2017-11-01 17:08:04 <nmeum> hm, go 1.9.2 seems to have build fine on armhf before
2017-11-01 17:11:47 <nmeum> in go 1.9 they introduced parallel compilation might be an issue with that
2017-11-02 05:00:43 <travis-ci> apk-tools:master [passed] |William Pitcock| solver: implement support for choosing default virtuals | https://travis-ci.org/alpinelinux/apk-tools/builds/296099225
2017-11-02 13:21:48 <rnalrd> duh
2017-11-02 16:40:24 <travis-ci> apk-tools:master [passed] |William Pitcock| solver: only select a default if there is at least one provider with a declared provider_priority | https://travis-ci.org/alpinelinux/apk-tools/builds/296357769
2017-11-02 17:26:42 <travis-ci> apk-tools:master [passed] |William Pitcock| solver: prefer highest requirer count above installed status | https://travis-ci.org/alpinelinux/apk-tools/builds/296378641
2017-11-03 16:22:41 <rdutra> algitbot: retry master
2017-11-04 18:36:57 <rdutra> algitbot: retry master
2017-11-04 18:44:31 <rdutra> algitbot: retry master
2017-11-05 00:37:43 <jirutka> retry master
2017-11-05 00:37:57 <jirutka> algitbot: retry master
2017-11-05 20:28:33 <nmeum> hmhm
2017-11-05 20:29:48 <nmeum> this worked fine on the travis CI runner
2017-11-05 20:29:48 <nmeum> > 
2017-11-05 20:29:53 <nmeum> > ALL TESTS SUCCESSFUL.
2017-11-05 20:29:55 <nmeum> strange…
2017-11-05 20:43:47 <nmeum> cannot even reproduce this on my machine
2017-11-05 20:52:19 <nmeum> aha!
2017-11-05 21:35:30 <rdutra> algitbot: retry master
2017-11-05 21:38:55 <rdutra> algitbot: retry master
2017-11-06 10:31:44 <rdutra> algitbot: retry master
2017-11-07 00:01:35 <rdutra> algitbot: retry master
2017-11-10 04:24:53 <kaniini> that's weird
2017-11-10 09:49:23 <clandmeter> algitbot: retry master
2017-11-10 20:05:59 <rdutra> algitbot: retry master
2017-11-10 20:35:38 <rdutra> algitbot: retry master
2017-11-10 21:27:33 <jirutka> retry master
2017-11-10 21:27:39 <jirutka> algitbot: retry master
2017-11-12 23:18:22 <jirutka> they really wants to piss me off! libressl detection in rust-openssl is fixed to exact version numbers, i.e. 2.6.2 is ok, 2.6.3 doesn’t work due to this stupid detection
2017-11-13 18:36:53 <rdutra> algitbot: retry master
2017-11-13 19:19:09 <rdutra> algitbot: retry master
2017-11-13 19:24:15 <rdutra> algitbot: retry master
2017-11-14 18:50:52 <rdutra> algitbot: retry master
2017-11-16 13:35:12 <rdutra> algitbot: retry master
2017-11-16 15:33:06 <ncopa> sigh...
2017-11-16 15:41:33 <ncopa> hum.. ok? mosh build second try...
2017-11-16 15:42:44 <rdutra> ncopa: seems community/lua-http 0.2-r0 is hanging on build-edge-ppc64le
2017-11-16 15:46:39 <ncopa> ok, i killed it
2017-11-16 15:46:58 <ncopa> it was lua5.1 /usr/bin/busted
2017-11-16 16:16:35 <rdutra> algitbot: retry master
2017-11-16 16:19:46 <rdutra> algitbot: retry master
2017-11-16 16:20:14 <rdutra> algitbot: retry master
2017-11-16 18:19:49 <rdutra> algitbot: retry master
2017-11-16 18:26:07 <rdutra> hmm. the_silver_searcher should build fine in ppc64le edge (as it does in 3.7 builder) as patch https://github.com/alpinelinux/aports/commit/bf0205104ba520e7d2fea17704a1cc131551c3d3  is merged
2017-11-17 12:23:28 <ncopa> algitbot: rebuild master
2017-11-17 12:24:01 <ncopa> algitbot: retry master
2017-11-17 15:45:19 <ncopa> jirutka: looks like may packages fails due to busted ^^^
2017-11-17 15:46:49 <rdutra> not finding  /usr/bin/busted-5.3 in 3.7 builders (as lua5.3 is been installed)
2017-11-17 16:04:59 <rdutra> algitbot: retry master
2017-11-17 16:25:03 <ncopa> looks like lua-busted is not built http://build.alpinelinux.org/buildlogs/build-3-7-ppc64le/community/lua-busted/
2017-11-17 16:29:00 <ncopa> oh
2017-11-17 16:29:08 <ncopa> i think i know what may happen
2017-11-17 16:32:05 <rdutra> ncopa: it is missing to build lua-busted?
