#archlinux32 | Logs for 2020-05-21

Back
[00:01:25] -!- yans has quit [Ping timeout: 264 seconds]
[00:02:44] <buildmaster> i686/python-orjson is broken (says eurobuild6-2): https://archlinux32.org
[02:02:26] -!- isacdaavid has quit [Read error: Connection reset by peer]
[02:30:46] -!- samantaz_ has quit [Ping timeout: 246 seconds]
[03:16:43] <buildmaster> i486/freeglut is broken (says nlopc46-i486bs1): https://archlinux32.org
[03:25:44] <buildmaster> i486/ftgl is broken (says nlopc46-i486bs0): https://archlinux32.org
[03:40:26] <buildmaster> i486/graphene is broken (says nlopc46-i486bs1): https://archlinux32.org
[03:50:01] <buildmaster> i486/kexec-tools are broken (says eurobuild6-7-i486): https://archlinux32.org
[03:51:01] <buildmaster> i486/jasper is broken (says nlopc46-i486bs1): https://archlinux32.org
[04:15:02] <buildmaster> i486/libgrss are broken (says nlopc46-i486bs0): https://archlinux32.org
[04:16:54] <buildmaster> i686/kexec-tools are broken (says eurobuild6-3): https://archlinux32.org
[04:18:14] <buildmaster> pentium4/kexec-tools are broken (says eurobuild6-2): https://archlinux32.org
[04:46:11] <buildmaster> i486/pango-perl is broken (says eurobuild6-7-i486): https://archlinux32.org
[05:11:07] <buildmaster> i486/sdl is broken (says nlopc46-i486bs1): https://archlinux32.org
[05:19:55] <buildmaster> i686/libsidplay is broken (says eurobuild6-6): https://archlinux32.org
[05:20:38] <buildmaster> pentium4/libsidplay is broken (says eurobuild6-3): https://archlinux32.org
[05:26:43] <buildmaster> i486/perl-dbd-mysql is broken (says nlopc46-i486bs1): https://archlinux32.org
[05:47:30] -!- sunshavi has quit [Ping timeout: 260 seconds]
[07:06:56] -!- DCyrax has joined #archlinux32
[07:13:53] -!- sunshavi has joined #archlinux32
[08:48:44] * buildmaster goes insane.
[09:00:49] -!- sunshavi has quit [Ping timeout: 264 seconds]
[09:00:58] -!- deep42thought has joined #archlinux32
[09:00:59] <buildmaster> Hi deep42thought!
[09:00:59] <buildmaster> !rq deep42thought
[09:01:00] <phrik> buildmaster: <deep42thought> because you're a pessimist with no hope - or you're a realist afraid of change ... pick your evil
[09:02:02] <deep42thought> buildmaster: pull yourself together!
[09:02:02] <buildmaster> I will check my sanity again.
[09:04:00] * buildmaster resumes sanity.
[09:04:52] <deep42thought> thx, buildmaster
[09:04:54] <buildmaster> np, deep42thought
[09:06:25] -!- sunshavi has joined #archlinux32
[09:40:17] -!- abaumann has joined #archlinux32
[09:40:17] <buildmaster> Hi abaumann!
[09:40:18] <buildmaster> !rq abaumann
[09:40:18] <phrik> buildmaster: <abaumann> so, the next compression algorithm in pacman will require 1 TB of RAM? ;-)
[09:40:23] <deep42thought> Hi abaumann!
[09:40:27] <abaumann> hi deep42thought
[09:41:13] <deep42thought> virtualbox fails with some bogus make target, it seems
[09:41:20] <deep42thought> I'm clueless
[09:41:40] <deep42thought> ... as usual ;-)
[09:41:56] <abaumann> yeah. some shell script, IIRC
[09:42:06] <deep42thought> no, I don't think so
[09:42:18] <deep42thought> kmk: *** No rule to make target '/build/virtualbox/src/VirtualBox-6.1.8/out/linux.x86/release/lib/VBoxCOM.a', needed by '/build/virtualbox/src/VirtualBox-6.1.8/out/linux.x86/release/obj/vboximg-mount/vboximg-mount'.
[09:42:26] <abaumann> mmh.
[09:42:28] <abaumann> that's new
[09:42:39] <deep42thought> maybe, we just never got that far?
[09:42:39] <deep42thought> ;-)
[09:44:32] <abaumann> ==> virtualbox-guest-utils is part of package virtualbox
[09:44:36] <abaumann> this is unfortunate
[09:44:45] <deep42thought> yeah
[09:44:49] <abaumann> nobody wants to run a virtualbox on a 32-bit host, at least I hope.
[09:45:06] <abaumann> the guest integrations are handy though to test 32-bit Archlinux32 on something..
