#archlinux32 | Logs for 2018-09-13

Back
[00:02:15] -!- ofara__ has quit [Ping timeout: 272 seconds]
[00:10:09] -!- ofara__ has joined #archlinux32
[00:46:34] -!- ofara__ has quit [Ping timeout: 272 seconds]
[00:48:03] -!- ofara__ has joined #archlinux32
[01:00:27] -!- ofara__ has quit [Ping timeout: 252 seconds]
[01:39:25] -!- woshty has quit [Ping timeout: 252 seconds]
[01:59:17] <elibrokeit> deep42thought, abaumann: just in case Firefox does not excite you enough, they're moving to clang in order to do PGO, and we'll be enabling both.
[02:02:46] -!- elibrokeit has quit [Quit: A random quit message]
[02:02:47] -!- City-busz has quit [Quit: http://quassel-irc.org - Csevegjen kényelmesen. Bárhol.]
[02:02:59] -!- elibrokeit has joined #archlinux32
[02:05:17] -!- ofara__ has joined #archlinux32
[02:19:44] -!- City-busz has joined #archlinux32
[05:38:47] -!- ofara__ has quit [Ping timeout: 240 seconds]
[08:08:35] -!- deep42thought has joined #archlinux32
[08:08:36] <buildmaster> Hi deep42thought!
[08:26:16] * buildmaster resumes sanity.
[08:40:49] -!- abaumann has joined #archlinux32
[08:40:50] <buildmaster> Hi abaumann!
[08:40:58] <deep42thought> Guten Morgen!
[08:41:06] <abaumann> deep42thought: Morns :-)
[08:41:19] <abaumann> Was a little bit busy lately, sorry for the late answers..
[08:41:24] <deep42thought> np
[08:41:34] <deep42thought> We all have a life (I hope).
[08:41:42] <abaumann> about building i486 on a i486 vm, this is completly optional.
[08:41:56] <abaumann> as we have seen: endbr32 runs on a qmeu 'i486'...
[08:42:02] <deep42thought> ok, so running staging-i686-build with i486 repositories enabled should be enough - right?
[08:42:07] <abaumann> so building in a march-i486 chroot should suffice.
[08:42:11] <abaumann> yep.
[08:42:15] <abaumann> exactly.
[08:42:24] <abaumann> has the benefit, it can run on all build slaves.
[08:42:26] <deep42thought> but we'd need the uname hack, wouldn't we?
[08:42:31] <deep42thought> yes
[08:42:41] <abaumann> that's a good point.
[08:42:58] <abaumann> though. gcc also acts as uname and says: "I can do SSE2, please use me" :-)
[08:43:15] <deep42thought> the current idea is, that the build slave tells the master what architecture it is capable to build and the master looks for a suitable build assignment
[08:43:25] <abaumann> ah. good.
[08:43:32] <deep42thought> but we could change that
[08:43:38] <abaumann> because you don't want to clogg up i686 builds with i486 builds maybe
[08:43:39] <deep42thought> (the idea)
[08:43:59] <deep42thought> when all build on the same slaves, this would not be a problem
[08:44:28] <deep42thought> e.g. firefox on i486 might fail, but then it is "failed" and will be scheduled after all not-yet-tried packages are tried
[08:44:47] <abaumann> apropos firefox: it builds on bare metal also for me.
[08:45:06] <abaumann> I suspect systemd-nspawnd to fiddle with the virtual address space
[08:45:34] <deep42thought> we should think about a "huge mem" option (stolen from archlinuxarm) and some slaves that _only_ build these packages in a vm (without systemd) or on bare metal
[08:45:36] <abaumann> the LTO rustc patch helps to reduce presure on RAM and makes rust parts succeed.
[08:45:44] <abaumann> the libxul.so linking issue still happens sometimes.
[08:45:59] <abaumann> yes. same for virtualbox btw.
[08:46:14] <deep42thought> I'm currently testing to build firefox in our archlinux32-vm with 10GB swap
[08:46:16] <abaumann> on the other hand. if it fails on one slave it will eventually end up on a slave which works..
[08:46:32] <deep42thought> what "same" for virtualbox?