2017-11-17 16:40:37 <ncopa> no its there
2017-11-17 16:40:59 <ncopa> i think it may be a bug in apk
2017-11-17 16:41:17 <ncopa> install_if is not installed if repository is specified on commandline
2017-11-17 16:41:26 <ncopa> lua-busted is in community repo
2017-11-17 16:41:33 <ncopa> the /etc/apk/repositories has only main
2017-11-17 16:41:35 <ncopa> not community
2017-11-17 16:41:48 <jirutka> sigh
2017-11-17 16:42:04 <ncopa> but while the dependencies are added, the repodest is added
2017-11-17 16:42:14 <ncopa> but for some reason it does not pull in the install_if
2017-11-17 16:42:21 <jirutka> why is community specified only on cli? and why it can even  affect install_if?
2017-11-17 16:42:54 <jirutka> and why it worked until now? lua-busted is not new
2017-11-17 16:42:56 <ncopa> because otherwise will we not catch dependencies main -> community
2017-11-17 16:43:40 <ncopa> apk would happily install deps from community while building packages in main
2017-11-17 16:43:58 <jirutka> hm, i do it differently on Travis, edits /etc/apk/repositories before each build
2017-11-17 16:44:25 <ncopa> i think for the edge builders, we have community in /etc/apk/repositories systemwide
2017-11-17 16:44:38 <ncopa> otherwise would not testing repo build...
2017-11-17 16:44:44 <ncopa> (yeah i know its a mess)
2017-11-17 16:44:54 <ncopa> i have a suggestion for a workaround
2017-11-17 16:45:08 <ncopa> do not rely on install_if if you have explicit depends
2017-11-17 16:45:23 <jirutka> https://github.com/alpinelinux/aports/blob/master/.travis/build-pkgs#L34-L50
2017-11-17 16:45:42 <jirutka> https://github.com/alpinelinux/aports/blob/master/.travis/build-pkgs#L73-L74
2017-11-17 16:45:49 <ncopa> sudo sh -c "printf '%s\n' $repos > /etc/apk/repositories"
2017-11-17 16:45:56 <ncopa> only works on disposable systems
2017-11-17 16:46:14 <jirutka> it updates it before each build
2017-11-17 16:46:40 <jirutka> yeah, it’s hackish, I don’t remember why I did it in this way, but our whole build system is messy hack, so…
2017-11-17 16:46:45 <ncopa> i though it builds the machine, the entire vm from scratch each build?
2017-11-17 16:46:59 <jirutka> yes, but that’s not important here
2017-11-17 16:47:14 <jirutka> if you don’t run multiple separated builds in parallel
2017-11-17 16:47:15 <ncopa> the weird thing is why its not gets installed
2017-11-17 16:47:30 <ncopa> because apk should find the install_if
2017-11-17 16:47:36 <jirutka> yeah, we should fix it
2017-11-17 16:48:20 <ncopa> but i dont like the indirect depends in makedepends
2017-11-17 16:48:23 <jirutka> wait a moment, lua5.1-busted and lua5.2-busted are installed, but not lua5.3-busted o.O http://build.alpinelinux.org/buildlogs/build-3-7-ppc64le/community/lua-lpeg_patterns/lua-lpeg_patterns-0.4-r1.log
2017-11-17 16:48:39 <ncopa> hum
2017-11-17 16:48:48 <ncopa> ok... i have no clue whats going on then
2017-11-17 16:49:08 <rdutra> jirutka: yes, it is missing just lua5.3-busted
2017-11-17 16:49:29 <rdutra> jirutka: and apk search for lua5.3-busted do not finds the package in 3.7 builders
2017-11-17 16:49:53 <jirutka> hm, so we have even bigger problem
2017-11-17 16:50:05 <ncopa> because /etc/apk/repositories does not contain "community", only "main"
2017-11-17 16:50:05 <jirutka> how the hell can happen that some subpackage get lost?
