2017-10-01 07:48:33 <kaniini> s390x needs abuild 3
2017-10-01 21:05:31 <rdutra> algitbot: retry master
2017-10-02 05:51:32 <kaniini> build-edge-armhf: retry
2017-10-02 05:51:39 <kaniini> etc
2017-10-02 05:51:46 <kaniini> algitbot: retry master
2017-10-02 06:01:21 <kaniini> hmm how long does it take the build log to upload
2017-10-02 07:52:00 <ncopa> i upgraded the s390x builder
2017-10-02 07:52:16 <ncopa> the problem was that abuild got new feature
2017-10-02 07:52:27 <ncopa> and apkbuild used the new feature before abuild got updated
2017-10-02 07:52:41 <ncopa> which caused build to fail, and if it fails it does not upgrade
2017-10-02 07:53:10 <ncopa> it is generally a good idea to implement a new feature in abuild, push it
2017-10-02 07:53:21 <ncopa> but wait a bit before actually using it
2017-10-02 19:43:07 <rdutra> algitbot: retry master
2017-10-04 01:02:23 <rdutra> algitbot: retry master
2017-10-04 14:06:45 <travis-ci> apk-tools:master [passed] |Timo Teräs| package: remove package script after use (fixes #7974) | https://travis-ci.org/alpinelinux/apk-tools/builds/283206774
2017-10-05 14:07:51 <travis-ci> apk-tools:master [passed] |Timo Teräs| libfetch: improve ssl connections | https://travis-ci.org/alpinelinux/apk-tools/builds/283708751
2017-10-06 10:33:48 <travis-ci> apk-tools:master [passed] |Timo Teräs| libfetch: improve openssl/libressl compatibility | https://travis-ci.org/alpinelinux/apk-tools/builds/284139837
2017-10-06 15:19:11 <travis-ci> apk-tools:master [passed] |Timo Teräs| libfetch: fix certificate host name check | https://travis-ci.org/alpinelinux/apk-tools/builds/284253512
2017-10-06 16:18:34 <travis-ci> apk-tools:v2.8.0 [passed] |Timo Teräs| apk-tools-2.8.0 | https://travis-ci.org/alpinelinux/apk-tools/builds/284280399
2017-10-06 16:24:13 <travis-ci> apk-tools:master [passed] |Timo Teräs| apk-tools-2.8.0 | https://travis-ci.org/alpinelinux/apk-tools/builds/284280416
2017-10-09 19:39:26 <clandmeter> algitbot: retry master
2017-10-09 20:01:16 <clandmeter> wtf, why is this not building?
2017-10-09 20:02:06 <clandmeter> fabled, ncopa?
2017-10-09 21:03:46 <ncopa> who is not building?
2017-10-09 21:19:59 <clandmeter> ncopa: sshfs
2017-10-09 21:30:52 <kaniini> it is possible that
2017-10-09 21:31:02 <kaniini> sshfs is not being pulled in because of openssh-client split
2017-10-09 21:58:31 <clandmeter> pulled in?
2017-10-09 22:37:11 <kaniini> to the list of things to build
2017-10-10 08:48:13 <travis-ci> apk-tools:master [passed] |Timo Teräs| db: handle default root correctly for /proc | https://travis-ci.org/alpinelinux/apk-tools/builds/285947232
2017-10-12 01:58:02 <rdutra> algitbot: retry master
2017-10-12 02:11:26 <kaniini> rdutra: i cant reproduce that build failure which makes me nervous
2017-10-12 02:11:56 <kaniini> rdutra: i suspect x86/x86-64 problem is grsec
2017-10-12 10:46:03 <travis-ci> apk-tools:master [passed] |Timo Teräs| io: fix skip and splice to detect unexpected end-of-file | https://travis-ci.org/alpinelinux/apk-tools/builds/286971958
2017-10-13 17:21:40 <ncopa> the py3-async-timeout issue is a bit annoying
2017-10-13 17:21:53 <ncopa> i think it needs paxmarking python3
2017-10-17 12:40:04 <fabled> hum. must be parallel build
2017-10-17 17:31:12 <rdutra> to unblock s390x builder: https://github.com/alpinelinux/aports/pull/2507
2017-10-17 18:01:29 <nmeum> applied. thanks!
2017-10-19 01:58:51 <kaniini> well, i would fix s390x if i had an s390x machine ;)
2017-10-19 02:03:04 <craftyguy[m]> In the meantime, algitbot is going to continue doing the exact same thing and expect a different result!
2017-10-19 04:42:00 <kaniini> note to self.
2017-10-19 04:42:04 <kaniini> check git branch
2017-10-19 04:42:05 <kaniini> lols
2017-10-19 05:26:52 <kaniini> ... it builds????
