2020-08-01 10:54:50 <PureTryOut[m]> algitbot: retry master
2020-08-01 11:35:44 <maxice8> !algitbot retry master
2020-08-01 13:52:20 <Ariadne> py3-ruamel.std.pathlib needs a distfiles purge
2020-08-01 13:52:37 <Ariadne> they changed the checksum without changing version
2020-08-01 18:03:26 <PureTryOut[m]> algitbot: retry master
2020-08-02 12:33:12 <ikke> algitbot: retry master
2020-08-02 12:37:11 <Cogitri> algitbot: retry master
2020-08-02 12:37:38 <Cogitri> something something cmake is a good buildsystem :P
2020-08-02 12:39:06 <PureTryOut[m]> Literally every other application with CMake works fine, I'd say it's a libphonenumber problem 😉
2020-08-02 12:39:38 <ikke> has it been reported upstream?
2020-08-02 12:41:51 <Cogitri> But it's so easy to mess up things with CMake :/
2020-08-02 12:45:19 <ikke> lua-argparse fails becaus errors include extra information (path/line number)
2020-08-06 08:42:20 <Cogitri> algitbot: retry master
2020-08-06 08:42:38 <Cogitri> weird that it still tries to build electrum on armv7 even though it's disabled
2020-08-06 08:45:16 <ikke> Cogitri: Probably because it's still building an older commit
2020-08-06 08:45:51 <ikke> Cogitri: the builders are set to continue on fail, so it will not fetch new changes until it built everything
2020-08-06 08:48:01 <Cogitri> ikke: Oh, can we force a pull somehow?
2020-08-06 08:49:02 <ikke> Cogitri: just let it continue. It will finish soonish and then it will update
2020-08-06 08:49:12 <Cogitri> Alrighty then, thanks for the info
2020-08-09 12:53:17 <ikke> algitbot: retry master
2020-08-09 16:16:33 <ikke> algitbot: retry master
2020-08-09 17:32:02 <ikke> algitbot: retry master
2020-08-10 04:11:20 <travis-ci> mkinitfs:master [failed] |Filippo Valsorda| initramfs-init: add cmdline consoles to inittab | https://travis-ci.com/FiloSottile/mkinitfs/builds/179064954
2020-08-10 04:28:26 <travis-ci> mkinitfs:master [failed] |Filippo Valsorda| initramfs-init: add cmdline consoles to inittab | https://travis-ci.com/FiloSottile/mkinitfs/builds/179065679
2020-08-10 08:12:47 <ikke> algitbot: retry 3.12-stable
2020-08-10 09:03:08 <ikke> algitbot: retry master
2020-08-10 09:06:23 <ikke> algitbot: retry 3.12-stable
2020-08-10 11:46:04 <Cogitri> algitbot: retry master
2020-08-10 12:02:37 <Cogitri> Ugh, GitHub doesn't provide stable checksums for patches?
2020-08-10 12:15:19 <ikke> no, not for patches
2020-08-10 12:15:54 <ikke> the git version is part of the patch
2020-08-10 12:16:19 <ikke> hmm, no
2020-08-10 12:16:31 <ikke> but the index can change
2020-08-10 12:16:41 <ikke> I think
2020-08-10 12:17:41 <Cogitri> Okie, I just changed it to use a pre downloaded patch as usual
2020-08-10 13:03:24 <nmeum> looks like that ppc64le disabling change also needs to be backported from community/firefox
2020-08-10 13:10:36 <ikke> I don't think firefox on ppc64le would make a lot of sense anyway, right?
2020-08-10 13:14:06 <Cogitri> Why not? 
2020-08-10 13:14:28 <Cogitri> There are ppc64le desktop machines, no? I think someone with a ppc64le machine worked on GNOME on that arch for some time
2020-08-10 13:14:39 <ikke> ah, ok
2020-08-10 13:15:47 <Cogitri> I mean it's not often used as desktop since those workstations are rather pricy and don't deliver as much performance as x86_64 per euro, but it's not like using FF on s390x :)
2020-08-10 13:16:57 <ikke> yeah, makes sense, I was confusing it more with s390x
2020-08-10 13:17:22 <nmeum> disabled it for now anyhow
2020-08-10 13:17:24 <ikke> But mac os of course was for a long time running on powerpc
2020-08-10 13:17:36 <nmeum> what's up with rustc segfaulting occassionally though?
