 <_ikke_> that makes sense
2018-08-14 15:03:44 <Shiz> clandmeter: for -distros we can't use the pgp remailer anyway
2018-08-14 15:03:45 <clandmeter> the implementation shiz worked on had some limitations.
2018-08-14 15:03:51 <Shiz> "Although membership is for distros rather than individuals, for each distro we're only subscribing individuals with their PGP keys (messages are re-encrypted to individual recipients), not exploders.
2018-08-14 15:03:53 <Shiz> "
2018-08-14 15:04:47 <clandmeter> so you can subscribe with them already without the need of a group address?
2018-08-14 15:05:08 <Shiz> yes
2018-08-14 15:05:16 <clandmeter> ok, thats new for me.
2018-08-14 15:05:21 <clandmeter> that has changed?
2018-08-14 15:05:42 <clandmeter> cause you also worked on that remailer and i did some work on another solution.
2018-08-14 15:34:53 <ncopa> Shiz: that was new to me too /o\
2018-08-14 15:35:05 <ncopa> we can just subscribe i guess
2018-08-14 15:35:08 <ncopa> and see what happens
2018-08-14 15:35:17 <Shiz> we need to find tasks to do
2018-08-14 15:35:22 <Shiz> and abide by 5)
2018-08-14 16:31:49 <_ikke_> ScrumpyJack: Do you know where the vulnerability will be disclosed? oss-security?
2018-08-14 16:32:27 <ollieparanoid[m]> hi folks! xfce4 (xfwm) is kind of broken with xorg-1.20.0 from edge. latest relese 1.20.1 fixes it, could someone please merge this? https://github.com/alpinelinux/aports/pull/4983
2018-08-14 16:34:46 <ollieparanoid[m]> test
2018-08-14 16:34:55 <_ikke_> Ok, it works
2018-08-14 16:35:39 <ollieparanoid[m]> thanks :)
2018-08-14 17:23:35 <ScrumpyJack> _ikke_: yeah. it's mainly xen as you can see
2018-08-14 17:59:33 <_ikke_> ScrumpyJack: see where?
2018-08-14 17:59:40 <_ikke_> http://www.openwall.com/lists/oss-security/2018/08/
2018-08-14 17:59:58 <_ikke_> It's just now 6pm utc, right?
2018-08-14 18:00:18 <_ikke_> Or did they disclose it 45 minutes before?>
2018-08-14 18:00:35 <liwakura> is this the exploit that you meant, ScrumpyJack?
2018-08-15 16:31:14 <andypost> anyone can fix out-of-space http://build.alpinelinux.org/buildlogs/build-edge-armhf/main/lucene++/lucene++-3.0.7-r7.log
2018-08-15 18:40:18 <auto716> Not sure if this is the venue or not but on Alpine 3.8 running apk add chromium chromium-chromedriver seems to leave me with chromedriver 2.38 which looks to have a maximize bug I am trying to work around.  Am I missing something here?
2018-08-15 19:29:02 <auto716> Not sure if this is the venue or not but on Alpine 3.8 running apk add chromium chromium-chromedriver seems to leave me with chromedriver 2.38 which looks to have a maximize bug I am trying to work around.  Am I missing something here?
2018-08-15 19:33:18 <liwakura> auto716: how do you estimate the chances that someone will have the exact same bug like you have?
2018-08-15 19:33:45 <andypost> btw looks http://chromedriver.chromium.org/ should be packaged separate
2018-08-15 19:36:40 <auto716> liwakura:  I assume anyone installing the apk chromium-chromedriver would see the chromedriver version
2018-08-15 19:37:54 <liwakura> ChromeDriver 2.38 (f91d32489882be7df38da3422a19713bfd113fa5)
2018-08-15 19:38:03 <auto716> Right
2018-08-15 19:38:44 <auto716> That has a maximize window bug I am just trying to upgrade o 2.41
2018-08-15 19:39:10 <liwakura> how to make it start a window?
2018-08-15 19:39:14 <andypost> auto716, you need to update chromium
2018-08-15 19:39:48 <auto716> Running that apk --upgrade chromium dosent seem to give me 2.41 either
2018-08-15 19:43:44 <liwakura> doesn't look like 2.41 is in the repo
2018-08-15 19:43:56 <liwakura> the apk version is the 68-ish one of chromium
2018-08-15 19:45:58 <auto716> I see the same, the repo shows 68.0.3440.75 but the chromedriver 2.38 shows support only to 67 and I am trying to figure out how to upgrade to 2.41 without the apk install
2018-08-15 19:47:42 <auto716> If the package installs chromium 68 and bundles with it the chromedriver 2.38 then I assume that is an issue but my build skills from source are lacking so I am trying to fight through this
2018-08-16 09:10:23 <liwakura> ???
2018-08-16 09:10:33 <liwakura> like, it shows the delays on mirrors.alpinelinux.org
2018-08-16 09:10:42 <_ikke_> right, but active alerts
2018-08-16 09:12:16 <clandmeter> and ones that are correct :)
2018-08-16 09:12:33 <clandmeter> seems my delay calculations are broken.