[09:45:30] <buildmaster> pentium4/sage-numerical-backends-coin is broken (says nlopc46): https://archlinux32.org
[09:47:38] <buildmaster> i486/sage-numerical-backends-coin is broken (says nlopc46-i486bs1): https://archlinux32.org
[09:50:23] <buildmaster> i686/sage-numerical-backends-coin is broken (says eurobuild6-4): https://archlinux32.org
[09:52:28] <buildmaster> i486/rime-cantonese is broken (says nlopc46-i486bs1): https://archlinux32.org
[10:10:31] <abaumann> kmk: *** No rule to make target '/build/virtualbox/src/VirtualBox-6.1.8/out/linux.x86/release/lib/VBoxCOM.a', needed by '/build/virtualbox/src/VirtualBox-6.1.8/out/linux.x86/release/obj/vboximg-mount/vboximg-mount'. Stop.
[10:10:35] <abaumann> kmk: Leaving directory '/build/virtualbox/src/VirtualBox-6.1.8/src/VBox/ImageMounter/vboximg-mount'
[10:10:38] <abaumann> yeah. confirmed
[10:12:06] <abaumann> kmk -C src/VBox/ImageMounter/vboximg-mount
[10:15:56] <abaumann> https://forums.virtualbox.org
[10:15:58] <phrik> Title: virtualbox.org • View topic - Compile VBOX 4.06 FAILED WHEN I DISABLE XPCOM (at forums.virtualbox.org)
[10:16:00] <abaumann> --disable-xpcom
[10:16:10] <abaumann> so the tool searches for a COM library
[10:16:19] <abaumann> are we implicitely disabling that somehow?
[10:17:23] <abaumann> ***** Checking libIDL *****
[10:17:23] <abaumann> found version 0.8.14
[10:17:31] <abaumann> libIDL is there for XPCOM, mmh.
[10:19:17] <deep42thought> *we* are only removin "-pipe"
[10:19:43] <deep42thought> well, explicitely removing
[10:19:46] <abaumann> kmk|kmk -j 1
[10:19:56] <abaumann> for non-parallel-out-of-memory builds.
[10:19:59] <abaumann> but otherwise?
[10:20:05] <deep42thought> there is a sed before that
[10:20:41] <abaumann> libXPCOM.a is not built. And this is the main library to do the Corba-thingy to communicate between clients as the mount-thingy with the VirtualBox server. I think..
[10:21:07] <deep42thought> maybe, it's not built on x86_64, either, but it is already there?
[10:21:18] <abaumann> there was a xxx.sh doing some library thing and failing, IIRC.
[10:21:19] <deep42thought> but for x86_64 only, obviously
[10:21:25] <deep42thought> ah, hmm
[10:21:28] <abaumann> I have to build again, with a log this time.
[10:21:32] <deep42thought> :-D
[10:21:41] <abaumann> maybe the build error gets silently ignored
[10:27:03] <abaumann> ERROR: recursively_umount_and_rm requires a directory as argument - "/data/home/slave1/builder/work/tmp.build-packages.5jLTOx" is not
[10:27:06] <abaumann> ERROR: Cannot lock ping-to-master - this should not happen.
[10:27:08] <abaumann> eeek
[10:27:17] <deep42thought> crappy code from me?
[10:27:22] * deep42thought hides
[10:27:28] <abaumann> just started a slave over another running slave
[10:27:30] <abaumann> me bad :-)
[10:27:51] <deep42thought> phew
[10:28:06] <abaumann> you can not take the blame all the time.. I also want my share. ;-)
[10:28:14] <deep42thought> !grab abaumann
[10:28:15] <phrik> deep42thought: Tada!
[10:28:15] <deep42thought> :-D
[10:28:26] * deep42thought hands abaumann a share of blame
[10:28:51] * abaumann thanks the donator
[10:48:25] <buildmaster> i686/haskell-sbv is broken (says nlopc46): https://archlinux32.org
[10:52:36] <buildmaster> pentium4/haskell-sbv is broken (says nlopc46): https://archlinux32.org
[11:18:28] <abaumann> Haskell auto-triggering doesn't seem to work properly: https://archlinux32.org
[11:18:32] <phrik> Title: Arch Linux 32 - List of Package Builds (at archlinux32.org)
[11:18:35] <abaumann> tons of packages on the failed list..
[11:19:49] <deep42thought> maybe autoscheduling works, but prioritizing those packages, that failed before, does not?
[11:20:04] <deep42thought> e.g. I suspect, haskell-network would build now
[11:20:11] <abaumann> let me try.