2017-10-19 05:38:27 <nmeum> wow
2017-10-19 05:38:44 <kaniini> it hasn't failed yet anyway
2017-10-19 05:38:50 <kaniini> i know it has to compile a ton of .go files
2017-10-19 05:38:57 <kaniini> but iirc it was crashing last time ;)
2017-10-19 05:39:03 <kaniini> so i am guessing it builds
2017-10-19 06:03:32 <kaniini> grudgingly, i sign up for an ibm linuxone trial
2017-10-19 06:03:39 <kaniini> my inbox is ready
2017-10-19 06:04:28 <kaniini> i expect to receive at least 5 emails per day about systems i do not have the credit rating to acquire
2017-10-19 06:26:20 <nmeum> hm, well
2017-10-19 06:26:43 <nmeum> kaniini: the guile build did fail after a while :(
2017-10-19 06:27:27 <kaniini> yes seems so
2017-10-19 06:27:44 <kaniini> linuxone is really slooooooow
2017-10-19 06:28:11 <kaniini> now i see why our builder is slow
2017-10-19 06:30:25 <kaniini> nmeum: my plan is to basically build it by hand
2017-10-19 06:30:35 <kaniini> nmeum: with ulimit -c unlimited and see if i can get a coredump
2017-10-19 06:30:50 <kaniini> but,t o be honest
2017-10-19 06:30:59 <kaniini> i think we need better s390x builder ;/
2017-10-19 06:31:08 <kaniini> but it's a chicken and egg problem
2017-10-19 06:31:43 <kaniini> for anyone to care about an adelie port (which could be used to bankroll alpine's s390x building needs too), the alpine port needs to actually be ok
2017-10-19 06:33:24 <nmeum> I believe the problem is not s390x builder but the fact that too few people are working on the s390x port
2017-10-19 06:33:57 <kaniini> yes/no.  s390x builder is pretty slow
2017-10-19 06:34:24 <kaniini> and, there's little non-commercial interest in s390x because to get an s390x machine you have to basically drop enough cash to buy a house
2017-10-19 06:34:28 <kaniini> which is a problem
2017-10-19 06:34:57 <nmeum> indeed
2017-10-19 06:35:52 <nmeum> I still don't understand though why you believe that compilation speed is a problem
2017-10-19 06:36:17 <kaniini> well it's a problem to me because my test environment hasn't even booted yet
2017-10-19 06:36:18 <kaniini> :P
2017-10-19 06:40:10 <nmeum> :)
2017-10-19 06:49:31 <kaniini> hmm, this linuxone vm has 20k bogomips
2017-10-19 06:49:36 <kaniini> i guess that's nothing to sneeze at
2017-10-19 07:01:12 <kaniini> ~/aports/main/guile/src/guile-2.2.2/bootstrap $ GUILE_AUTO_COMPILE=0 ../meta/build-env guild compile --verbose --target="s390x-ibm-linux-musl" -O1 -L "/home/buildozer/aports/main/guile/src/guile-2.2.2/module" -L "/home/buildozer/aports/main/guile/src/guile-2.2.2/guile-readline" -o "language/cps/closure-conversion.go"
2017-10-19 07:01:14 <kaniini> 
ACTION waits patiently

2017-10-19 07:10:28 <kaniini> hmm, we use -fomit-frame-pointer on s390x
2017-10-19 07:10:30 <kaniini> that seems bad
2017-10-19 07:27:01 <kaniini> ok so from what ig ather
2017-10-19 07:27:01 <kaniini> it is stack overflow issue
2017-10-19 07:27:46 <fabled> still?
2017-10-19 07:27:54 <fabled> kaniini, we added a patch for that
2017-10-19 07:27:58 <fabled> but maybe it's not enough
2017-10-19 07:28:15 <fabled> we should file a bug
2017-10-19 07:28:19 <fabled> it's doing unlimited recursion
2017-10-19 07:28:43 <kaniini> yes, that seems to be the problem
2017-10-19 07:29:34 <kaniini> fwiw
2017-10-19 07:29:41 <kaniini> most distributions seem to be staying at guile 2.0
2017-10-19 07:30:15 <fabled> kaniini, there's other issues than the recursion problem?
2017-10-19 07:30:26 <kaniini> i'm not sure
2017-10-19 07:30:30 <kaniini> i just observed that
2017-10-19 07:30:45 <kaniini> however
2017-10-19 07:30:59 <kaniini> stack frames on s390x are a lot bigger than on other archs
2017-10-19 07:31:07 <kaniini> so yes, i think our patch doesn't expand it enough
2017-10-19 07:32:12 <fabled> yeah, the size needed depends on stack frame size. the recursion level was huge what we saw
2017-10-19 07:32:18 <fabled> but also depends on instance
2017-10-19 07:32:26 <fabled> apparently related to some hash randomization or similar
2017-10-19 07:32:31 <kaniini> when it crashed, was about 16k frames deep
2017-10-19 07:32:50 <fabled> on x86 it worked half of the time without any patch
2017-10-19 07:33:03 <fabled> so it's definitely some worst case scenario that triggers it
2017-10-19 07:33:25 <kaniini> on s390x it fails 100% of the time
2017-10-19 07:33:33 <kaniini> but guile 2.0 seemed ok
2017-10-19 07:34:00 <kaniini> i feel like we should downgrade guile
2017-10-19 07:34:21 <kaniini> i have a gut feeling that guile 2.2 is going to cause a major problem in 3.7 release
2017-10-19 07:34:41 <kaniini> i will write to alpine-devel about it
2017-10-19 07:34:52 <fabled> kaniini, maybe post on the github PR?