2018-08-16 13:31:49 <ncopa> seems like ceph 13 requires gcc7+
2018-08-16 13:32:04 <ncopa> we need upgrade gcc sooner than later :-/
2018-08-16 14:27:45 <danieli> ncopa: how much of a pain is it going to be to upgrade python to 3.7 stable?
2018-08-16 15:11:17 <ncopa> dunno
2018-08-16 15:11:49 <ncopa> i need fix ceph first due to boost upgrade
2018-08-16 15:11:53 <ncopa> but it depends on gcc7
2018-08-16 15:11:58 <ncopa> so i need fix gcc7 first
2018-08-16 15:12:10 <ncopa> after that can i fix ceph
2018-08-16 15:12:21 <ncopa> and after that i'd like to upgrade boost to 1.68
2018-08-16 15:12:33 <ncopa> and then maybe i can start look at python 3.7
2018-08-16 15:12:56 <ncopa> need to fix some security vulns before all the above though
2018-08-16 15:13:25 <ncopa> and i'd like to get alpine 3.8.1 out
2018-08-16 15:59:10 <danieli> ncopa: I was just wondering if you knew how much effort it would be, I understand that you have other priorities, of course
2018-08-16 16:03:18 <_ikke_> at least all libraries need to be bumped for 3.7
2018-08-16 20:19:55 <dorvan> hi all, we are evaluating to use alpine linux, but we need a updated mono package, it's possible make a request?
2018-08-16 20:23:08 <_ikke_> dorvan: It sure is possible to ask for an updated version, but it depends on when people have time for it and how difficult it is to update. And then it would probanly only be part of the next major release
2018-08-16 20:24:44 <_ikke_> dorvan: https://bugs.alpinelinux.org
2018-08-16 20:26:23 <dorvan> ah ok
2018-08-17 07:05:07 <liwakura> ping ncopa: i made some PR's to mkinitfs
2018-08-17 08:04:49 <ncopa> liwakura: if you did the PRs as nero, then they look really good.
2018-08-17 08:05:12 <AinNero> i did change my nick to make the association more obvious
2018-08-17 08:05:32 <AinNero> i'd take "nero" if it wasn't already taken
2018-08-17 08:06:07 <ncopa> I have been busy with the boost upgrade and I have problems with keeping up
2018-08-17 08:06:10 <ncopa> Sorry
2018-08-17 08:06:31 <AinNero> im offering to take over bug fixing maintenance for mkinitfs
2018-08-17 08:06:50 <ncopa> That would be great
2018-08-17 08:07:42 <ncopa> I'm on a train now and will most likely be busy today
2018-08-17 08:07:50 <AinNero> its not urgent anyways
2018-08-17 08:08:25 <ncopa> Ping me next week if you don't hear from me today
2018-08-17 08:08:32 <AinNero> i was a bit worried because there was no responses to the PR's (except from one of my pals putting likes under it to troll me)
2018-08-17 08:08:35 <AinNero> alright
2018-08-17 12:36:27 <andypost> Can someone get this in? https://github.com/alpinelinux/aports/pull/4798
2018-08-17 14:08:41 <ncopa> fcolista: do you think you can help me with upgrade tomahawk or rebuild against boost-1.67
2018-08-17 14:08:48 <ncopa> it failed on first try for me
2018-08-17 14:09:00 <ncopa> and i notice that gentoo ships a 0.9.0 prerelease
2018-08-17 14:09:08 <ncopa> im fine with either
2018-08-17 16:11:57 <andypost> ncopa, this change looks better & more backward compatible https://github.com/alpinelinux/aports/pull/5021#pullrequestreview-147289773
2018-08-17 16:12:44 <andypost> I think both commits needs backport to 3.8
2018-08-17 23:18:54 <sleepycat> I'm making my first Alpine package and while my compilation completes my package build fails after abuild trys to make a -dev package. `arangodb-dev*: Running split function dev...` & then `can't cd to /home/mike/arangodb/pkg/arangodb-dev`
2018-08-17 23:19:26 <sleepycat> Can anyone shed a little light on what's going on?
2018-08-17 23:21:33 <sleepycat> My repo for this project is here: https://gitlab.com/sleepycat/arangodb-alpine
2018-08-17 23:23:18 <sleepycat> Comments, critiques and constructive criticism welcome... also pull requests. 😃
2018-08-18 09:43:47 <fcolista> ncopa, I've tried to work on tomahawk
2018-08-18 09:44:04 <fcolista> https://dpaste.de/ujRm/raw
2018-08-18 09:44:34 <fcolista> I'm now hitting this problem: https://github.com/tomahawk-player/tomahawk/issues/578
2018-08-18 09:44:41 <fcolista> which appears to be related to libboost
2018-08-18 09:45:04 <fcolista> it's more tha one year old
2018-08-18 09:46:48 <fcolista> I don't have cpp skills to understand what is this problem related to...
2018-08-18 16:03:39 <_ikke_> ^ is not really a request for alpine is it
2018-08-18 16:04:39 <skarnet> why wouldn't it be
2018-08-18 16:05:02 <_ikke_> Do we maintain musl-utils>
2018-08-18 16:05:05 <_ikke_> ?
2018-08-18 16:09:31 <skarnet> yes?
