#archlinux32 | Logs for 2018-05-17

[00:05:59] -!- AndrevS has quit [Remote host closed the connection]
[00:08:35] <eschwartz> on the topic of fun, https://bugs.archlinux.org
[00:08:36] <phrik> Title:FS#40713 : [python-pip] pip does not use /etc/ssl/certs/ca-certificates.crt (at bugs.archlinux.org)
[00:08:45] <eschwartz> I fixed it!
[00:09:03] <eschwartz> ish. I don't have push access to extra
[00:14:30] <buildmaster> haskell-wai-extra is broken (says buildknecht).
[00:23:42] <buildmaster> haskell-wai-logger is broken (says buildknecht).
[00:29:09] -!- City-busz has joined #archlinux32
[00:41:33] -!- City-busz has quit [Quit: http://quassel-irc.org - Csevegjen knyelmesen. Brhol.]
[00:42:14] -!- City-busz has joined #archlinux32
[00:44:02] <deep42thought> eschwartz: even with 5713cd629c97c7a12a600a1dd73ad81d87374eb1 I still get "sudo: PAM account management error: Authentication service cannot retrieve authentication info" on one box
[00:55:54] <eschwartz> this is exceedingly odd, are you sure you used the right thing?
[00:56:16] <eschwartz> what does the shadow file in the build chroot look like
[00:56:29] <deep42thought> ah, I should try -c, at some time
[00:56:32] <deep42thought> gimme a sec
[00:58:03] <deep42thought> "builduser" is missing from that file
[00:58:32] <deep42thought> !ptpb
[00:58:32] <phrik> “<command> 2>&1 | curl -F c=@- https://ptpb.pw” OR “curl -F c=@path/to/a/file https://ptpb.pw” OR install community/pbpst
[00:58:44] <deep42thought> https://ptpb.pw
[00:59:17] <deep42thought> let me try devtools with -c
[00:59:40] <eschwartz> What does /usr/bin/makechrootpkg actually say
[00:59:56] <eschwartz> And do you have a different one in your path anyway
[01:00:12] <eschwartz> You might not be running what you think you are :)
[01:00:30] <deep42thought> damn
[01:00:38] <deep42thought> nooooo
[01:00:49] <deep42thought> $ which makechrootpkg
[01:00:49] <deep42thought> /usr/local/bin/makechrootpkg
[01:00:54] * deep42thought is a fool
[01:01:46] <deep42thought> thanks!
[01:02:13] <eschwartz> I've done the same thing. ;) Why do you think I caught it?
[01:02:20] <deep42thought> ... resolution: I ran "sudo make install" inside devtools' source folder some time in the past to quickly deploy tried changes
[01:03:58] <deep42thought> "sudo: staging-i686-build: command not found" - this one I can solve by my own :-)
[01:05:53] <buildmaster> firefox is broken (says rechenknecht).
[01:08:13] <deep42thought> ok, I'm off for bed
[01:08:16] <deep42thought> cu tomorrow
[01:08:21] -!- deep42thought has quit [Quit: Leaving.]
[03:47:19] -!- oaken-source has quit [Ping timeout: 256 seconds]
[04:21:05] <tyzoid> deep42thought: will do (re. vagrant)
[06:40:12] -!- deep42thought has joined #archlinux32
[06:40:12] <buildmaster> Hi deep42thought!
[06:40:21] <deep42thought> Hi buildmaster, what's up?
[06:40:21] <buildmaster> up? I'm up for 19 hours, 30 minutes, load average: 0.51, 0.83, 1.51
[06:52:14] <tyzoid> hey deep42thought: new vagrant box is uploading right now
[06:52:21] <deep42thought> tyzoid: Thanks!
[06:53:13] <deep42thought> there was someone upperstream trying to debug i686 bugs and I thought, the easiest way for them would be to use our vagrant file
[07:12:27] <eschwartz> So file seccomp issues seem to have broken us too
[07:17:45] <tyzoid> deep42thought: Let me know if that vagrant box works for you
[07:30:10] <eschwartz> So you could remove your own core/file overrides now, I guess
[07:52:03] <deep42thought> eschwartz: we need to wait with removal of our patch until yours hits [core]
[08:12:44] -!- deep42thought has quit [Quit: Leaving.]
[09:38:53] -!- deep42thought has joined #archlinux32
[09:38:54] <buildmaster> Hi deep42thought!
[10:02:53] -!- abaumann has joined #archlinux32
[10:02:54] <buildmaster> Hi abaumann!