2020-08-10 13:21:05 <Cogitri> Rust likes to trigger edge cases in LLVM apparently
2020-08-10 13:22:21 <Cogitri> Because it has super aggressive optimization (due to it being able to infer lifetimes, types, doing strict aliasing etc.) and CLang is still mainly developed for C(++) apparently
2020-08-10 13:22:55 <Cogitri> But fortunately that got a lot better with LLVM10, I don't think I saw it SIGSEGVing ever since that came along (up until now at least :)
2020-08-10 13:23:29 <Cogitri> But I don't have a ppc64le machine so I can't really debug this/report it upstream with a proper backtrace
2020-08-10 13:23:30 <nmeum> hehe
2020-08-10 16:30:30 <ncopa> algitbot: retry 3.9-stable
2020-08-12 20:32:01 <mps> algitbot: retry master
2020-08-13 07:14:26 <ikke> algitbot: retry master
2020-08-14 06:28:03 <mps> algitbot: retry master
2020-08-14 19:05:54 <ikke> algitbot: retry master
2020-08-14 19:09:35 <ikke> algitbot: retry master
2020-08-14 19:26:24 <ikke> algitbot: retry master
2020-08-15 18:50:21 <ikke> algitbot: retry master
2020-08-15 18:53:28 <ikke> algitbot: retry master
2020-08-16 22:47:45 <andypost> algitbot: retry master
2020-08-16 23:01:20 <andypost> algitbot: retry master
2020-08-17 12:05:26 <ikke> algitbot: retry master
2020-08-18 10:33:05 <ncopa> algitbot: retry master
2020-08-18 15:27:06 <ikke> algitbot: retry master
2020-08-18 20:05:21 <Ariadne> algitbot: retry master
2020-08-19 14:43:04 <ikke> Cogitri: PureTryOut[m] ^ ;-)
2020-08-19 14:43:11 <PureTryOut[m]> Lol wtf
2020-08-19 14:43:12 <PureTryOut[m]> I saw it
2020-08-19 14:43:36 <PureTryOut[m]> I disabled that arch but on the wrong package 🙈
2020-08-19 14:43:48 <ikke> heh
2020-08-19 14:44:01 <PureTryOut[m]> If you can do a direct push, just disable that package on x86
2020-08-19 14:44:12 <PureTryOut[m]> Probably quicker than making a MR
2020-08-19 14:44:25 <ikke> nod, will do
2020-08-19 14:44:37 <ikke> which one did you disable by accident?
2020-08-19 14:44:59 <PureTryOut[m]> maliit-keyboard, which should also be disabled because it depends on libpinyin. But when disabling in my mind I thought that was the only package
2020-08-19 14:45:16 <ikke> aha, ok
2020-08-19 14:46:40 <ikke> done
2020-08-19 14:47:29 <PureTryOut[m]> Thanks
2020-08-19 14:47:32 <Cogitri> Oh, thanks for fixing :) 
2020-08-19 14:47:37 <PureTryOut[m]> Will try to work with upstream to resolve the issue now 😉
2020-08-20 11:11:12 <Ariadne> algitbot: retry master
2020-08-21 09:31:35 <Ariadne> algitbot: retry master
2020-08-21 11:02:55 <ikke> keeps segfaulting
2020-08-21 11:56:26 <Ariadne> yes
2020-08-21 11:56:37 <Ariadne> i "fixed" it
2020-08-21 11:56:44 <Ariadne> and by "fixed" i mean, i committed serious crimes
2020-08-21 11:57:23 <ikke> I was following your monologue :)
2020-08-21 13:43:02 <Ariadne> algitbot: retry master
2020-08-21 13:43:25 <Ariadne> crimes complete
2020-08-21 13:46:39 <Ariadne> that's related to crimes, phase 2
2020-08-21 13:46:59 <Ariadne> i'm now rebuilding main/gcc *again* with the second main/gcc
2020-08-21 13:57:52 <Ariadne> algitbot: retry master
2020-08-23 17:02:50 <Cogitri> Ariadne,PureTryOut[m]: Seems like GCC ICEs on akonadi on mips
2020-08-23 17:03:35 <Ariadne> just disable it there
2020-08-23 17:12:58 <Cogitri> Will probably mean to disable the entirety of KDE on mips
2020-08-23 19:15:48 <mps> algitbot: retry master
2020-08-24 07:25:35 <Ariadne> Cogitri: yes but only temporary.  gcc devs already fixing the bug.