[08:46:42] <abaumann> huge mem
[08:46:58] <abaumann> vm.mmap_min_addr=0
[08:47:02] <deep42thought> ah, virtualbox fails to build due to oom?
[08:47:11] <abaumann> this is also an option you want to set with sysctl
[08:47:19] <deep42thought> dammit
[08:47:21] <abaumann> yes. the SOAP service usually fails.
[08:47:22] <deep42thought> I forgot about that
[08:47:30] <abaumann> might work without that.
[08:47:49] <deep42thought> can linux on 32 bit use >4GB swap, actually?
[08:47:54] <abaumann> firefox sometimes builds with /usr/bin/ld (gold) and sometimes with a local gold linker.
[08:48:11] <deep42thought> sounds ... well ... fishy
[08:48:11] <abaumann> I'm not sure, why it picks one sometimes and the other the other times.
[08:48:28] <abaumann> I have to look closer.
[08:48:42] <abaumann> I'm more worried about mesa/qt5 crashes with sddm on testing/staging
[08:48:52] <abaumann> and our good old friend librsvg failing
[08:49:10] <abaumann> but if it is only SSE2, then we don't patch that anymore for i686, right? but in i486.
[08:49:20] <deep42thought> for now: yes
[08:49:25] -!- oaken-source has joined #archlinux32
[08:49:28] <deep42thought> for later: introduce "pentium3"
[08:49:46] <deep42thought> and patch sse2 for i686 but not pentium3
[08:49:57] <abaumann> ah. right.
[08:50:07] <abaumann> I remember the discussions. :-)
[08:50:26] <deep42thought> when the automation really works, another architecture should be painless
[08:50:45] <abaumann> true.
[08:50:59] <abaumann> So, what do I have to do with the packages i486 branch?
[08:50:59] <deep42thought> but we will see, how the buildmaster swallows i486 - which I set up yesterday and which already uncovered a few bugs
[08:51:20] <deep42thought> merge (not "git merge) your modifications into the master branch
[08:51:28] <abaumann> as I understood, you are ready in the buildmaster (minus bugs) :-)
[08:51:33] <abaumann> ah. ok.
[08:51:34] <deep42thought> yes
[08:51:46] <abaumann> and what's the difference between ARCH and CARCH?
[08:51:56] <deep42thought> that was a typo
[08:52:01] <abaumann> ah. good. :-)
[08:52:02] <deep42thought> it's CARCH and has always been
[08:52:22] <abaumann> so, ok, I merge manually. git merges are mis-merging things anyway in this case..
[08:52:29] <deep42thought> (for "pentium3", we will need to patch pacman/makepkg, though)
[08:52:55] <abaumann> another issue: you need a i486 starting point on the mirrors?
[08:53:03] <deep42thought> the only downside for merging i486 and i686 into a single PKGBUILD is, that every change will rebuild _both_ packages
[08:53:06] <abaumann> or do we use eurobuild3 for now, as it is the only mirror containing i486?
[08:53:09] <deep42thought> ah, right
[08:53:39] <deep42thought> my plan was to use a pacman.conf with the official repos first and yours at the end
[08:53:46] <deep42thought> this way, we do not need to bootstrap again
[08:53:53] <abaumann> I still have a manual build with non-CET and gcc 8.2 running, so I finish that one..
[08:54:03] <deep42thought> and as soon as the build slave(s) have rebuilt all i486 packages, they should be on the official mirror
[08:54:04] <abaumann> I would be grateful not to have to bootstrap anymore :-)
[08:54:12] <deep42thought> :-)
[08:54:16] <abaumann> good plan. :-)
[08:54:26] <deep42thought> you will need to bootstrap again as soon as we break some core component ;-)
[08:54:46] <abaumann> well.. yeah.. but that's tough luck..
[08:55:02] <deep42thought> and you might be able to use the archive in that case ...
[08:55:27] <deep42thought> tyzoid: can you include "i486" in the archives, please?
[08:57:08] <abaumann> bootstrapping has another problem: some 20 packages in base/base-devel don't build currently..
[08:57:20] <deep42thought> :-/
[08:57:30] <abaumann> gnulib, glibc 2.28 stuff mainly
[08:57:31] <deep42thought> also on i686?
[08:57:33] -!- ofara has quit [Ping timeout: 245 seconds]
[08:57:34] <abaumann> yes.