[10:03:02] <deep42thought> good morning
[10:03:06] <abaumann> Hi deep42thought.
[10:03:18] <abaumann> Busy nights and days for you. :-)
[10:04:02] <deep42thought> umm
[10:04:05] <deep42thought> yeah :-/
[10:04:22] <deep42thought> I brewed my big cup of coffee this morning O.O
[10:05:22] <abaumann> The "causa glibc": I would really like to build a debug version of the toolchain with all symbols attached in order to produce a nice stacktrace of the segfault in glibc.
[10:05:34] <abaumann> This increases the probability it gets fixed. :-)
[10:05:46] <deep42thought> sounds reasonable
[10:06:39] <deep42thought> I *think* you can put a modified PKGBUILD in build-support/gcc-debug/ and make it "provide=gcc"
[10:06:49] <deep42thought> this way, it would naturally enter our repositories
[10:06:54] <deep42thought> (I hope so)
[10:07:04] <abaumann> mmh. I can also build a local version on a VM.
[10:07:15] <deep42thought> yeah, works too
[10:07:32] <deep42thought> but this means, we/I need to inject it manually into our mirror(s)
[10:07:41] <abaumann> I don't like to disturb the build process more than necessary, otherwise the buildmaster gets angry at me :-)
[10:07:53] <deep42thought> :-D
[10:07:54] <deep42thought> ok
[10:07:57] <deep42thought> fair point
[10:07:57] <abaumann> I just want to get a backtrace basically.
[10:10:35] <deep42thought> I was just thinking, that it might be worth the effort to have a gcc/glibc/binutils regularly updated in build-support - and for that, separate PKGBUILDs would be ideal
[10:11:12] <buildmaster> tensorflow is broken (says buildknecht).
[10:12:40] <abaumann> actually. good idea.
[10:14:21] <abaumann> I really like redhat's approach of -debuginfo packages, then you can just install the packages when you hit a problem.
[10:16:33] <deep42thought> archlinux has -debug packages, too
[10:16:41] <deep42thought> but I think, the buildmaster ignores them :-D
[10:17:36] -!- eduardoeae has quit [Ping timeout: 255 seconds]
[10:19:00] <abaumann> ah.
[10:33:56] <deep42thought> abaumann: see "man PKGBUILD" -> "debug"
[10:35:58] <deep42thought> but OTOH: https://github.com
[10:36:00] <phrik> Title:builder/build-packages at master · archlinux32/builder · GitHub (at github.com)
[10:39:33] <abaumann> oh. :-)
[10:41:43] <abaumann> systemd: why can three virtual machines with different macs get the same IP?
[10:41:53] <abaumann> /etc/machine-id?
[10:42:36] <buildmaster> nvidia is broken (says buildknecht2).
[10:42:48] <abaumann> systemd-machine-id-setup is your friend after cloning :-)
[10:44:17] <abaumann> even then they get the same.. I wonder, what "random source" is used to generate the machine-id?!
[10:45:19] <abaumann> ah. the /etc/machine-id file must be deleted before calling systemd-machine-id-setup..
[10:45:48] <deep42thought> who's assigning the ips?
[10:45:59] <abaumann> systemd-networkd
[10:48:25] <abaumann> I don't know how systemd-networkd manages this, but last time I had a look at DHCP, a unique MAC lead to a unique IPv4. Obviously if /etc/machine-id is identical after cloning, the VMs get the same IP.
[10:48:34] <abaumann> strange new systemd-world. :-)
[10:52:02] <abaumann> mmh, :: x264: requires libavcodec.so=57-32
[10:53:08] <abaumann> mplayer: error while loading shared libraries: libswscale.so.4: cannot open shared object file: No such file or directory
[10:53:34] <abaumann> but only on one machine. *puzzle*
[10:56:44] -!- Polichronucci has quit [Ping timeout: 268 seconds]
[10:56:46] <abaumann> ah. testing is the problem. so maybe mplayer needs a recompile as ffpeg and x264 had an ABI bump.
[10:57:02] <abaumann> *ffmpeg
[10:57:40] -!- Polichronucci has joined #archlinux32
[11:06:10] <abaumann> ah: mplayer is broken: cpudetect.c:60:63: error: parameter 2 ('sc') has incomplete type
[11:06:13] <abaumann> static void sigill_handler_sse( int signal, struct sigcontext sc )
[11:06:15] <abaumann> ^~
[11:21:12] <buildmaster> python-scipy is broken (says rechenknecht).