2018-08-18 16:09:48 <_ikke_> Thought it came from musl
2018-08-18 16:10:09 <_ikke_> https://git.alpinelinux.org/cgit/aports/tree/main/musl/APKBUILD#n16
2018-08-18 16:10:46 <_ikke_> ah, apparently we add getent
2018-08-18 16:10:50 <skarnet> yes
2018-08-18 16:10:55 <skarnet> it's not provided by musl
2018-08-18 16:11:01 <_ikke_> right, missed that
2018-08-20 11:24:03 <mksully22> is the ppc64le edge builder down, I don't see any status being reported?
2018-08-20 12:31:39 <Fusl> any alpine dev able to take a look at this six month old PR for telegraf and either get this or #4383 merged? https://github.com/alpinelinux/aports/pull/3377
2018-08-20 13:23:44 <Shiz> yea, when i get at work
2018-08-20 14:12:33 <_ikke_> Wouldn't this match an include dir anywhere, even if it's not a header dir? https://git.alpinelinux.org/cgit/abuild/tree/abuild.in#n1686 (so a non-revelent dir might be moved)
2018-08-20 14:21:02 <AinNero> anywhere in $pkgdir, yeah...
2018-08-20 14:23:32 <_ikke_> Can that be considered a bug? (not that I'm running into it, was just reading the code)
2018-08-20 14:32:18 <Fusl> _ikke_: many find commands there are very loosely defined
2018-08-20 14:32:25 <Fusl> such as https://git.alpinelinux.org/cgit/abuild/tree/abuild.in#n569
2018-08-20 14:32:31 <Fusl> its missing a -type f
2018-08-20 14:34:29 <AinNero> but what are the chances of having an directory named config.sub in the build tree?
2018-08-20 14:35:17 <ncopa> Fusl: thanks for good work with telegraph
2018-08-20 14:35:28 <ncopa> and thanks for your patience....
2018-08-20 14:37:06 <_ikke_> Fusl: right, this was just one that I though would have the most chance of false positives
2018-08-20 14:38:50 <ncopa> i think there was some package that had /usr/something/include or similar
2018-08-20 14:39:04 <ncopa> probably /usr/lua5.x/include
2018-08-20 14:39:32 <ncopa> i dont think i have seen any /include/ yet that is not for *.h files
2018-08-20 14:41:47 <_ikke_> Alright, it's probably very rare, though it might be supprising when it does happen
2018-08-20 15:12:41 <andypost> ncopa, where to start discussion about php5 removal? it goes EOL in Dec
2018-08-20 15:13:14 <andypost> and after that we can rename php7 back to php-
2018-08-20 15:13:18 <ncopa> on alpine-devel mailing ilst
2018-08-20 15:13:49 <ncopa> are there any package that still require php5?
2018-08-20 15:14:17 <ncopa> would be nice to be able to get rid of php5
2018-08-20 15:16:57 <_ikke_> 4 packages seem to directly depend on php5
2018-08-20 15:18:28 <andypost> _ikke_, https://pkgs.alpinelinux.org/package/edge/community/x86_64/php5 shows 5
2018-08-20 15:22:01 <_ikke_> Oh, I accidentally choose armhf
2018-08-20 15:23:06 <ncopa> mksully22: it looks like ppc64le is up and running, its just busy
2018-08-21 07:50:15 <ncopa> rnalrd: thanks for fixing ldb tests!
2018-08-21 07:51:09 <ncopa> i think we should also remove the `unset CFLAGS`, which i think was added due to those failing tests
2018-08-21 08:54:11 <rnalrd> ncopa, sounds good
2018-08-21 08:54:34 <rnalrd> didn't see the comment
2018-08-21 13:07:37 <sleepycat> I'm building my package and it's failing at the end. I'm doing `install -o arangodb...` in the package() function... but getting: install: unknown user arangodb. Does the pre-install script get run during package creation? It looks like "no" is the answer, so how do you create folders owned by users that don't exist yet?
2018-08-21 13:09:13 <sleepycat> The redis package is doing the same thing: https://git.alpinelinux.org/cgit/aports/tree/main/redis/APKBUILD#n62
2018-08-21 13:09:46 <sleepycat> As is the postgres package: https://git.alpinelinux.org/cgit/aports/tree/main/postgresql/APKBUILD#n134
2018-08-21 13:11:54 <sleepycat> The only users I can see being created in those packages are in the pre-install. 🤔
2018-08-21 13:21:43 <ncopa> sleepycat: pkgusers and pkggroups: https://git.alpinelinux.org/cgit/aports/tree/main/redis/APKBUILD#n10
2018-08-21 13:22:20 <ncopa> https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#pkgusers
2018-08-21 13:35:06 <sleepycat> ncopa: Awesome!
2018-08-21 13:35:10 <sleepycat> Thank you!
2018-08-21 15:38:31 <sleepycat> My package has a brutally long compilation and is currently failing to build after the compilation. Is there a way to get abuild to keep the stuff that was just compiled? `-K  Keep buildtime temp dirs and files (srcdir/pkgdir/deps)` looks like the right thing, but everything gets compiled again anyway.