[08:57:46] <abaumann> everything with GNU in it. :->
[08:57:50] <abaumann> coreutils...
[09:02:36] <abaumann> return-assignment: Wrong number of arguments (5)
[09:02:37] <abaumann> call either:
[09:02:37] <abaumann> return-assignment ABORT
[09:02:37] <abaumann> return-assignment pkgbase git_revision mod_git_revision repository arch ERROR
[09:02:39] <abaumann> return-assignment pkgbase git_revision mod_git_revision repository arch sub-pkgrel
[09:02:42] <abaumann> unknown return code 5 from 'return-assignment'
[09:02:47] <deep42thought> update your script
[09:02:50] <deep42thought> sry
[09:02:52] <abaumann> though so :-)
[09:03:15] <deep42thought> it is now expected to tell the build-master for which architecture it was building
[09:03:23] <deep42thought> (actually, the buildmaster should know this)
[09:03:25] <abaumann> yep.
[09:07:45] -!- oaken-source has quit [Ping timeout: 252 seconds]
[09:43:41] -!- abaumann has quit [Quit: leaving]
[09:55:59] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T09:53:47.WXqspa.stdin"?.
[10:00:29] * buildmaster resumes sanity.
[10:02:21] -!- abaumann has joined #archlinux32
[10:02:21] <buildmaster> Hi abaumann!
[10:13:35] -!- woshty has joined #archlinux32
[10:17:30] <buildmaster> firefox-developer-edition is broken (says eurobuild3).
[10:31:35] -!- oaken-source has joined #archlinux32
[10:35:21] <abaumann> https://bugs.archlinux32.org swr doesn't build and swrast gets illegal instructions again.. this starts to get a nightmare.
[10:35:23] <phrik> Title:FS#39 : xorg modesetting module fails with illegal instruction (at bugs.archlinux32.org)
[10:39:55] <abaumann> I seriously don't know what's going on with asp. when I check out mesa, I get 18.1.1, built on testing is 18.1.8, new version in trunk is 18.2.0?!
[10:40:18] <deep42thought> there is some intermediate git repository
[10:40:27] <deep42thought> I think, there is a command to let asp update that
[10:40:48] <deep42thought> "asp update"
[10:40:55] <abaumann> ah. asp update mesa
[11:06:18] -!- morealaz has joined #archlinux32
[11:12:56] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T11:10:33.VvozEZ.stdin"?.
[11:20:14] * buildmaster resumes sanity.
[11:20:38] <abaumann> I had modern software. xz uses up all virtual memory and kills effectively the i486 build.
[11:20:52] <deep42thought> xz O.o
[11:21:01] <abaumann> why?
[11:21:09] <abaumann> used for packages, yes.
[11:21:12] <deep42thought> yes
[11:21:21] <deep42thought> I just wonder, why it tries to use as much memory as possible
[11:22:50] <abaumann> it has -a -T flag, for using threads, but the default is 1
[11:22:51] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T11:20:29.hBHMuO.stdin"?.
[11:23:07] <abaumann> guile2.0 seems to be hell of a package to compress.
[11:23:11] <abaumann> * abaumann shrugs
[11:29:08] * buildmaster resumes sanity.
[11:41:43] -!- woshty has quit [Ping timeout: 245 seconds]
[11:44:04] -!- ofara has joined #archlinux32
[11:57:27] -!- morealaz has quit [Remote host closed the connection]
[12:00:33] -!- abaumann has quit [Quit: leaving]
[12:55:44] <Ingar> "To prevent uncomfortable surprises, xz
[12:55:46] <Ingar> has a built-in memory usage limiter, which is disabled by default.
[12:56:16] <Ingar> --memlimit=1GiB
[12:56:26] <Ingar> for some sane value of 1GiB
[14:52:03] -!- woshty has joined #archlinux32
[15:17:58] -!- oaken-source has quit [Ping timeout: 245 seconds]
[15:25:53] -!- abaumann has joined #archlinux32
[15:25:53] <buildmaster> Hi abaumann!
[15:25:59] <deep42thought> Hi abaumann!