[11:35:52] <buildmaster> python-eventlet is broken (says buildknecht).
[12:02:15] -!- eduardoeae has joined #archlinux32
[12:03:00] -!- oaken-source has joined #archlinux32
[12:11:35] -!- eduardoeae has quit [Ping timeout: 240 seconds]
[12:15:20] -!- eduardoeae has joined #archlinux32
[12:34:06] -!- deep42thought has parted #archlinux32
[12:34:12] -!- deep42thought has joined #archlinux32
[12:35:09] <buildmaster> Hi deep42thought!
[13:37:13] -!- deep42thought has quit [Remote host closed the connection]
[14:35:46] -!- oaken-source has quit [Ping timeout: 264 seconds]
[14:45:40] -!- buildmaster has quit [Ping timeout: 256 seconds]
[14:48:13] -!- buildmaster has joined #archlinux32
[14:52:31] -!- abaumann has quit [Quit: leaving]
[14:55:18] <bill-auger> it looks like arch added a package 'ir-lv2' - parabola has this package already so we need to remove it - but we may need to keep building it for i686 - does anyonw know if srchlinux32 willtake this package anytim soon?
[15:06:05] <tyzoid> bill-auger: If it builds with i686, and Arch mainline has it, we'll start building it at some point
[15:07:43] -!- yans has quit [Ping timeout: 268 seconds]
[15:11:53] -!- abaumann has joined #archlinux32
[15:11:53] <buildmaster> Hi abaumann!
[15:12:55] <tyzoid> hey abaumann
[15:13:31] <tyzoid> Are you able to add new packages to the repo?
[15:13:43] <tyzoid> and/or can you look at ir-lv2?
[15:14:26] <abaumann> currently I get a buildmaster.archlinux32.org port 22: Network is unreachable
[15:14:44] <abaumann> ah. there again.
[15:14:45] <abaumann> mmh.
[15:15:03] <abaumann> ir-lv2, lemme check. No I cannot add them, but I can have a look where it is currently..
[15:16:58] <abaumann> Upstream I see only one in AUR.
[15:17:27] <tyzoid> abaumann: interesting. bill-auger: Where are you seeing this package?
[15:23:36] <abaumann> tyzoid: also nothing in the git mirror, maybe it has to appear there first..
[15:24:11] <tyzoid> abaumann: Did you see bill's message in the logs from right before you hopped on?
[15:24:18] <abaumann> yep. :-)
[15:24:34] -!- yans has joined #archlinux32
[15:25:06] <abaumann> https://bbs.archlinux32.org
[15:25:08] <phrik> Title:linux-pae / Kernel & Hardware / Arch Linux 32 Forums (at bbs.archlinux32.org)
[15:25:17] <abaumann> actually, is there a place we keep packages like this?
[15:25:25] <abaumann> I have my own repo, but officially?
[15:25:42] <abaumann> Sadly I have no 32-bit server with more than 4 GB RAM to test PAE.
[15:25:52] <tyzoid> In general, If arch mainline doesn't have it, it usually makes little sense for us to add.
[15:26:10] <tyzoid> abaumann: I can spin you up a kvm system, if you want
[15:26:53] <tyzoid> if you think that might work
[15:27:22] <abaumann> Well, it's really just a rebuild of the kernel with another option.
[15:27:48] <abaumann> What I was thinking, can we put PAE as default, or can it be enabled/disabled via kernel command line?
[15:28:43] <abaumann> ah. no. PAE is quite on or off-ish.
[15:29:26] <tyzoid> well, I mean for testing, if a qemu/kvm system will work for that.
[15:29:40] <abaumann> Ah. I can do that on my 8 GB 64-bit machine, no problems.
[15:29:51] <abaumann> I meant, I don't have real hardware for testing this.
[15:30:17] <abaumann> I think, non-PAE is the safer option for the default kernel (most people run it on a laptop or system with less than 4 GB of RAM).
[15:31:15] <abaumann> I meant: do we have something official like: http://archlinux32.andreasbaumann.cc
[15:31:17] <phrik> Title:Index of /aur/i686 (at archlinux32.andreasbaumann.cc)
[15:31:31] <abaumann> I could put a kernel-pae there..
[15:31:41] <abaumann> I would also not like to build it per default..
[15:31:54] <tyzoid> Nothing official like that
[15:32:08] <tyzoid> I mean, it'd basically be the same as if a TU on arch made a custom repo
[15:32:18] <abaumann> right.