2017-10-19 07:36:15 <kaniini> ah github.  of course this originates from there
2017-10-19 07:36:43 <fabled> the original update to new guile was github pr
2017-10-19 07:38:12 <kaniini> yes.  from vakartel.
2017-10-19 07:38:28 <kaniini> i hate to do this, but i think we need to downgrade it
2017-10-19 07:38:45 <kaniini> and require a test build of the distribution to touch it
2017-10-19 07:39:07 <kaniini> in other words, i don't think we should accept github PRs for it
2017-10-19 07:46:41 <kaniini> i think a lesson here is that we need to get the other builders jenkins stuff live
2017-10-19 07:46:56 <kaniini> because ppc64le and s390x were canaries for this
2017-10-19 07:47:06 <kaniini> and we didn't see it until post-merge
2017-10-19 07:47:31 <fabled> yes
2017-10-19 08:01:05 <fabled> huh: make[3]: *** [Makefile:42629: moddep.lst] Error 1
2017-10-19 08:01:09 <fabled> on armhf grub
2017-10-19 08:01:12 <fabled> worked for me
2017-10-19 08:52:25 <fabled> algitbot: retry master
2017-10-19 15:55:42 <kaniini> humm
2017-10-19 15:55:52 <kaniini> s390x builder hasn't uploaded to community
2017-10-19 15:58:00 <kaniini> cool
2017-10-20 08:31:34 <clandmeter> wow nice commit msg...
2017-10-20 08:31:40 <rnalrd> :D
2017-10-20 08:32:13 <clandmeter> atleast its clear :D
2017-10-24 14:56:43 <travis-ci> lua-aports:v0.7.0 [passed] |Natanael Copa| ==== release 0.7.0 ==== | https://travis-ci.org/alpinelinux/lua-aports/builds/292123691
2017-10-24 14:57:11 <travis-ci> lua-aports:master [passed] |Natanael Copa| ==== release 0.7.0 ==== | https://travis-ci.org/alpinelinux/lua-aports/builds/292123685
2017-10-26 15:42:23 <rdutra> ncopa: ^done
2017-10-26 15:43:50 <kaniini> i was going to say
2017-10-26 15:43:51 <kaniini> uhhhhhhhh
2017-10-26 15:43:57 <kaniini> but then i saw it was prepare, not package
2017-10-26 15:45:00 <ncopa> kaniini: LOL!
2017-10-26 15:45:17 <rdutra> :)
2017-10-27 06:15:14 <fabled> huh, worked on my lxc...
2017-10-27 06:16:45 <fabled> smells related to recent perl upgrde
2017-10-27 06:32:41 <fabled> algitbot: retry master
2017-10-27 06:34:58 <fabled> ?!
2017-10-27 06:35:52 <fabled> bld2 just crashed?
2017-10-27 06:42:01 <kaniini> pmOS crew probably hacked it so they could build their stuff
2017-10-27 06:42:03 <kaniini> ;)
2017-10-27 10:26:06 <fabled> working on gdnsd
2017-10-27 10:30:47 <fabled> i think it needs !checkroot
2017-10-27 10:30:52 <fabled> testing
2017-10-27 12:20:29 <rdutra> fabled: gdsnd test is failing in ppc64le when building in a LXC container (I reproduced that in my environment), but when I tried it in a ppc64le VM all tests passed
2017-10-27 12:32:22 <travis-ci> apk-tools:master [passed] |Timo Teräs| solver: consider virtual provides to exclude non-provides transitively | https://travis-ci.org/alpinelinux/apk-tools/builds/293649691
2017-10-27 13:23:01 <travis-ci> apk-tools:master [passed] |Timo Teräs| apk-tools-2.8.1 | https://travis-ci.org/alpinelinux/apk-tools/builds/293670187
2017-10-27 13:27:17 <travis-ci> apk-tools:v2.8.1 [passed] |Timo Teräs| apk-tools-2.8.1 | https://travis-ci.org/alpinelinux/apk-tools/builds/293670193
2017-10-30 10:18:45 <ncopa> did i kill it?
2017-10-30 21:34:34 <przemoc> yay, not everything is red ;)
2017-10-31 14:15:12 <ncopa> the many errors in community is mainly due to missing dependencies from main
2017-10-31 14:15:17 <ncopa> not yet built
2017-10-31 16:25:05 <rdutra> ncopa: ppc64le 3-7 builder failed to build weechat (http://build.alpinelinux.org/buildlogs/build-3-7-ppc64le/main/weechat/weechat-1.9.1-r0.log)
2017-10-31 16:25:05 <rdutra> I also reproduced the same problem in a x86_64 LXC container and I opened a PR: https://github.com/alpinelinux/aports/pull/2618
2017-10-31 17:44:09 <rdutra> algitbot: retry master
2017-10-31 21:06:45 <ncopa> i will try fix those build errors in edge tomorrow