2018-08-21 16:16:01 <sleepycat> At the end of my compilation process I get two errors: `fatal: No names found, cannot describe anything.`
2018-08-21 16:16:37 <sleepycat> And `ERROR: APKINDEX.tar.gz: UNTRUSTED signature`
2018-08-21 16:17:42 <sleepycat> Anyone have any thoughts on what's going wrong?
2018-08-21 16:33:00 <sleepycat> the build was done with `abuild -r`.
2018-08-21 17:31:34 <skarnet> sleepycat: got your packager's key in ~/.abuild and correctly pointed to by ~/.abuild/abuild.conf ? got the pubkey available in /etc/apk/keys?
2018-08-21 17:32:58 <skarnet> also, I don't remember exactly what the problem was, but I've hit the "No names found" thing before, it's abuild taking a wrong branch and running "git describe". Look for instances of "git" in /usr/bin/abuild, there's a describe there, and something's happening
2018-08-21 17:33:09 <skarnet> making it take that branch when it shouldn't - a variable missing or something.
2018-08-22 08:38:58 <clandmeter> ncopa, thanks for fixing chromium :)
2018-08-22 09:23:25 <BitL0G1c1> anyone used the 1GB Alpine AWS ami ?
2018-08-22 10:13:32 <ncopa> i havent
2018-08-22 10:23:44 <xsteadfastx> i want to create a apk for bitlbee-mastodon but its hosted on a own cgit. i have no idea how to get a source out of this
2018-08-22 10:24:15 <xsteadfastx> there is a github mirror... but i dont know if its official
2018-08-22 10:54:00 <ncopa> we are thinking about doing monthly snapshots from edge
2018-08-22 10:54:14 <ncopa> release images every month from edge
2018-08-22 10:54:21 <ncopa> but i think we will only do minirootfs
2018-08-22 10:54:26 <ncopa> and netboot
2018-08-22 12:24:23 <clandmeter> xsteadfastx, https://github.com/kensanata/bitlbee-mastodon ?
2018-08-22 12:29:26 <xsteadfastx> clandmeter: exactly. thats what im using now. i checked his mastodon account and its the same username... so i guess its official
2018-08-22 12:31:06 <xsteadfastx> i created a PR on github for the package
2018-08-22 13:56:56 <AinNero> what to do about the console=ttyS0 thing on virt?
2018-08-22 13:57:07 <AinNero> this is not the first time someone complained about it
2018-08-22 13:57:21 <AinNero> should we tell people to remove the part from the kernel command line?
2018-08-22 14:10:46 <AinNero> or create an option for optional_console=, that dont have affect if the tty is not found
2018-08-22 14:10:51 <AinNero> not sure whats the right thing to do here
2018-08-22 14:21:30 <AinNero> strfry: ping
2018-08-22 14:26:23 <ncopa> AinNero: i wanted to ask you the same
2018-08-22 14:27:15 <awilfox> we solved it in Adélie, but I don't know if it is something Alpine can use tbh
2018-08-22 14:27:25 <awilfox> I say "we"
2018-08-22 14:27:30 <awilfox> skarnet is the one who did all the work
2018-08-22 14:28:29 <AinNero> awilfox: how?
2018-08-22 14:28:30 <ncopa> we ship alpine-virt with console=ttyS0 iirc, since you often will want access the booloader via serial console
2018-08-22 14:28:31 <awilfox> sysvinit doesn't do getty at all now, it runs a s6 supervision tree for gettys, and then a one-shot service creates getty services for all the ttys that exist on the system
2018-08-22 14:29:28 <awilfox> dynamic gettys, without systemd \o/
2018-08-22 14:29:39 <AinNero> wat
2018-08-22 14:29:40 <AinNero> \o/
2018-08-22 14:29:47 <AinNero> what a random feature.
2018-08-22 14:29:56 <AinNero> ncopa: btw, my patch didn't solve the issue
2018-08-22 14:30:21 <AinNero> the kernel still has the non-existing tty's in the sysfs and waits until they hotplug
2018-08-22 14:30:35 <ncopa> didn solve #9324?
2018-08-22 14:30:40 <ncopa> or other issue
2018-08-22 14:31:14 <AinNero> does not.
2018-08-22 14:31:28 <AinNero> the "correct" way would be to remove the console= parameter from the kernel command line
2018-08-22 14:32:18 <AinNero> but then alpine would become uninstallable on machines where ttyS0 is the only way to interact with the system
2018-08-22 14:33:41 <AinNero> i'll spend some brain cycles on it later
2018-08-22 14:34:41 <Shiz> awilfox: when's adelie gonna switch to s6 completely ;)
2018-08-22 14:34:52 <awilfox> Shiz: hopefully by beta3
2018-08-22 14:34:57 <awilfox> Shiz: but it may be post-1.0
2018-08-22 14:35:03 <awilfox> depends on what we can get done this month
2018-08-22 14:39:07 <ncopa> AinNero: and it would not help to test if tty exists?