2017-11-17 16:50:13 <ncopa> its not lost
2017-11-17 16:50:22 <ncopa> it just fail pull in the install_if as expected
2017-11-17 16:51:07 <ncopa> but i dont like to depend on install_if for makedepends
2017-11-17 16:51:14 <jirutka> hm, right, lua5.3-busted is in the 3.7 repository
2017-11-17 16:51:33 <jirutka> why?
2017-11-17 16:51:50 <ncopa> because it makes calculating the build order significantly more complicated
2017-11-17 16:52:21 <jirutka> hm, yeah, install_if is not even considered in dependency graph
2017-11-17 16:52:50 <ncopa> lua-busted it
2017-11-17 16:52:51 <ncopa> is*
2017-11-17 16:53:00 <jirutka> btw it is not a make depend, but check depend ;)
2017-11-17 16:53:01 <ncopa> so it should work anyway
2017-11-17 16:53:03 <jirutka> but that doesn’t matter
2017-11-17 16:53:17 <ncopa> build order calculation gets more compilcated
2017-11-17 16:53:47 <ncopa> and its trivial to make it depend explicitly on lua5.3-busted
2017-11-17 16:54:31 <jirutka> no, on lua5.1-busted, lua5.2-busted, lua5.3-busted
2017-11-17 16:54:34 <jirutka> https://github.com/alpinelinux/aports/blob/master/community/lua-lpeg_patterns/APKBUILD
2017-11-17 16:54:49 <ncopa> you loop luaver
2017-11-17 17:00:43 <ncopa> http://tpaste.us/DkVl
2017-11-17 17:04:16 <jirutka> grrr, we need lua-busted (w/o) version for binary on PATH
2017-11-17 17:04:51 <jirutka> supporting multiple versions of lang is insane
2017-11-17 17:05:03 <jirutka> unfortunately we need lua5.1 for luajit
2017-11-17 17:05:10 <jirutka> but I’d like to get rid at least of lua5.2
2017-11-17 17:07:11 <jirutka> and maybe support lua5.1 *only* for LuaJIT
2017-11-17 17:07:41 <jirutka> I wish upstream *finally* implement Lua 5.3 into LuaJIT, but it doesn’t look like it’ll be soon :(
2017-11-17 17:13:12 <ncopa> jirutka: thanks for cleaning that up
2017-11-17 17:13:36 <ncopa> there might be other build failures now since i enabled halt and report on errors
2017-11-17 17:16:56 <jirutka> any idea why openjdk7 fails to build?
2017-11-17 17:21:08 <ncopa> i havent looked at it
2017-11-17 17:21:12 <ncopa> its weekend for me now
2017-11-17 17:21:18 <ncopa> will check it on monday
2017-11-19 01:36:05 <kaniini> algitbot: retry master
2017-11-19 01:38:47 <kaniini> algitbot: retry master
2017-11-19 01:39:38 <kaniini> algitbot: retry master
2017-11-19 01:40:02 <kaniini> algitbot: retry master
2017-11-19 01:41:14 <kaniini> grr
2017-11-19 12:46:28 <clandmeter> ollieparanoid: ^
2017-11-19 16:17:29 <ollieparanoid[m]> clandmeter: thanks! \o/
2017-11-19 19:12:26 <ollieparanoid[m]> clandmeter: now build-edge-armhf fails at gns3-server. I guess because it misses this commit? https://github.com/alpinelinux/aports/commit/40d074c249ab8664d462b112900a0d37c9487a28 -- (lately all .tar.gz archives generated by github had the checksum changed)
2017-11-19 23:22:07 <clandmeter> algitbot: retry master
2017-11-20 11:27:09 <jirutka> algitbot: retry master
2017-11-20 16:56:22 <travis-ci> mkinitfs:master [passed] |Andrej| mkinitfs: fix building initfs without a kernel | https://travis-ci.org/alpinelinux/mkinitfs/builds/304842084
2017-11-21 11:40:43 <ncopa> #!python3 -s
2017-11-21 11:40:52 <ncopa> itstool is broken for some reason
2017-11-21 11:41:57 <ncopa> bah
2017-11-21 11:42:23 <fabled> ?