2020-08-24 07:26:39 <Cogitri> Ah neat 
2020-08-24 07:48:12 <Ariadne> mps: the linux-edge is due to gcc 10 regression
2020-08-24 07:48:29 <Ariadne> I'm working on an updated gcc 10 package to fix it.
2020-08-24 07:56:22 <mps> Ariadne: this is triggered by TRIM_UNUSED_KSYMS=y I think, but it works with gcc 9.x
2020-08-24 07:56:40 <Ariadne> no.  it is a vectorization bug
2020-08-24 07:56:52 <Ariadne> I've already cherry picked the fix
2020-08-24 07:57:35 <mps> aha, ok
2020-08-24 07:57:53 <mps> I didn't yet found time to look more at it
2020-08-24 07:58:34 <mps> and also have this for custom build kernel on armv7 with gcc 10.2: `crypto/aegis128-neon-inner.c:151:3: error: incompatible types when initializing type 'unsigned char' using type 'uint8x16_t'`
2020-08-24 07:58:54 <mps> 151 |   k ^ vld1q_u8(const0),
2020-08-24 07:59:15 <mps> but will try again when you push fixed gcc
2020-08-24 08:07:25 <Ariadne> yes, that is due to vectorization :)
2020-08-24 08:07:35 <Ariadne> the RHS is a vector type
2020-08-24 08:07:40 <Ariadne> k is unsigned char
2020-08-24 08:11:35 <Ariadne> algitbot: retry master
2020-08-24 10:22:18 <ikke> Ariadne: looks like you missed a mkdir -p "$subpkdir" for busybox
2020-08-24 10:22:35 <Ariadne> oops :)
2020-08-24 10:22:38 <ikke> right, s/pkgdir/subpkgdir/
2020-08-24 10:23:44 <Ariadne> hooray, openrc fails to build with -fno-common
2020-08-24 10:24:09 <ikke> What does -fno-common do? 
2020-08-24 10:26:01 <Ariadne> it is default for gcc 10
2020-08-24 10:26:07 <Ariadne> it disallows you from declaring the same variable twice
2020-08-24 10:27:46 <ikke> aha, that explains the error message then :)
2020-08-24 10:28:42 <ikke> https://wiki.gentoo.org/wiki/Gcc_10_porting_notes/fno_common
2020-08-24 10:29:25 <Ariadne> according to gentoo, there is now openrc 0.41.2
2020-08-24 10:29:32 <Ariadne> of course, this is nowhere on openrc github
2020-08-24 10:29:49 <Ariadne> just going to use this as an additional datapoint that openrc might be in a unhealthy state :)
2020-08-24 10:30:31 <ikke> Do we have a proper alternative already?
2020-08-24 10:30:48 <ikke> From the discussions I'e seen, not yet
2020-08-24 10:31:47 <Ariadne> not yet, no.
2020-08-24 10:31:56 <Ariadne> algitbot: retry master
2020-08-24 10:32:55 <Ariadne> it has been low priority, because writing a service manager (especially one that is backwards compatible with openrc) is a pain in the ass :)
2020-08-24 10:33:17 <Ariadne> and, we would want to replace busybox init and such too
2020-08-24 10:35:58 <mps> 'we'?