[15:26:08] <abaumann> Ingar: thanks, everything below 1 GB is a sane value for xz. :0>
[15:26:15] <abaumann> hi deep42thought
[15:26:26] <deep42thought> I'm currently trying to append you i486 changes to our modifications
[15:26:33] <abaumann> in builder?
[15:26:41] <deep42thought> no, in archlinux32/packages
[15:26:48] <abaumann> uh.
[15:26:50] <deep42thought> but it looks like a lot of them are only temporary
[15:26:53] <abaumann> yes.
[15:26:57] <deep42thought> :-)
[15:27:03] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T15:24:48.u9H0iw.stdin"?.
[15:27:16] <abaumann> most are just temporary cycle breakers, or some packages with really insane dependencies.
[15:27:34] <deep42thought> how about gcc-ada?
[15:27:37] <abaumann> ah.. good I didn't start yet to do the merging too.
[15:27:46] <abaumann> has to be bootstrapped with a i486 version of gcc-ada.
[15:27:54] <abaumann> a binary gnat
[15:28:10] <deep42thought> it looks really hard to only blacklist one part of a split package for only one architecture
[15:28:19] <deep42thought> makepkg is not made for this
[15:28:45] <abaumann> so. it would be better to build gcc as vanilla as possible.
[15:28:51] <deep42thought> yes
[15:28:57] <deep42thought> or package an empty gcc-ada for now
[15:30:32] <abaumann> also the python pip setuptools wheel bootrapping in most python packages is obsolete now, as python 3.7 is bootstrapped on i486.
[15:30:48] <abaumann> where are you current merging, in which packgage?
[15:30:52] <abaumann> *package
[15:31:19] <deep42thought> I'm currently just looking at it and guessing what problems might occur
[15:31:25] <abaumann> ah. good.
[15:31:27] <deep42thought> but the plan was to merge directly into master
[15:31:52] <deep42thought> I wanted to start with gcc and hit the blacklist-only-one-split-package-for-one-architecture problem
[15:31:59] <abaumann> I'm almost thru with the last i486 build. I had some regressions around icu 62
[15:32:37] <abaumann> mmh. A smaller package is simpler for testing. gcc has a lot of nastiness in it..
[15:32:59] <abaumann> on the other hand.. if that one works, so does the rest :)
[15:33:04] <deep42thought> I wanted to try the hard stuff first :-D
[15:33:36] <abaumann> Ok. I'll try what I can do about gnat..
[15:40:53] <deep42thought> abaumann: how about that?: https://ptpb.pw
[15:42:10] <abaumann> ah. that should work.
[15:42:18] <deep42thought> let's see :-)
[15:42:48] <abaumann> the binary gnat works on the i486 vm
[15:42:57] <abaumann> so, it should be possible to build that package.
[15:44:02] <abaumann> oh. a stopid doinstall script with interaction, installing into /usr/gnat.. *grmpf*
[15:44:45] <abaumann> so, I think I install that manually into a chroot on the i486 VM..
[15:45:51] <abaumann> I should maybe move staging stuff from my mirror to stable?
[15:46:11] <abaumann> or do we get into trouble with two staging mirrors?
[15:46:19] <deep42thought> not at all
[15:46:24] <abaumann> ah. then.
[15:46:30] <deep42thought> we need to create a suitable pacman.conf anyway
[15:47:31] * buildmaster resumes sanity.
[15:48:57] <deep42thought> what's the url of your mirror (again)?
[15:49:30] <deep42thought> nvm, found it :-)
[15:50:39] <abaumann> Server = http://archlinux32.andreasbaumann.cc
[15:50:40] <deep42thought> umm, you might want to move your i486 folder out-of-the-way and activate syncing of the i486 folder from the master mirror
[15:51:05] <abaumann> You are faster searching than I'm able to copy-paste with this stupid Lenovo mousepad..-stick-thingy..
[15:51:14] <deep42thought> :-D
[15:52:09] <abaumann> you have a suggestion?
[15:52:24] <deep42thought> i486-bootstrap
[15:52:37] <abaumann> as architecture?
[15:52:40] <deep42thought> yes
[15:52:46] <deep42thought> or bootstrap/i486
[15:52:52] <abaumann> maybe better.
[15:53:21] <abaumann> ok.