2017-11-21 11:43:07 <ncopa> oh
2017-11-21 11:43:29 <ncopa> i thought it was me who failed to verify that itstool actually worked
2017-11-21 11:43:49 <ncopa> i did test that itstool fixes actually work - i thought
2017-11-21 11:43:53 <fabled> it works for me
2017-11-21 11:44:01 <ncopa> and it looks like i did
2017-11-21 11:44:02 <ncopa> however
2017-11-21 11:44:18 <fabled> ppc64le specific issue?
2017-11-21 11:44:25 <ncopa> looks like someone cleaned up itstool after me
2017-11-21 11:44:34 <ncopa> commit b34db16729ab78563258d16e0be4d5d67cb2f071
2017-11-21 11:44:34 <ncopa> Author: Jakub Jirutka <jakub@jirutka.cz>
2017-11-21 11:44:34 <ncopa> Date:   Sat Nov 18 01:14:19 2017 +0100
2017-11-21 11:44:34 <ncopa>     main/itstool: don't rely on python version autodetection
2017-11-21 11:44:38 <ncopa> that wone broke it
2017-11-21 11:45:13 <fabled> no pkgrel bumb
2017-11-21 11:45:29 <fabled> different binary packages on different arch
2017-11-21 11:45:46 <ncopa> after that commit it ended up as #!python -s
2017-11-21 11:45:54 <ncopa> instead of #!/usr/bin/python ...
2017-11-21 11:45:57 <ncopa> python3 that is
2017-11-21 11:46:21 <fabled> ah
2017-11-21 11:46:25 <fabled> explains it
2017-11-21 11:46:34 <fabled> so the var needs full path
2017-11-21 11:47:03 <ncopa> seems so yes
2017-11-21 11:47:30 <fabled> does not still explain the error in https://github.com/alpinelinux/aports/pull/2782
2017-11-21 11:47:42 <fabled> maybe thats older issue
2017-11-21 11:48:08 <ncopa> i think that might be solved by my previous commits
2017-11-21 11:48:17 <ncopa> i'll have a look at that after i fixed itstool
2017-11-21 11:49:45 <fabled> ok, so linux-rpi failure due to missing 'bc' is different still
2017-11-21 11:50:44 <ncopa> yup
2017-11-21 11:50:47 <ncopa> different issue
2017-11-21 12:21:34 <ncopa> i have no clue why girara test fails
2017-11-21 13:33:08 <fabled> ncopa, build-3-7-armhf does not report failures here?
2017-11-21 13:56:12 <ncopa> fabled: apparently not
2017-11-21 13:56:15 <ncopa> i'll fix it
2017-11-21 14:00:05 <fabled> whee, build-3-7-armhf / main finished
2017-11-21 14:54:39 <jirutka> ncopa: fabled: about error in itstool, the problem was on armhf builder, there was python2 installed, so it picked python2 instead of python3
2017-11-21 14:54:52 <ncopa> jirutka: i figured
2017-11-21 14:54:58 <ncopa> thanks for fixing it
2017-11-21 14:55:12 <ncopa> or....
2017-11-21 14:55:19 <jirutka> ncopa: I tried to build it locally and it worked, where does it set #!python3?
2017-11-21 14:55:24 <ncopa> thanks for figuring out the proper fix
2017-11-21 14:55:38 <jirutka> you’re welcome
2017-11-21 14:55:48 <ncopa> *shrug*, it should be ok now
2017-11-21 14:55:53 <ncopa> teamwork ;)
2017-11-21 14:55:57 <jirutka> the question i got no answer for is how the hack get python2 installed on the builder and not cleaned
2017-11-21 14:56:22 <ncopa> edge builder?
2017-11-21 14:56:45 <jirutka> yes
2017-11-21 14:57:35 <ncopa> thats definitively a good question
2017-11-21 14:57:44 <ncopa> i dunno
2017-11-21 14:57:49 <ncopa> its building openjdk7 now
2017-11-21 14:57:54 <ncopa> will probably do so til tomorrow
2017-11-21 14:58:42 <ncopa> but i have no clue why the girara test fails
2017-11-21 14:58:55 <ncopa> i dunno if the test is broken or if its a bug in the lib
2017-11-21 22:10:59 <rdutra> algitbot: retry master
2017-11-22 12:33:04 <rdutra> algitbot: retry master
2017-11-22 12:35:39 <rdutra> algitbot: retry master
2017-11-22 15:36:47 <ncopa> fcolista: can you please have a look at the tomahawk build? or move it to unmaintained
2017-11-22 15:48:36 <fcolista> ncopa, sure
2017-11-22 15:54:16 <fcolista> trying to build against qt5
2017-11-22 16:15:34 <clandmeter> ncopa: are you working on geany?