2020-08-24 10:37:23 <Ariadne> mps: there is a desire to reduce busybox where possible, to make container size smaller
2020-08-24 10:37:44 <Ariadne> also busybox init is largely overhead when using a proper service manager
2020-08-24 10:37:53 <Ariadne> even openrc has its own init now, we just do not use it
2020-08-24 10:38:07 <mps> yes, that's true
2020-08-24 10:39:10 <mps> but I'm not sure switch will go without problems. So this should be done carefully and maybe in few steps
2020-08-24 10:39:26 <mps> and, thank you for explanation
2020-08-24 10:42:00 <Ariadne> yes, incremental steps :)
2020-08-24 10:44:47 <Ariadne> but basically, where we want to wind up is, a simple service manager that uses process supervision, supports systemd unit files (so we can make use of unit files included in packages), supports openrc services for backwards compatibility
2020-08-24 10:58:05 <Ariadne> ikke: is the ppc64le builder stuck
2020-08-24 10:58:39 <ikke> checking
2020-08-24 10:59:16 <ikke> Ariadne: no, seems idle
2020-08-24 10:59:23 <Ariadne> weird
2020-08-24 10:59:33 <Ariadne> the build page says it is on busybox
2020-08-24 11:17:55 <ikke> algitbot: retry master
2020-08-24 11:20:57 <ikke> hmm? "building v20200626-2820-g0c4a13be77"
2020-08-25 12:40:52 <fabled> bah
2020-08-25 12:41:48 <ikke> zlib module missing
2020-08-25 12:42:41 <ikke> scdoc ;-)
2020-08-25 13:59:50 <Ariadne> HECK YEAH
2020-08-25 13:59:54 <Ariadne> REBASE TIME
2020-08-25 14:00:10 <ikke> heh
2020-08-25 14:00:45 <fabled> cool. algitbot is smart!
2020-08-25 14:02:57 <PureTryOut[m]> 🎉
2020-08-27 12:57:04 <Cogitri> Huh, the builders are trying to mock me I think, everything works on CI and locally and then blows up on the builders :D 
2020-08-27 12:57:13 <Cogitri> algitbot: retry master
2020-08-27 12:57:27 <Cogitri> Let's just make sure that the libreoffice build failure isn't on some star being misaligned or something
2020-08-27 13:00:42 <Cogitri> wat
2020-08-27 13:00:55 <Cogitri> ikke: Can you fetch me the config.log of that maybe? 😅
2020-08-27 13:01:17 <Cogitri> The builders are very weird since the GCC10 upgrade, both LO and LDC work just fine on CI and locally
2020-08-27 13:01:20 <ikke> Later today 
2020-08-27 13:01:39 <Cogitri> Okie 👍
2020-08-27 13:39:38 <mps> Cogitri: last night I noticed that the Mars, Saturn and Jupiter are somewhat aligned on the sky ;)
2020-08-27 13:43:19 <Cogitri> Heh, gotta be cosmic rays :D
2020-08-27 15:10:21 <ikke> Cogitri: for what do you need the output? libreoffice or ldc?
2020-08-27 15:11:27 <Cogitri> For libreoffice please
2020-08-27 15:13:22 <ikke> Cogitri: https://tpaste.us/5ELj
2020-08-27 15:13:28 <ikke> This is from aarch64
2020-08-27 15:14:14 <Cogitri> Thanks 👍
2020-08-27 15:15:02 <Cogitri>  /usr/bin/ld: /usr/lib/libboost_date_time.so: undefined reference to `__cxa_free_exception'
2020-08-27 15:15:13 <Cogitri> What, how does that only happen on the builders
2020-08-27 15:16:41 <ikke> no clue
2020-08-28 06:49:29 <nmeum> weird, dpkg passed on the CI on s390x and aarch64
2020-08-28 06:49:49 <nmeum> looks like a permission mismatch in the tar output
2020-08-28 06:49:50 <nmeum> > got: '. mode=42755
2020-08-28 06:49:54 <nmeum> > expected: '. mode=40755
2020-08-28 06:51:43 <nmeum> I guess S_ISGID is set in the directory where the test are exexcuted?
2020-08-28 06:56:17 <ikke> nmeum: I see some directories in the source with the s bit set
2020-08-28 06:56:27 <ikke> https://tpaste.us/QnO4
2020-08-28 06:58:06 <nmeum> hm, but that wouldn't explain why it passes locally and on the ci
2020-08-28 06:58:18 <ikke> /home/buildozer has it set
2020-08-28 06:58:28 <ikke> So probably inheritted
2020-08-28 06:58:34 <ikke> Which is not the case on CI
2020-08-28 06:58:41 <nmeum> ah, that could explain it
2020-08-28 06:59:02 <nmeum> but how would I go about fixing this then? :D
2020-08-28 06:59:11 <ikke> Very good question
2020-08-28 06:59:26 <ikke> Should the tests be fixed somehow?