[11:20:20] <deep42thought> but it was not built since 2020-05-17
[11:21:09] <deep42thought> yup, haskell-quickcheck-io (the one breaking haskell-network) was built in the evening of 2020-05-17
[11:21:26] <deep42thought> ~1h after haskell-network failed
[11:21:32] <abaumann> haskell-network builds fine now on i686
[11:22:04] <deep42thought> I prioritized all of haskell
[11:22:06] <deep42thought> let's see ...
[11:23:25] <deep42thought> prioritizing formerly-broken packages, when a rescheduled haskell-* package returns successfully, is quite hard
[11:23:41] <deep42thought> there is no information, *why* the returned package was scheduled
[11:23:49] <deep42thought> attached to that build, I mean
[11:23:55] <abaumann> yep.
[11:24:29] <abaumann> or we need a watchdog of sorts, which scans Haskell build files and tries to reschedule packages which fail to fullfil the dependencies.
[11:24:35] <abaumann> but that's quite some work to do.
[11:25:04] <abaumann> last time haskell built, but not necessarily worked, that's another issue.
[11:25:07] <deep42thought> we could scan those files on returning haskell packages
[11:25:20] <abaumann> I start to get the feeling that haskell packages need to be built in a very specific order
[11:25:44] <deep42thought> I have that feeling for a long time
[11:41:48] -!- samantaz_ has joined #archlinux32
[12:03:27] <deep42thought> ok, I think, I see the problem with haskell
[12:03:51] <deep42thought> we have a, makedepending on b, depending on c
[12:04:00] <deep42thought> c gets rebuilt, so b is broken
[12:04:19] <deep42thought> but a thinks, b is still available (which is true, but it's broken) for building
[12:04:38] <abaumann> sounds reasonable
[12:04:54] <deep42thought> this is fallout from "we do not need the newest git to build something, that has makedepends=(git)"
[12:05:11] <deep42thought> we had once an exception for haskell packages
[12:05:21] <deep42thought> that treated all makedepends of haskell packages like depends
[12:05:31] <deep42thought> but this introduced a lot of cycles, naturally
[12:05:31] <abaumann> ah
[12:06:14] <deep42thought> the interesting part is, that we *do* have all the info for b to know, that it is currently broken
[12:06:45] <abaumann> so it could be retriggerd?
[12:06:46] <deep42thought> e.g. it does not simply "depend on c", but in the database it depends on "libHSc-cryptic-hash.so"
[12:07:00] <deep42thought> it *was* still in the build-list, I think
[12:07:01] <buildmaster> i486/bash-language-server is broken (says nlopc46-i486bs0): https://archlinux32.org
[12:07:14] <deep42thought> we need to withhold a as long as b cannot be installed
[12:07:27] <deep42thought> I need to check, why that did not happen (it seems obvious to do so)
[12:08:55] <abaumann> as long as a doesn't require a _new_ b to build sucessfully?
[12:09:02] <deep42thought> it does
[12:09:06] <deep42thought> because the old is broken
[12:09:10] <abaumann> ah.
[12:09:36] <deep42thought> btw: I just saw, that we *do* prioritize the broken-due-to-other-package package on rescheduling
[12:09:50] <deep42thought> which should suffice
[12:09:58] <abaumann> that's good, so we get out fast..
[12:10:09] <deep42thought> if the local mirrors sync up fast ;-)
[12:10:36] * deep42thought looked at the wrong script
[12:10:49] <abaumann> hence the sync_mirrir.php script hooks :-)
[12:10:54] <deep42thought> yeah
[12:11:04] * abaumann heads over to check if that one is still running on his mirror..
[12:11:06] <deep42thought> but we do not wait for the sync to finish, so ...
[12:11:09] <deep42thought> :-D
[12:11:17] <abaumann> oh.
[12:11:22] <abaumann> that's maybe not ideal.. :-)
[12:11:37] <abaumann> though, it just adds some additional build rounds..
[12:11:49] <deep42thought> the priorization gets lost, then, too
[12:12:04] <abaumann> ..that's bad again..
[12:12:19] <abaumann> ..and explains, why some packages wait for a build for a long time
[12:12:22] <deep42thought> we could bail out by not setting the priority to 0 but by decreasing it
[12:12:42] <abaumann> "reversing the polarity" ;-)
[12:12:42] * deep42thought ponders this
[12:12:49] <deep42thought> huh?
[12:13:02] <abaumann> Dr Who quote.