2018-08-22 14:39:10 <ncopa> http://tpaste.us/kW0w
2018-08-22 14:40:46 <AinNero> that might by racey with hotpluging (think of ttyUSB0)
2018-08-22 14:42:18 <AinNero> i'll test it later
2018-08-22 14:42:23 <ncopa> which is sort of the point
2018-08-22 14:42:47 <ncopa> if serial does not exist then dont add it to /etc/inittab
2018-08-22 14:43:07 <ncopa> if you need it, then plug it in and reboot
2018-08-22 14:43:53 <AinNero> in the first moments of the kernel, none of the ttys exist, stuff gets hotplugged at some point
2018-08-22 14:44:17 <AinNero> and some hardware might only get ready half a second after we made that check
2018-08-22 14:44:20 <ncopa> nplug-findfs does it
2018-08-22 14:44:27 <ncopa> right
2018-08-22 14:44:35 <AinNero> i mean not creating the device node, i mean the kernel detecting it
2018-08-22 14:45:05 <AinNero> usb hubs are a good canidate for it, but ttyS0 is uncritical since its per convention always at the same place and interrupt
2018-08-22 14:45:22 <AinNero> so yeah, we can do it, but it might randomly break in exotic cases
2018-08-22 14:46:06 <skarnet> the joys of working around sysvinit infelicities (that bb init shares, in this case)
2018-08-22 14:47:13 <ncopa> indeed
2018-08-22 14:47:36 <AinNero> ncopa: oh, btw, you need to check against /dev/$tty
2018-08-22 14:47:40 <skarnet> honestly dynamic gettys is the real solution
2018-08-22 14:48:07 <ncopa> replacing init system
2018-08-22 14:48:14 <ncopa> solves alot issues
2018-08-22 14:48:20 <skarnet> yeah. /etc/inittab isn't flexible enough for this
2018-08-22 14:48:33 <ncopa> but creates a bunch of new issues
2018-08-22 14:48:37 <skarnet> just get your supervision tree from somewhere else (hint, hint)
2018-08-22 14:49:04 <skarnet> Adélie hasn't replaced init yet
2018-08-22 14:49:14 <skarnet> still running on sysvinit + openrc
2018-08-22 14:49:22 <AinNero> lets wait how it turns out for Adélie
2018-08-22 14:49:32 <ncopa> but use s6 for supervision?
2018-08-22 14:49:43 <skarnet> yeah, init supervises a s6 tree
2018-08-22 14:50:07 <AinNero> skarnet: dynamic getty's also dont solve the full problem
2018-08-22 14:50:13 <skarnet> so in "openrc sysinit" you can add early services and they will be spawned by the time "openrc default" runs
2018-08-22 14:50:19 <ncopa> AinNero: since it happens after install, i'd think he can just edit the /etc/inittab
2018-08-22 14:50:26 <AinNero> definitely
2018-08-22 14:50:32 <ncopa> and edit the boot config
2018-08-22 14:50:43 <AinNero> skarnet: what to do with wrong console= parameters in the kernel command line?
2018-08-22 14:51:11 <AinNero> the kernel has the same issues with invalid ttys, it just fails better and quiter
2018-08-22 14:51:17 <AinNero> *quieter
2018-08-22 14:51:23 <skarnet> don't hardcode console? >.>
2018-08-22 14:51:27 <skarnet> I mean
2018-08-22 14:51:43 <skarnet> it's such a basic setting you want to hardcode it... except it's wrong on some systems
2018-08-22 14:51:50 <skarnet> so you need to make it configurable somehow
2018-08-22 14:51:55 <ncopa> it is
2018-08-22 14:52:06 <ncopa> boot options is configurable
2018-08-22 14:52:12 <skarnet> then what's the problem
2018-08-22 14:52:34 <ncopa> thats what im saying, user should just change the config
2018-08-22 14:52:40 <AinNero> skarnet: users cant/wont configure it
2018-08-22 14:52:53 <ncopa> they want it to work out of the box
2018-08-22 14:52:56 <AinNero> we could add menu.c32 to the iso to make command line editing easier
2018-08-22 14:53:17 <skarnet> ¯\_(ツ)_/¯ I can fix tech that doesn't work, I can't fix dumb
2018-08-22 14:53:31 <skarnet> and yeah, a bootloader menu would help
2018-08-22 14:54:04 <skarnet> I've learned that the answer to "users won't do it" is often "give them a nicer interface"
2018-08-22 14:55:14 <clandmeter> use alpine netboot and you will get a nice boot menu :)
2018-08-22 14:55:37 <skarnet> with various choices for console? :)
2018-08-22 14:56:01 <clandmeter> exactly :)
2018-08-22 14:59:26 <AinNero> ncopa: updated patch http://0x0.st/syf_.diff
2018-08-22 15:15:08 <ncopa> or we could wait 3 seconds for it to show up
2018-08-22 15:15:11 <ncopa> http://tpaste.us/O4DX
2018-08-22 15:15:19 <ncopa> but i find that ugly
2018-08-22 15:16:02 <AinNero> yeah, ugly, just do the quick check
2018-08-22 15:16:13 <AinNero> it will probably close more issues than it will cause to be created
2018-08-22 15:16:43 <ncopa> i suppose we could also do it from setup-alpine
2018-08-22 15:17:06 <ncopa> check if current, logged in tty is
2018-08-22 15:18:00 <ncopa> and if its not ttyS0, then remove it from boot options and from /etc/inittab
2018-08-22 15:18:58 <ncopa> it goes with the "give them a nicer interface"
2018-08-22 15:19:36 <AinNero> the bootoptions aren't transferred to the diskless rootfs, or are they?