2020-08-28 07:02:09 <ikke> They should ideally not be affected by outside factors like these
2020-08-28 07:03:12 <nmeum> it's a tar extraction test
2020-08-28 07:03:21 <nmeum> I guess we colud untar the tarball in a different directory?
2020-08-28 07:03:32 <nmeum> e.g. in /tmp/<something>
2020-08-28 07:04:03 <nmeum> the mytmpdir perl variable (line 35) in ./lib/dpkg/t/t-tarextract.t seems to specify the directory used
2020-08-28 07:04:31 <ikke> Or remove the S_ISGID on $BUILDDIR?
2020-08-28 07:04:48 <nmeum> could you change that to line to: my $tmpdir = '/tmp/t-tarextract' and see if that helps?
2020-08-28 07:05:23 <nmeum> not sure what implications removing the S_ISGID bit would have, I don't know why it was set in the first place
2020-08-28 07:05:26 <nmeum> anyway
2020-08-28 07:05:40 <nmeum> I have to focus on dayjob stuff now :/
2020-08-28 07:06:01 <nmeum> if it get's to annoying just disable the test
2020-08-28 07:06:08 <nmeum> *gets
2020-08-28 07:06:15 <nmeum> *tests
2020-08-28 07:15:58 <nmeum> ikke: I reproduced this locally by setting the S_ISGID bit on my home, changing the variable outlined above fixes the extract test but there is one more test which needs a similar fix
2020-08-28 07:37:00 <nmeum> ikke: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/11819 this fixes it for me :D
2020-08-28 07:42:38 <ikke> nmeum: looks good :)
2020-08-28 07:53:42 <Cogitri> Right, seems like ff-esr needs some additional patches to do without dbus
2020-08-28 07:56:03 <nmeum> https://github.com/kisslinux/repo/blob/master/extra/firefox-esr/patches/no-dbus.patch ?
2020-08-28 07:56:18 <Cogitri> Yup
2020-08-28 07:56:34 <Cogitri> 
ACTION should really not push things before sleep :D

2020-08-28 07:58:03 <nmeum> hehe
2020-08-28 08:10:51 <mps> Cogitri: yes
2020-08-28 10:08:51 <nmeum> oh boy, more firefox fun
2020-08-28 10:10:25 <nmeum> looks like libgkrust needs to be patched as well?
2020-08-28 10:20:07 <Cogitri> Oh, the patch from KISS Linux doesn't do that? 🤔
2020-08-28 10:29:47 <nmeum> doesn't look like it there are a bunch of linking errors due to undefined dbus symbols used in libgkrust
2020-08-28 10:30:07 <nmeum> maybe that could also be fixed using some linker? idk how kiss solves that
2020-08-28 10:30:12 <nmeum> *linker flag
2020-08-28 10:53:32 <Cogitri> I doubt a linker flag would magically makes those undefined errors go awah
2020-08-28 10:58:12 <nmeum> probably, just guessing why it works on kiss but doesn't work for us
2020-08-28 12:54:43 <mps> algitbot: retry master
2020-08-28 14:28:56 <mps> algitbot: retry master
2020-08-29 07:55:27 <_ikke_> algitbot: retry master
2020-08-29 07:55:39 <_ikke_> algitbot: retry 3.12-stable
2020-08-29 07:59:11 <_ikke_> Is that a -fno-common issue with xorg-server?
2020-08-29 08:30:25 <_ikke_> Hmm, only happens on mips
2020-08-29 18:36:02 <ikke> algitbot: retry master
2020-08-29 18:44:38 <ikke> algitbot: retry master
2020-08-29 21:58:35 <afontain_> huh, it got merged?
2020-08-30 12:55:10 <ikke> algitbot: retry master
2020-08-30 14:11:30 <Cogitri> Hm, I think the x86_64 is getting some download mirror that's currently down, downloading works just fine here
2020-08-30 14:22:09 <ikke> Not sure if it's related, but CenturyLink has large network issues. 
2020-08-30 14:23:46 <mps> that explains why I can't download xorg-server to build locally for 3.12