[12:13:48] <abaumann> pacman -S qtcreator
[12:13:56] <abaumann> "Reverse the polarity of the neutron flow"
[12:15:21] <deep42thought> wtf? https://git.archlinux32.org
[12:15:22] <phrik> Title: mysql-functions ÂŤ lib - builder - Archlinux32 build system (at git.archlinux32.org)
[12:16:06] <abaumann> special treatement? :-)
[12:16:11] <deep42thought> yeah
[12:16:27] <deep42thought> haskell dependencies do *not* get ignored if they failed to build for >1 day :-D
[12:16:33] <deep42thought> probably for the best
[12:16:50] <deep42thought> # ... or (b) not relevant for binary_packages (e.g. a makedepends) has
[12:16:50] <deep42thought> # no built provider
[12:16:51] <deep42thought> aha
[12:17:04] <deep42thought> we need to check, that the provider is not only built, but also installable ...
[12:19:22] <deep42thought> hmm, I know, why I did not include this check
[12:19:33] <deep42thought> the dependency problem could be several layers down
[12:20:00] <deep42thought> e.g. a makedepends on b->c->d->e; e was recompiled, so d is broken, but c and b don't know
[12:20:17] <abaumann> "Tres Leches Cake"
[12:20:40] <abaumann> would be transitive..
[12:20:46] <abaumann> and more than three layers. :-)
[12:23:22] <deep42thought> I get more and more the feeling, I should start to write another thesis (besides the one about relativistic optics): one about correct dependency tracking and build scheduling of linux packages
[12:24:04] <buildmaster> i686/haskell-lifted-async is broken (says nlopc46): https://archlinux32.org
[12:27:09] <abaumann> :-)
[12:28:17] <abaumann> I pushed qtcreator to stable (because of the clang breakage)
[12:29:47] <deep42thought> thx
[12:30:32] <abaumann> I broke it, so.. :-)
[12:39:09] <abaumann> aha. Also clazy needs a nudge..
[12:48:29] <T`aZ> \o/
[13:02:55] <abaumann> yep: comparing build logs of virtualbox shows that xpcom is simply not built on 32-bit..
[13:20:43] -!- drathir_tor has quit [Ping timeout: 240 seconds]
[13:31:29] -!- drathir_tor has joined #archlinux32
[13:48:03] -!- drathir_tor has quit [Quit: Lost terminal]
[14:03:00] <buildmaster> i686/haskell-simple-sendfile is broken (says nlopc46): https://archlinux32.org
[14:04:41] -!- drathir_tor has joined #archlinux32
[14:52:03] <buildmaster> i686/haskell-hjsonpointer is broken (says eurobuild6-1): https://archlinux32.org
[14:59:52] <buildmaster> i686/haskell-aeson-yaml is broken (says eurobuild6-1): https://archlinux32.org
[15:00:32] <abaumann> Makefile.kmk:68: *** missing 'endef', unterminated 'define'. Stop.
[15:00:43] <abaumann> cool this kmk makefile even has English bugs.. :-)
[15:02:31] <deep42thought> :-)
[15:02:39] <abaumann> we could simply not build vboximg-mount
[15:02:53] <deep42thought> if that works for now :-)
[15:03:23] <abaumann> it's just a FUSE mount so you can mount a virtual box image on the host.
[15:03:51] <abaumann> "vboximg-mount is a command line utility for Mac OS X hosts that provides raw access to an Oracle VM VirtualBox virtual disk image on the host system. Use this utility to mount, view, and optionally modify the disk image contents."
[15:04:15] <abaumann> most likely because on Linux you have qemu-img
[15:04:19] <deep42thought> only on macosx, right?
[15:04:36] <abaumann> somebody upstream thought it's a good idea to build it.
[15:04:39] * abaumann calls a difflog
[15:04:44] <buildmaster> i686/haskell-path is broken (says nlopc46): https://archlinux32.org
[15:05:48] <abaumann> 9c552c55d21dcd08f69e24a97b5e59ec7d97ca7c build & install vboximg-mount
[15:08:37] <buildmaster> i686/haskell-ipynb is broken (says nlopc46): https://archlinux32.org
[15:11:41] <buildmaster> pentium4/haskell-aeson is broken (says eurobuild6-1): https://archlinux32.org
[15:17:47] <abaumann> cool llvm: gold linker issues on 32-bit
[15:18:27] <abaumann> the gold linker is causing nothing than problems everywhere IMHO
[15:18:33] <abaumann> *buT*
[15:35:55] <trotz> 2020/05/21 15:35 CRIT jeti100 Master Mirror HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 6 - Couldn't resolve host name
[15:36:56] <trotz> 2020/05/21 15:36 CRIT buildmaster Git HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 28 - Timeout was reached
[15:45:56] <trotz> 2020/05/21 15:45 OK jeti100 Master Mirror HTTP OK: HTTP/1.1 200 OK - 1408 bytes in 0.034 second response time
[15:56:57] <trotz> 2020/05/21 15:56 OK buildmaster Git HTTP OK: HTTP/1.1 200 OK - 9495 bytes in 0.032 second response time
[15:57:58] <abaumann> install: cannot stat 'VBox.sh': No such file or directory
[15:58:10] <abaumann> aha. that one I meant. No clue what goes wrong there...