[15:33:13] <abaumann> but TUs have signing keys. Archlinux32 users usually not.
[15:33:22] <bill-auger> the package data is here https://www.archlinux.org
[15:33:24] <phrik> Title:Arch Linux - ir.lv2 1.3.3-2 (x86_64) (at www.archlinux.org)
[15:33:35] -!- buildmaster has quit [Ping timeout: 256 seconds]
[15:33:55] <abaumann> ah. a dot.
[15:34:12] <bill-auger> does archlinux32 normally take everything that arch packages? or do you sometimes filter stuff?
[15:35:08] <abaumann> we take everything, but things listed in https://github.com
[15:35:10] <phrik> Title:packages/blacklist at master · archlinux32/packages · GitHub (at github.com)
[15:35:20] <tyzoid> bill-auger: The only things we filter are things that won't build / won't work once built
[15:35:32] <abaumann> exactly.
[15:36:26] <tyzoid> bill-auger: looks like we have it in our repos already? https://packages.archlinux32.org
[15:36:27] <phrik> Title:Arch Linux 32 - ir.lv2 1.3.3-2.0 (i686) (at packages.archlinux32.org)
[15:36:53] <abaumann> the one with a dot. is ir.lv2 and ir-lv2 (AUR) the same?
[15:37:13] <tyzoid> idk, bill's package reference has the dot, which is what we're also showing
[15:37:36] <tyzoid> the description looks similar
[15:37:41] <abaumann> yeah.
[15:39:07] <bill-auger> why donmt i see it ? https://packages.archlinux32.org
[15:39:08] <phrik> Title:Arch Linux 32 - Package Search (at packages.archlinux32.org)
[15:39:40] <abaumann> because it's: https://packages.archlinux32.org
[15:39:41] <bill-auger> the dot not a dash
[15:39:41] <phrik> Title:Arch Linux 32 - Package Search (at packages.archlinux32.org)
[15:39:48] <bill-auger> gee whiz i missed that too
[15:39:58] <abaumann> :-)
[15:40:16] <abaumann> it's not the most mnemonic name..
[15:40:33] <bill-auger> who ever though of putting a dot in an identier - that should be illegal
[15:40:38] <bill-auger> or sumfin
[15:40:50] <abaumann> it will run havock somewhere, I'm sure. ;-)
[15:43:33] <bill-auger> i just read the policy - names can have any of @, ., _, +, -.
[16:30:12] -!- buildmaster has joined #archlinux32
[16:31:48] -!- buildmaster has quit [Remote host closed the connection]
[16:57:20] <eschwartz> [09:33:09 AM] <abaumann> but TUs have signing keys. Archlinux32 users usually not.
[16:57:37] <eschwartz> but you're not a user, you're an arch32 dev...
[16:57:53] <eschwartz> most archlinux users don't have signing keys either
[16:58:45] <eschwartz> bill-auger: why should it be illegal
[16:59:07] <eschwartz> BTW pacman will accept the weirdest nonsense, but makepkg will refuse to *create* it
[16:59:23] <eschwartz> pretty sure newlines are valid as far as pacman itself cares
[17:04:39] -!- yans has quit [Remote host closed the connection]
[17:06:41] <bill-auger> period chars in package names
[17:09:48] <eschwartz> hehe https://wiki.archlinux.org
[17:09:50] <phrik> Title:User:Apg - ArchWiki (at wiki.archlinux.org)
[17:10:01] <eschwartz> so newlines are invalid, but they're one of the few things
[17:10:47] <eschwartz> you can use ":=<>"but it'll break things, like dependencies
[17:13:18] <eschwartz> pkgver can contain "-" even (which breaks *hard* when {foo-1}-{1.0} clashes with {foo}-{1-1.0})
[17:13:25] <abaumann> eschwartz: my problem was more: where would we put a 'linux-pae' package and how to tell the buildmaster that he has a depending package (it should built when 'linux' is built).
[17:13:37] <abaumann> I don't like NOT to automatize things :-)
[18:07:56] -!- abaumann has quit [Quit: leaving]
[18:13:40] -!- Asiff has joined #archlinux32
[18:13:55] <Asiff> hello
[18:14:10] <eschwartz> hello
[18:14:41] -!- erkin has joined #archlinux32
[18:14:52] -!- erkin has parted #archlinux32
[18:26:05] <tyzoid> Asiff: Did you need something?