2017-11-22 16:21:38 <fcolista> looks that for domino effect other packages needs to be rebuilt against qt5
2017-11-22 16:21:41 <fcolista> working on it
2017-11-22 18:56:01 <fcolista> qt5 in main: when are we going to do this big jump?
2017-11-22 23:46:09 <rdutra> algitbot: retry master
2017-11-23 07:29:26 <ncopa> only aarch64 is left now
2017-11-23 12:52:24 <ncopa> i think i have a fix for aarch64 in a minute
2017-11-23 15:07:29 <travis-ci> mkinitfs:master [passed] |Natanael Copa| ==== release 3.2.0 ==== | https://travis-ci.org/alpinelinux/mkinitfs/builds/306361919
2017-11-23 15:27:06 <ncopa> i aborted the build of firefox-esr on aarch64 because i updated to 52.5.0
2017-11-23 15:27:10 <ncopa> no need to build it twice
2017-11-23 15:48:07 <ScrumpyJack> algitbot: help
2017-11-23 16:45:14 <rdutra> algitbot: retry master
2017-11-23 17:04:50 <rdutra> algitbot: retry master
2017-11-24 14:55:05 <clandmeter> jirutka: thx
2017-11-25 02:02:38 <jirutka> algitbot: retry master
2017-11-27 21:29:38 <rdutra> algitbot: retry master
2017-11-28 14:34:56 <travis-ci> apk-tools:master [passed] |Timo Teräs| fetch: print additional error diagnostics in verbose mode | https://travis-ci.org/alpinelinux/apk-tools/builds/308457269
2017-11-28 18:31:11 <ncopa> seems now
2017-11-28 18:31:57 <rdutra> ncopa :(
2017-11-28 18:32:48 <ncopa> bah
2017-11-28 18:33:08 <rdutra> ncopa: (04:28:21 PM) algitbot: build-edge-s390x: failed to build libassuan: http://build.alpinelinux.org/buildlogs/build-edge-s390x/main/libassuan/libassuan-2.4.3-r2.log
2017-11-28 18:33:31 <rdutra> ncopa: I think this is the same  a problem that happened in ppc64le edge builder
2017-11-28 18:33:38 <ncopa> yeah
2017-11-28 18:33:40 <ncopa> looks so
2017-11-28 18:33:42 <rdutra> ncopa: I think musl did not upgraded there fine and you manually upgraded
2017-11-28 18:33:44 <ncopa> needs upgrade
2017-11-28 18:34:26 <rdutra> algitbot: retry master
2017-11-28 18:38:48 <ncopa> i doubbt make -j1 helps
2017-11-28 18:39:06 <rdutra> ncopa: :( . In my ppc64le container it worked
2017-11-28 18:40:22 <rdutra> I mean, it built fine
2017-11-28 18:40:37 <ncopa> it does not look like it runs the tests any faster withiout the -j1
2017-11-28 18:40:57 <ncopa> oh, im wrong
2017-11-28 18:44:12 <ncopa> rdutra: do you have coreutils installed on your system?
2017-11-28 18:45:30 <ncopa> yeah, coreutils seems to solve it
2017-11-28 18:45:38 <rdutra> ncopa: umm I dont have it installed
2017-11-28 18:57:33 <ncopa> whatta
2017-11-28 18:57:44 <ncopa> ncopa-edge-x86_64:~/aports$ groups
2017-11-28 18:57:44 <ncopa> ncopa wheel readproc kvm abuild racktables-lsm
2017-11-28 18:57:54 <ncopa> how did I end up in the racktables-lsm group?
2017-11-28 18:58:11 <ncopa> buildozer is there too
2017-11-28 19:13:42 <ncopa> ok i thikn i know whats wrong with augeas
2017-11-28 19:14:03 <ncopa> apparently the testuite check a group you are in 
2017-11-28 19:14:16 <ncopa> the it `ls -l something`
2017-11-28 19:14:31 <ncopa> and check the group name from ls output
2017-11-28 19:15:42 <ncopa> but busybox ls truncate the groupname