[16:00:22] <buildmaster> pentium4/virtualbox is broken (says eurobuild6-1): https://archlinux32.org
[16:04:24] <buildmaster> i686/virtualbox is broken (says nlopc46): https://archlinux32.org
[16:08:17] <abaumann> cd "VirtualBox-$pkgver/out/linux.$BUILD_PLATFORM_ARCH/release/bin"
[16:08:22] <abaumann> install -m0755 VBox.sh "$pkgdir/usr/bin/VBox"
[16:09:46] <abaumann> ./src/VBox/Installer/linux/VBox.sh
[16:12:42] <abaumann> somthing doesn't add up here.. :-)
[16:16:57] <trotz> 2020/05/21 16:16 CRIT szilassi PkgApi JSON HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[16:37:01] <trotz> 2020/05/21 16:36 OK szilassi PkgApi JSON HTTP OK: HTTP/2.0 200 - 1050 bytes in 0.708 second response time
[16:47:01] <trotz> 2020/05/21 16:45 CRIT buildmaster Git HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 28 - Timeout was reached
[17:02:50] <abaumann> out/linux.amd64/release/bin/VBox.sh but none on 32-bit..
[17:03:38] <deep42thought> the sh script should not depend on the architecture, should it?
[17:04:14] <abaumann> no.
[17:04:25] <deep42thought> so why not simply copy it over, then?
[17:04:27] <abaumann> and it's in the release folder but then not coppied over.
[17:04:36] <abaumann> because other stuff might be missing.
[17:04:45] <abaumann> I don't trust this kmk-build system a bit..
[17:05:00] <abaumann> who would write his own make tool anyway for a project.
[17:05:10] * deep42thought has done so before
[17:05:14] <deep42thought> it was a mess
[17:05:21] <abaumann> shell, I suppose. :-)
[17:05:26] <deep42thought> no, pascal
[17:05:27] <deep42thought> :-D
[17:05:29] <abaumann> ah.
[17:05:43] <abaumann> well, my newest project also has just a build.sh and a test.sh script
[17:06:18] <buildmaster> i686/haskell-tree-diff is broken (says eurobuild6-1): https://archlinux32.org
[17:07:02] <trotz> 2020/05/21 17:06 OK buildmaster Git HTTP OK: HTTP/1.1 200 OK - 9496 bytes in 0.032 second response time
[17:08:08] <abaumann> aha, it might try to change to ls out/linux.x86/release/bin/
[17:08:13] <abaumann> but BUILD_PLATFORM_ARCH is not x86
[17:08:22] <abaumann> where does this BUILD_PLATFORM_ARCH come from anyway?
[17:08:28] * abaumann asks a grep
[17:08:57] <buildmaster> pentium4/haskell-http2 is broken (says nlopc46): https://archlinux32.org
[17:09:16] <abaumann> makepkg or some library would be a logical place
[17:10:32] <abaumann> oh. or in the devtools scripts
[17:10:49] <deep42thought> I haven't seen this variable before
[17:11:00] <deep42thought> sounds like some configure variable inside the PKGBUILD to me
[17:11:04] <abaumann> ah.
[17:11:06] <abaumann> env.sh
[17:11:10] <deep42thought> :-)
[17:11:21] <abaumann> ARCH stadns for Architecture not for Archlinux
[17:11:25] <abaumann> my wrong. :-)
[17:11:33] <abaumann> BUILD_PLATFORM_ARCH="x86"
[17:11:39] * abaumann scratches his head
[17:12:10] <abaumann> if build() does a source ./env.sh, should that also be set in package_virtualbox()?
[17:12:19] <abaumann> I think I read somewhere something about something.
[17:12:22] <deep42thought> yes, probably
[17:12:28] <deep42thought> but I think, it might get lost
[17:12:37] <abaumann> because it's not good if enviroments get shared between the phases.
[17:12:41] <deep42thought> yes
[17:12:55] <abaumann> Why this works on 64-bit, is a miracle to me.
[17:12:57] <deep42thought> elibrokeit is the expert on this terrain
[17:13:09] <abaumann> he should consider himself summoned ;-)
[17:13:09] <deep42thought> maybe, the variable is not needed on x86_64?
[17:13:13] <deep42thought> lol
[17:13:28] * eschwartz is summoned
[17:13:38] <abaumann> :-)
[17:13:39] <deep42thought> Hi eschwartz
[17:13:44] <abaumann> hi
[17:13:50] <abaumann> we should check the devtools versions..