[15:53:30] <deep42thought> pacman.conf: https://ptpb.pw
[15:53:31] <abaumann> and this is the last mirror entry then in pacman.d/mirrorlist..
[15:53:37] <abaumann> too fast :-)
[15:53:51] <deep42thought> no, not in mirrorlist
[15:54:05] <abaumann> ah. yeah.
[15:54:18] <abaumann> isn't is a problem to have duplicate repo names?
[15:54:32] <deep42thought> oh, yeah, probably :-D
[15:54:53] <abaumann> so we prefix the later ones with boostrap-?
[15:54:59] <abaumann> *bootstrap-
[15:55:45] <deep42thought> https://ptpb.pw
[15:56:41] <deep42thought> I'm still unsure, how we trick "uname -m" on the build slaves
[15:56:51] <abaumann> ok. I'll put that on the i486 vm and I see, if I can build a simple package this way.
[15:56:58] <deep42thought> thx
[15:59:26] <abaumann> mmh. I have to rename the bootstrap databases
[15:59:40] <deep42thought> why?
[15:59:42] <deep42thought> ah
[15:59:44] <deep42thought> yeah
[15:59:45] <deep42thought> crap
[16:00:01] <abaumann> or simply build new ones.. :-)
[16:00:08] <abaumann> np
[16:02:55] <abaumann> what's in mirrorlist32?
[16:03:05] <deep42thought> any mirror you like
[16:03:13] <deep42thought> well
[16:03:19] <abaumann> I have to have two mirrors, one for bootstrap, one for future i486
[16:03:23] <deep42thought> any mirror that already syncs i486
[16:03:25] <abaumann> I cannot take my own :-)
[16:03:31] <deep42thought> why not?
[16:03:36] <abaumann> Server = http://archlinux32.andreasbaumann.cc
[16:03:36] <abaumann> Server = http://archlinux32.andreasbaumann.cc
[16:03:37] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T16:01:13.kqzB4M.stdin"?.
[16:03:40] <deep42thought> you need to sync i486, too obviously
[16:03:48] <abaumann> ah.
[16:03:59] <abaumann> but does this work? two mirrors?
[16:04:14] <deep42thought> no
[16:04:27] <deep42thought> that's why I didn't include the mirrorlist for the bootstrap repos
[16:04:36] <deep42thought> but simply put a "Server = ..." line there
[16:04:49] <abaumann> ah. in pacman.conf.
[16:04:56] <deep42thought> yes
[16:04:58] <abaumann> right, I have to sync and use the upper one.
[16:05:00] <abaumann> ok.
[16:05:29] <deep42thought> and you need to rename your *.db.tar.gz
[16:05:36] <deep42thought> and *.files.tar.gz
[16:05:45] <abaumann> yep. I'm on it. :-)
[16:05:48] <deep42thought> ok :-)
[16:06:31] <abaumann> huh. i486 already contains packages?
[16:06:40] <abaumann> cool :-)
[16:06:50] <deep42thought> only "any" packages
[16:07:33] <abaumann> tomcat7 and tomcat8 :-)
[16:09:30] -!- woshty has quit [Ping timeout: 252 seconds]
[16:11:07] <deep42thought> the real fix for the missing documentation would be to have arch=(i486 i686) in package_whatever() and arch=(any) in package_whatever-doc()
[16:11:29] <abaumann> true.
[16:12:11] <deep42thought> the build master should support that - one only needs to make the i486 slave not fail on building this (e.g. let it build an empty *-doc package)
[16:12:24] <abaumann> Instead of merging everything from i486 into master, maybe it's easier just to let it roll and see where it fails, and then port back?
[16:12:41] <deep42thought> very true
[16:13:14] <abaumann> though this generates a lots of build errors..
[16:13:23] <deep42thought> and?
[16:13:30] <abaumann> bad statistics :-)
[16:14:12] * buildmaster resumes sanity.
[16:16:41] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T16:14:29.MGpyfZ.stdin"?.
[16:17:02] <deep42thought> man, how many bugs did I introduce today ...
[16:17:17] <abaumann> ..only good bugs :-)
[16:17:51] <deep42thought> I thought, I tested that stupid query, but apparently I didn't ...