2018-08-22 15:19:48 <AinNero> cant see code doing it, so the new update-extlinux.conf would be fresh then
2018-08-22 15:21:29 <AinNero> ah, setup-disk grabs them from /proc/cmdline..
2018-08-22 15:22:10 <AinNero> i'd go with the "strip it during setup-alpine" method
2018-08-22 15:44:43 <ncopa> yeah, i think that makes sense
2018-08-22 15:44:48 <ncopa> i'll look at it tomorrow
2018-08-23 04:27:45 <_ikke_> heh ^
2018-08-23 04:27:52 <_ikke_> first ticket about it
2018-08-24 09:18:00 <AinNero> ^ can we offer that guy a full refund?
2018-08-24 09:18:40 <clandmeter> :)
2018-08-24 09:18:50 <clandmeter> i like his openvpn issue too
2018-08-24 09:19:38 <AinNero> or #9324, where he demands to break installing on my headless qemu setups
2018-08-24 09:21:17 <danieli> do we close 9324 as a wontfix then?
2018-08-24 09:21:28 <danieli> they are indeed interesting issues though
2018-08-24 09:21:49 <AinNero> danieli: there will be something happening, but definitely not what he is demanding
2018-08-24 09:22:09 <danieli> well, it is intentional behavior, and the slight noise in dmesg can easily be removed
2018-08-24 09:22:24 <AinNero> idea was to make setup-alpine strip console= boot options to invalid terminals
2018-08-24 09:22:32 <AinNero> but yeah, can be closed as invalid
2018-08-24 09:24:22 <danieli> that would be a good solution
2018-08-24 09:37:27 <mstanich_> danieli: wontfix is good idea
2018-08-24 09:37:53 <mstanich_> ttyS0 is useful on VM installs
2018-08-24 09:38:05 <AinNero> mstanich_: its an hard requirement for headless
2018-08-24 09:38:54 <AinNero> the qemu on alpine still has vnc, but no gtk or sdl frontend
2018-08-24 09:38:58 <mstanich_> personally I would like it on the bare metal also
2018-08-24 09:39:06 <AinNero> i agree with that
2018-08-24 09:39:44 <mstanich_> AinNero: qemu on alpine have gtk frontend
2018-08-24 09:40:17 <mstanich_> qemu-ui-gtk
2018-08-24 09:40:27 <AinNero> oh, as separate package.. mh
2018-08-24 09:41:04 <mstanich_> yes, and it works
2018-08-24 09:41:15 <AinNero> it even has sdl then :3
2018-08-24 09:41:32 <AinNero> the point with headless still holds..
2018-08-24 09:41:44 <mstanich_> of course, qemu-ui-sdl
2018-08-24 09:42:55 <mstanich_> having console on serial port doesn't hurt and is often quite useful
2018-08-24 09:44:46 <AinNero> things without display available aren't that uncommon
2018-08-24 09:45:00 <AinNero> recently, there was an alpine user who wanted to install alpine on an pcengine apu2
2018-08-24 09:45:16 <AinNero> in the end, i made him an custom iso with console=ttyS0 enabled
2018-08-24 09:49:26 <mstanich_> AinNero: I installed AL on some ARM single board computers, different VMs and bare metal PCs, and serial console is always helpful to me
2018-08-24 09:49:50 <danieli> it's very useful for various embedded devices
2018-08-24 09:50:45 <mstanich_> danieli: agree with you, I would like it to be default on all installation images
2018-08-24 09:51:07 <danieli> using setup-alpine to semi-dynamically do it on an install basis would be nice
2018-08-24 10:02:58 <AinNero> im not fully satisfied with that solution
2018-08-24 10:03:53 <danieli> it's not going to work for completely headless installations, an alternative would be to supply separate installation media
2018-08-24 11:15:16 <AinNero> danieli: im generally not satisfied that creating an own boot medium rarely works on non-alpine systems
2018-08-24 11:15:24 <AinNero> maybe i'll put some effort into it someday
2018-08-24 11:15:32 <AinNero> so people can create their own boot media more easily
2018-08-24 12:21:29 <AinNero> for the last days, i've been studing the initramfs code
2018-08-24 12:21:49 <AinNero> and there are quite some things that should be cleaned up
2018-08-24 12:22:07 <AinNero> there is some sort of splash support, but only for diskless boot
2018-08-24 12:23:04 <AinNero> what the general direction we want to go with splash support?
2018-08-24 12:23:46 <danieli> probably in the direction of supporting plymouth
2018-08-24 12:28:00 <AinNero> that would be like 120k more in the initramfs, but fine if its optional as an mkinitfs feature
2018-08-24 12:30:19 <AinNero> i would change it into something configured at initramfs build time, instead of something configured by the apkovl
2018-08-24 12:32:23 <AinNero> on the other side, plymouth is an freedesktop thing...