[17:13:54] <eschwartz> if build() sources anything, and it is needed in package() too, then it needs to be sourced in package()
[17:13:54] <abaumann> ..might be a difference there.
[17:14:06] <eschwartz> but this is because you might run makepkg --repackage
[17:14:13] <deep42thought> ah, right
[17:14:15] <eschwartz> and makepkg --repackage does not execute build()
[17:14:17] <abaumann> ah, yes.
[17:14:30] <deep42thought> ok, so we will try sourcing env.sh in package_*(), too
[17:14:38] * abaumann is on it..
[17:15:34] <abaumann> actually, I have another suspision
[17:15:36] <abaumann> there is a source "VirtualBox-$pkgver/env.sh"
[17:15:39] <abaumann> in package_virtualbox
[17:15:56] <abaumann> in build cd "VirtualBox-$pkgver"
[17:15:59] <abaumann> and source ./env.sh
[17:16:00] <abaumann> mmh.
[17:16:47] <eschwartz> virtualbox, eh?
[17:16:53] <abaumann> yep.
[17:17:05] <eschwartz> that's an exciting one
[17:17:13] <abaumann> in various ways :-)
[17:17:44] <abaumann> first of all building macX vboximg-mount on 64-bit which fails miracoulsly on 32-bit (by not building an XPCOM library)
[17:18:47] <abaumann> if "cd "VirtualBox-$pkgver/out/linux.$BUILD_PLATFORM_ARCH/release/bin" fails in PKGBUILD, does the build silently continue?
[17:19:07] <deep42thought> no, it shouldn't, I think
[17:19:16] <abaumann> or not anymore
[17:19:19] <eschwartz> it should fail like any other command
[17:19:24] <abaumann> ok
[17:19:41] <eschwartz> unless you have a secret cd program which doesn't return failures :p
[17:19:49] <abaumann> of course ;-)
[17:20:01] <trotz> 2020/05/21 17:19 CRIT buildmaster Git Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:20:04] * abaumann likes his AMD Ryzen 7 2700X and uses it at it's fullest for Virtualbox
[17:20:11] <deep42thought> :-)
[17:20:13] <abaumann> *its
[17:21:01] <trotz> 2020/05/21 17:20 CRIT archlinux32.org Home Certificate HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:21:13] -!- thalos has joined #archlinux32
[17:22:01] <trotz> 2020/05/21 17:21 CRIT archlinux32.org Home HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 6 - Couldn't resolve host name
[17:22:01] <trotz> 2020/05/21 17:21 CRIT archlinux32.org Home Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:22:01] <trotz> 2020/05/21 17:21 CRIT buildmaster Git HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 6 - Couldn't resolve host name
[17:22:45] <abaumann> aha: BUILD_PLATFORM_ARCH: x86
[17:22:50] <abaumann> and the directory exists
[17:23:01] <trotz> 2020/05/21 17:21 CRIT buildmaster Buildmaster Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:23:01] <trotz> 2020/05/21 17:22 CRIT jeti100 Archive Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:23:05] <abaumann> just VBox.sh doesn't get installed into that out directory
[17:23:05] <deep42thought> what's the deal about all those dns resolution errors on the buildmaster?
[17:23:20] <abaumann> oh.
[17:23:43] <abaumann> checking..
[17:23:56] <abaumann> btw. Hetzner says we are not on a blacklist (mail).
[17:24:01] <trotz> 2020/05/21 17:23 CRIT jeti100 Master Mirror HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 6 - Couldn't resolve host name
[17:24:01] <trotz> 2020/05/21 17:23 CRIT jeti100 Master Mirror Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:24:02] <abaumann> another issue
[17:24:15] <thalos> Hi! I want to install Arch32 on a platform that are missing the cmov instruction (i.e not i686 compatible). Archlinux32.org talks about a i486 package repository but I can't find any boot images for i486. I only see dual and i686. Are there no bootable iso for i486?
[17:24:24] <deep42thought> that's good (mx on hetzner)
[17:25:00] <deep42thought> thalos: we don't have isos for i486, but you can use pacstrap to install from a different host
[17:25:00] <abaumann> http://archlinux32.andreasbaumann.cc
[17:25:02] <trotz> 2020/05/21 17:24 CRIT jeti100 Archive HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 6 - Couldn't resolve host name
[17:25:02] <phrik> Title: Index of /arch486isos (at archlinux32.andreasbaumann.cc)
[17:25:16] <deep42thought> ah, right, we *do* have i486 isos, but they're "experimental" :-)
[17:25:22] <abaumann> and old.
[17:25:33] <abaumann> you might have to fiddle with pgp keys before you can install.