[16:19:50] -!- woshty has joined #archlinux32
[16:26:18] <abaumann> ok. renaming through
[16:27:53] <abaumann> ui. kernel i486 changes. that's brave. :-)
[16:28:12] <deep42thought> I want to break some things before I leave in 2 minutes :-D
[16:28:21] <abaumann> good plan :-)
[16:28:32] <deep42thought> has worked a few times already
[16:29:53] <abaumann> don't we need an adapted chroot for i486 bootstrap?
[16:30:13] <deep42thought> well, it needs to use the other repositories
[16:30:15] <deep42thought> that's all
[16:30:24] <deep42thought> and "uname -m" must report i486
[16:30:36] <deep42thought> (so $arch in pacman.conf / mirrorlist works)
[16:31:53] <deep42thought> cu, I hae to leave
[16:31:55] <abaumann> pacman-staging-i486.conf?
[16:31:58] <abaumann> ok.
[16:32:03] <abaumann> I try something.. :-)
[16:32:05] <deep42thought> yeah, sound good
[16:32:12] -!- deep42thought has quit [Quit: Leaving.]
[16:56:20] <abaumann> yep. my manual scripts build with the new scripts and repo layout. So the only thing remaining is to put the i486 VM on srv0.tyzoid.com onto the list of build slaves. :-)
[16:57:23] <abaumann> ls
[16:57:37] <abaumann> * abaumann types in the wrong window :-)
[17:08:34] <abaumann> mmh. ok. we get a small downgrade issue in i486, because I had a build epoch of .2 in pkgver for the bootstrapping process.
[17:32:59] -!- abaumann has quit [Remote host closed the connection]
[18:12:42] -!- dwilc has quit [Ping timeout: 252 seconds]
[18:25:45] -!- dwilc has joined #archlinux32
[18:50:45] -!- oaken-source has joined #archlinux32
[19:27:04] -!- ofara_ has joined #archlinux32
[19:27:22] -!- Cthulu201 has quit [Quit: Nowhere special. I always wanted to go there.]
[19:27:29] -!- ofara_ has quit [Client Quit]
[19:31:06] -!- Cthulu201 has joined #archlinux32
[20:25:03] -!- oaken-source has quit [Ping timeout: 245 seconds]
[20:56:09] -!- ofara has quit [Ping timeout: 252 seconds]
[20:57:44] -!- ofara has joined #archlinux32
[21:07:07] * buildmaster resumes sanity.
[21:09:01] -!- deep42thought has joined #archlinux32
[21:09:01] <buildmaster> Hi deep42thought!
[21:09:36] <deep42thought> abaumann: downgrade should be no problem, I assume, the people actually using i486 are pretty few and pretty used to experimenting ;-)
[21:10:12] <deep42thought> for includeing the i486 vm as build slave, I only need the public ssh key :-)
[21:13:27] -!- abaumann has joined #archlinux32
[21:13:28] <buildmaster> Hi abaumann!
[21:13:48] <abaumann> deep42thought: I was more concerned about the building when packages are suddenly older.
[21:13:51] <deep42thought> good evening, abaumann!
[21:13:55] <abaumann> hi.
[21:14:09] <deep42thought> no problem, either - they get picked by repository from top to bottom
[21:14:19] <deep42thought> e.g. earlier repository in pacman.conf wins
[21:14:46] <abaumann> didn't think about that.. yep.
[21:17:07] <abaumann> I forgot about the i486 fork of devtools32. There is the pacman.conf for i486 and some local hacks on the i486 VM.
[21:17:36] <abaumann> now I checked in.
[21:21:15] <deep42thought> I think, the i486 branch is not really, what we need now
[21:21:24] <deep42thought> but feel free to change that branch to our needs :-)
[21:22:37] <deep42thought> brb, rebooting router
[21:24:07] <abaumann> I'm tyred and cannot type straight anymore. :-)
[21:24:12] <abaumann> so cu
[21:24:14] -!- abaumann has quit [Quit: leaving]
[21:26:51] -!- deep42thought has quit [Ping timeout: 252 seconds]
[21:41:08] <buildmaster> tdlib is broken (says buildknecht2).
[22:00:36] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2018-09-13T21:58:34.FgLGWV.stdin"?.
[22:18:50] * buildmaster resumes sanity.
[23:38:06] -!- isacdaavid has joined #archlinux32