2018-08-24 12:45:25 <TBB> now that grsecurity is not actively in the way of it, plymouth would be quite okay. at least until it too absolutely requires systemd (for now it's optional)
2018-08-24 13:07:52 <danieli> it's not like you'd use plymouth *and* expect a tiny initramfs
2018-08-24 13:08:38 <AinNero> im not comfy with plymouths "abstract everything away" interface, but it could be cool for impressing potential new users
2018-08-24 13:09:28 <skarnet> I'd like Alpine to be less about impressing people and more about doing things right and small
2018-08-24 13:09:57 <AinNero> not having any splash code in the initramfs would also be cool, skarnet
2018-08-24 13:10:05 <skarnet> indeed
2018-08-24 13:10:16 <skarnet> and that's my favorite solution :)
2018-08-24 13:13:58 <awilfox> plymouth is useful for "user friendly" scenarios
2018-08-24 13:14:07 <awilfox> it's entirely the opposite on servers
2018-08-24 13:14:08 <AinNero> awilfox: you were with the adelie people?
2018-08-24 13:14:19 <awilfox> AinNero: am
2018-08-24 13:16:07 <AinNero> it would be an optional mkinitfs feature, but the shell code driver would still be in the initramfs and needs to be maintained
2018-08-24 13:16:15 <AinNero> thanks for your input, i'll think a night about it
2018-08-24 13:28:34 <AinNero> ncopa: your opinion on plymouth in the initramfs?
2018-08-24 13:49:51 <mstanich_> please keep AL far from unneeded 'crap'
2018-08-24 14:05:57 <TBB> now if I was a devel, I'd have a pretty nasty response to that
2018-08-24 14:06:21 <TBB> cleanly stated tho, how very unprofessional
2018-08-24 14:07:05 <AinNero> im not an devel either, im probably just the only person in the world who bothers to read the mkinitfs code :P
2018-08-24 14:08:06 <TBB> don't worry, I've had to go through it a hundred times by now as I've needed some customizations
2018-08-24 14:09:13 <AinNero> there are some things that are utterly hacked together
2018-08-24 14:09:45 <AinNero> and some unreachable code sections...
2018-08-24 14:09:54 <AinNero> "organically grown" i guess
2018-08-24 14:10:05 <TBB> yup. but overall what it achieves is really nice, so the concept has been proven :)
2018-08-24 14:10:42 <AinNero> TBB: whats your committer ident?
2018-08-24 14:10:45 <TBB> now the only thing resultwise I'm still missing is some sort of a modular/feature-based initramfs-init construction dracut-style
2018-08-24 14:11:01 <AinNero> personally im fine with the single script
2018-08-24 14:11:07 <AinNero> easier to debug
2018-08-24 14:11:22 <TBB> I've so far only managed to commit a bug report or two and some APKBUILDs; my project restricts heavily how I can contribute to Alpine
2018-08-24 14:11:46 <TBB> I basically need a written permission for every contribution, and getting those has proven to be too slow to be of much value
2018-08-24 14:12:43 <TBB> and the punishment for a misjudgement from my part can be pretty severe
2018-08-24 14:13:22 <TBB> good thing is, I'll probably be at the end of this project soon so if I still have enthusiasm I can actually contribute without fear...
2018-08-24 14:13:42 <TBB> other than the fear of getting ridiculed, of course!
2018-08-26 13:22:14 <awilfox> just spent two hours trying to help a user debug why apk would print "I/O ERROR" during `apk fetch`, it turns out it was a DNS resolution error... that error message could use some work
2018-08-26 18:04:53 <mps> please ignore issue 9339, pressed 'Return' to early by mistake
2018-08-26 18:05:52 <mps> btw, is it possible to delete wrong bug/issue on bugs.a.o
2018-08-26 18:15:09 <AinNero> they can be closed
2018-08-26 18:15:16 <AinNero> closed issues disappear from most lists
2018-08-26 18:16:40 <mps> AinNero: already closed by 'Nero'. are you the Nero?
2018-08-26 18:16:46 <AinNero> probably
2018-08-26 18:17:00 <mps> AinNero: :)
2018-08-26 18:17:30 <mps> but, can submitter delete/close wrong reports somehow?
2018-08-26 18:17:39 <AinNero> i though you would be able to
2018-08-26 18:17:51 <AinNero> but i have no idea how the redmine privileges work
2018-08-26 18:18:57 <mps> I looked how to do that but couldn't find anything on bugs.a.o
2018-08-26 21:02:23 <terra> Guys, how can I protect some files in APKBUILD installed other places than /etc ?