[17:26:01] <trotz> 2020/05/21 17:25 CRIT jeti100 Master Mirror Certificate HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:26:01] <trotz> 2020/05/21 17:25 CRIT szilassi PkgApi JSON HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[17:26:03] <abaumann> the question is really, can you boot on that machine (then ISO) or do can you install a compact flash or hard disk and just pacstrap from another machine.
[17:26:13] <thalos> So what would you recommened? Pacstrap or the old/experimental isos?
[17:26:19] <deep42thought> pacstrap
[17:26:32] <abaumann> especially, if you have no cdrom
[17:26:59] <abaumann> putting that i486 iso on an USB stick is even more experimental (and you'll need plopp or a similar helper to get USB booting the right way)
[17:27:01] <trotz> 2020/05/21 17:26 CRIT jeti100 Archive Certificate HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[17:27:20] <abaumann> old hardware has USB, but usually you can only boot from media via quirks.
[17:27:59] <abaumann> edit your hosts /etc/pacman.conf, set Architecture=i486, do a pacman -Sy (do not update your host!), then do a normal pacstrap.
[17:28:18] <abaumann> arch-chroot is no problem, as i486 is a subset architecture.
[17:28:29] <abaumann> so you can customize stuff in the chroot on the host.
[17:29:19] <abaumann> cURL returned 6
[17:29:25] <abaumann> sounds like my check_curl
[17:29:27] <abaumann> :-)
[17:29:59] <thalos> Hmm, ok. I'll have to work something out, I don't have arch on any other of my hosts. But I guess I can install arch in a VM and pass-through the HDD to the VM and pacstrap the drive from there
[17:30:08] <deep42thought> nah
[17:30:11] <deep42thought> just boot a live iso
[17:30:20] <thalos> True
[17:30:21] <abaumann> you can also dd a alix-i486.img.xz http://archlinux32.andreasbaumann.cc
[17:30:23] <phrik> Title: Index of /arch486isos (at archlinux32.andreasbaumann.cc)
[17:30:25] <deep42thought> but make sure, you install to the right hdd ;-)
[17:30:38] <abaumann> which is an installed image with root and password xx (I think)
[17:31:24] <abaumann> PXE is another (experimental) option. Tricky, unless you are familiar with NBD, TFTP, DHCP and friends. :-)
[17:31:45] <abaumann> and you need a network card doing PXE properly.
[17:32:01] <trotz> 2020/05/21 17:31 OK buildmaster Git HTTP OK: HTTP/1.1 200 OK - 9496 bytes in 0.023 second response time
[17:32:51] <thalos> I did PXE/tftp boot back in the day on SUN workstations :D But just dd'ing alix sounds vert convenient!
[17:32:54] <thalos> *very
[17:33:42] <abaumann> just to warn you: i486 is basically text-mode only currently.
[17:34:02] <trotz> 2020/05/21 17:33 OK jeti100 Master Mirror HTTP OK: HTTP/1.1 200 OK - 1408 bytes in 0.034 second response time
[17:34:03] <trotz> 2020/05/21 17:33 OK jeti100 Master Mirror Secure HTTP OK: HTTP/1.1 200 OK - 1408 bytes in 0.080 second response time
[17:34:03] <thalos> Who uses GUI's? Yuck ;D
[17:34:11] <deep42thought> my words
[17:34:15] <abaumann> :-)
[17:34:42] <thalos> Although, I guess X is good for having more than 1 terminal on the screen at the same time :)
[17:35:02] <trotz> 2020/05/21 17:34 OK jeti100 Archive HTTP OK: HTTP/1.1 200 OK - 821 bytes in 6.714 second response time
[17:35:45] <abaumann> mmh. the DNS@Hetzner didn't change
[17:36:02] <trotz> 2020/05/21 17:35 OK jeti100 Master Mirror Certificate OK - Certificate 'arch.eckner.net' will expire on Tue 11 Aug 2020 03:57:58 PM GMT +0000.
[17:36:02] <trotz> 2020/05/21 17:35 OK szilassi PkgApi JSON HTTP OK: HTTP/2.0 200 - 1050 bytes in 0.681 second response time
[17:36:13] <deep42thought> these ocassional dns issues are not new, either
[17:36:31] <abaumann> I always blamed them on systemd-resolved
[17:36:39] <abaumann> but there is no systemd-resolved on the buildmastre
[17:37:02] <trotz> 2020/05/21 17:36 OK jeti100 Archive Certificate OK - Certificate 'arch.eckner.net' will expire on Tue 11 Aug 2020 03:57:58 PM GMT +0000.