[18:31:54] <Asiff> yope
[18:32:18] <Asiff> can i use archlinux32 for my server and hope to not run into problem?
[18:32:44] <tyzoid> Asiff: I recommending not running servers on hopes and dreams
[18:33:11] <tyzoid> That said, you can use archlinux32 (or any arch system) on servers - just note that there is a higher chance for breakages
[18:33:46] <eschwartz> https://wiki.archlinux.org
[18:33:48] <phrik> Title:User:Seblu - ArchWiki (at wiki.archlinux.org)
[18:34:13] <eschwartz> oh.. I made a factoid, right
[18:34:16] <eschwartz> !server
[18:34:17] <phrik> Should I run Archlinux on my server? -- https://wiki.archlinux.org
[18:34:28] <tyzoid> one eschwartz: Well, keep in mind, we're talking AL32 here, not AL
[18:34:50] <tyzoid> breakages are a bit more common, by virtue of having such a small team.
[18:34:51] <eschwartz> It should not be substantially different with regard to AL32, except that there's less testing and more automated toolchains
[18:35:09] <eschwartz> Which as tyzoid says, has its own unique risks
[18:35:53] <tyzoid> Asiff: Anyway, if you're goal is to run an arch system as a server without problems, best to run it as a container / vm on another host, like centos, or debian (I use debian for this)
[18:36:12] <tyzoid> s/without problems/without as much risk of problems/
[18:36:47] <tyzoid> That said, it'll function just fine as a server, there's no technical limitations in the software
[18:38:07] -!- abaumann has joined #archlinux32
[18:38:19] <tyzoid> wb abaumann
[18:38:30] <tyzoid> Buildmaster is down, so didn't send you a message
[18:38:41] <abaumann> ah.
[18:39:02] <tyzoid> abaumann: see https://status.archlinux32.org
[18:39:16] <tyzoid> I can add you to the email alert list for that system, if you wan
[18:39:33] <abaumann> tyzoid: yeah. that's a good idea.
[18:40:17] <abaumann> Asiff: what I do on servers I have test versions around and I always do the upgrade first on test machines.
[18:40:26] <abaumann> This is good practice anyway. :-)
[18:40:26] <tyzoid> ^
[18:41:14] <tyzoid> abaumann: Check your email, you need to accept the invitation to receive alert mail
[18:41:16] <abaumann> Where is buildmaster.archlinux32.org? It's quite unstable..
[18:41:20] <abaumann> ok.
[18:41:27] <tyzoid> iirc it's on rewby's stack
[18:42:48] <abaumann> email? to @archlinux32 or my personal one?
[18:43:07] <tyzoid> abaumann: my last buildmaster backup is 12h:30m ago, so if we ever need, we can work on restoring that elsewhere.
[18:43:12] <tyzoid> yeah, sent to archlinux32
[18:43:20] <tyzoid> I can use a different one if you want
[18:43:44] <abaumann> no, it's ok. but I didn't receive anything so far.
[18:44:15] <tyzoid> huh
[18:44:26] <tyzoid> oh, right
[18:44:27] <tyzoid> greylisting
[18:44:31] <tyzoid> it'll be a few mins
[18:44:42] <abaumann> easy :-)
[18:47:10] <tyzoid> abaumann: I thought you were using mutt?
[18:47:43] <abaumann> well :-)
[18:50:49] <tyzoid> weird that I haven't seen it yet
[18:51:59] <tyzoid> I could use the internal delivery address, but that's cheating
[18:56:27] <tyzoid> abaumann: looks like it just got delivered, check now
[19:39:36] <abaumann> thanks. :-)
[19:45:25] -!- oaken-source has joined #archlinux32
[21:03:37] <tyzoid> abaumann: added to the email alert list
[21:15:22] -!- eduardoeae has quit [Ping timeout: 264 seconds]
[21:16:31] -!- eduardoeae has joined #archlinux32
[21:25:59] -!- oaken-source has quit [Ping timeout: 276 seconds]
[21:36:01] -!- abaumann has quit [Quit: leaving]
[22:40:55] -!- yans has joined #archlinux32
[23:42:25] -!- deep42thought has joined #archlinux32
[23:43:33] <deep42thought> If anyone wonders, what's up with buildmaster: I'm in contact with rewbycraft, here is part of what he wrote: "I've hit a bug in oVirt which is causing the cluster to crash and reboot if the combined disk I/O of all vms goes past a certain level"