2018-08-26 21:07:37 <[[sroracle]]> /etc/apk/protected_paths.d/ i think
2018-08-26 21:08:31 <[[sroracle]]> an example is in ca-certificates: https://git.alpinelinux.org/cgit/aports/tree/main/ca-certificates/APKBUILD#n40
2018-08-26 21:08:55 <[[sroracle]]> i think those start with - though because it's unprotecting them
2018-08-26 21:12:53 <terra> hmm
2018-08-26 21:12:58 <terra> thanks
2018-08-26 21:14:47 <[[sroracle]]> it supports other stuff too i guess https://git.alpinelinux.org/cgit/apk-tools/tree/src/database.c#n1268
2018-08-26 21:17:12 <terra> sadly, I'm trying to figure this out this like trying to find a hidden treasure
2018-08-27 07:58:44 <ncopa> AinNero: i dont have any strong opinion re plymouth, as long as it is optional
2018-08-27 07:59:12 <ncopa> i know that fabled had a graphical progressbar at somepoint using fbsplash
2018-08-27 08:46:28 <mps> ncopa: isn't simplycity one of main feature of the AL
2018-08-27 08:48:49 <mps> also I have no strong opinion about plymouth, but such things should be of the lower priorities IMHO
2018-08-27 09:04:49 <AinNero> blaudioslave: ^^
2018-08-27 09:06:04 <AinNero> ncopa: in the last hours i was thinking of some sort of hook system, so people can hack some features in
2018-08-27 09:06:12 <AinNero> but im not sure if the addition of complexity is worth it
2018-08-27 09:43:08 <ncopa> AinNero: I have also been thinking of hook ssytem
2018-08-27 09:43:35 <ncopa> so far i dont think it has been worth it
2018-08-27 09:53:55 <AinNero> there have been recent things to move functionality from the initramfs into the firstboot service
2018-08-27 09:54:06 <AinNero> i guess to remove openssl from the initramfs
2018-08-27 09:54:20 <AinNero> but might we need it anyways for fetching the apkovl?
2018-08-27 09:54:51 <clandmeter> apk needs ssl
2018-08-27 09:55:09 <clandmeter> or you mean openssl cmd?
2018-08-27 09:55:49 <clandmeter> i dont remember we include it?
2018-08-27 09:56:39 <AinNero> the busybox wget client needs it or the ssl_client from busybox
2018-08-27 09:56:46 <AinNero> the latter was rather broken, afaik
2018-08-27 09:57:30 <clandmeter> ncopa made a helper
2018-08-27 10:00:34 <clandmeter> AinNero, https://git.alpinelinux.org/cgit/aports/commit/main/busybox?id=1d0560a9b6b5597b191e5aff69a31c2fe0aba273
2018-08-27 10:01:11 <AinNero> ah
2018-08-27 10:01:31 <AinNero> this removes the rationale for moving the fetching of ssh keys from the firstboot service, right?
2018-08-27 10:01:43 <AinNero> *from the initramfs into the firstboot service
2018-08-27 10:01:53 <AinNero> because im not entirely satisfied how its currently split up
2018-08-27 10:06:31 <clandmeter> firstboot was implement just before this commit iirc
2018-08-27 10:06:59 <clandmeter> so the idea was to keep initramfs small and handle things over to firstboot
2018-08-28 11:46:55 <Bluefoxicy> Any chance of dotnet core 2.1 showing up in edge some day?
2018-08-28 12:07:54 <danieli> not just yet, but soon enough, sure
2018-08-28 12:08:00 <danieli> upstream put work into it https://github.com/dotnet/corert/issues/4552
2018-08-28 12:14:31 <Bluefoxicy> Nice
2018-08-28 12:14:50 <Bluefoxicy> I'm looking at building electronic voting machines, and I want to write the software in .NET and run it on Alpine Linux
2018-08-28 12:15:43 <Bluefoxicy> the ideal scenario is a minimal image running on arm64 booting on a Raspberry Pi, with the initrd copying the SD card to a tmpfs to run from RAM
2018-08-28 12:16:36 <Bluefoxicy> besides having Element14 remove the bluetooth/wifi radio chip in a custom batch so there's no network, this nets me a very small, fully-open code base
2018-08-28 12:53:36 <danieli> if only I used .NET I'd help out a bit more, but I cannot prioritize it
2018-08-28 17:31:30 <mps> anyone can help help with cmake APKBUILD, I want to disable include dir in built apk
2018-08-28 17:33:09 <mps> removed $pkgname-dev from subpackages definition but the abuild still build '-dev' and puts it in pkg subdir
2018-08-28 17:33:39 <mps> trying to make svxlink apk for AL
2018-08-30 07:30:18 <AinNero> chisa: hi!
2018-08-30 09:09:05 <chisa> hey AinNero :3
2018-08-30 16:44:47 <sleepycat> I am working on an alpine package and was wondering if people could give me a little feedback on it. Any suggestions? https://gitlab.com/sleepycat/arangodb-alpine/tree/master/arangodb
2018-08-30 17:21:24 <_ikke_> sleepycat: let me check
2018-08-30 17:24:03 <_ikke_> sleepycat: still here?
2018-08-30 17:49:33 <sleepycat> _ikke_: Yup still here.
2018-08-30 17:51:49 <_ikke_> 1) Indentation is off (probably mixed tab / spaces. 2) builddir should be set to the directory you go to on line 28. 3) Not sure if you're supposed to use the msg / msg2 functions. 4) The project itself does not have an installation step?
2018-08-30 17:52:42 <_ikke_> 5) you should not pass -j to make. abuild will do that based on what is set in /etc/abuild.conf
2018-08-30 18:07:13 <sleepycat> _ikke_: Nice! Thank you.
2018-08-30 18:38:43 <terra> ncopa: I'm trying to compile chromium but I got this: https://paste.ubuntu.com/p/S8Hb7g8JCQ/
2018-08-30 18:40:06 <terra> I remember you have mentioned about fontconfig/freetype issues during upgrading process
2018-08-31 05:59:09 <ncopa> terra: I dont know