[17:37:15] <deep42thought> there is
[17:37:33] <deep42thought> at least in /etc/nsswitch.conf
[17:37:36] <abaumann> Active: inactive (dead)
[17:37:53] <deep42thought> hosts: files mymachines myhostname resolve [!UNAVAIL=return] dns
[17:38:05] <deep42thought> doesn't that mean, resolved is asked, first?
[17:38:13] <deep42thought> well, "first" = "before dns"
[17:38:22] <abaumann> mmh
[17:38:51] <abaumann> resolve is not even in the manpage *grmpf*
[17:39:13] <abaumann> this is bonkers.
[17:39:32] <abaumann> I removed it
[17:40:02] <trotz> 2020/05/21 17:39 OK archlinux32.org Home HTTP OK: HTTP/1.1 200 OK - 22910 bytes in 0.780 second response time
[17:40:02] <trotz> 2020/05/21 17:39 OK buildmaster Git Secure HTTP OK: HTTP/1.1 200 OK - 9257 bytes in 0.034 second response time
[17:40:06] <abaumann> aha. :-)
[17:40:08] <abaumann> there you go. :-)
[17:40:31] <deep42thought> that's one of the first things I do on all of my boxes :-)
[17:40:48] <abaumann> I wonder how it worked till now?
[17:41:01] <trotz> 2020/05/21 17:39 OK archlinux32.org Home Certificate OK - Certificate 'archlinux32.org' will expire on Fri 31 Jul 2020 11:43:23 PM GMT +0000.
[17:41:01] <trotz> 2020/05/21 17:40 OK archlinux32.org Home Secure HTTP OK: HTTP/2.0 200 - 22998 bytes in 0.850 second response time
[17:42:01] <trotz> 2020/05/21 17:41 OK buildmaster Buildmaster Secure HTTP OK: HTTP/2.0 200 - 2903 bytes in 0.234 second response time
[17:42:22] <abaumann> aha.
[17:42:23] <deep42thought> my problems with systemd-resolved mainly circle around the caching
[17:42:30] <abaumann> this expains my funny problems in my network!
[17:42:34] <abaumann> resolve [!UNAVAIL=return]
[17:43:01] <trotz> 2020/05/21 17:42 OK jeti100 Archive Secure HTTP OK: HTTP/1.1 200 OK - 619 bytes in 0.081 second response time
[17:43:11] <abaumann> not only that, it has wrong timeout handling (so the first request fails, the second succeeds)
[17:43:19] <abaumann> "product ripes at the end user"
[17:43:32] * abaumann doesn't want to start another rant about the project starting with an s.
[17:44:07] <abaumann> but this means, the standard nsswitch.conf _requires_ systemd-resolved to be configured in order not to have hickups on the system?
[17:45:00] <abaumann> local DNS caching is always tried again, till it causes unwanted side effects
[17:45:18] <abaumann> DNS itself is caching and designed to do so, so I never understood the need of local DNS caches.
[17:45:26] <abaumann> also not the old Redhat ones..
[17:47:55] <abaumann> nscd, that was the name
[17:50:08] <deep42thought> https://www.freedesktop.org
[17:50:09] <phrik> Title: nss-resolve (at www.freedesktop.org)
[17:50:47] <deep42thought> "but are routed to nss-dns if this service is not available."
[17:51:11] <deep42thought> so not-running systemd-resolved will also disable the "resolve"
[17:51:42] <abaumann> yeah, so why the hickups?
[17:52:03] <abaumann> at least there is some timeout and time loss involved..
[17:52:27] <abaumann> and nss is supposed an optional thing.
[17:53:17] <abaumann> yp, nis+, nss, those are things for big companies..
[17:53:23] <abaumann> not for a laptop.
[17:53:35] <abaumann> yeah, just a personal opinion. :-)
[17:57:50] <abaumann> kBuild: Installing /build/virtualbox/src/VirtualBox-6.1.8/out/linux.amd64/release/bin/VBox.sh
[17:57:54] <abaumann> completely missing on 32-buit
[17:57:57] <abaumann> *32-bit
[17:58:21] <deep42thought> should this be in out/linux.x86?
[17:58:30] <abaumann> yes
[17:58:43] <abaumann> out/linux.x86/release/bin
[17:59:01] <trotz> 2020/05/21 17:58 CRIT buildmaster Git HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 28 - Timeout was reached
[17:59:48] -!- drathir_tor has quit [Remote host closed the connection]
[17:59:55] <abaumann> in the end it's a bug in kmk
[18:09:03] <trotz> 2020/05/21 18:08 OK buildmaster Git HTTP OK: HTTP/1.1 200 OK - 9496 bytes in 0.031 second response time
[18:14:56] <abaumann> VBoxTunctl
[18:15:04] <abaumann> ok. laters. I'll try another day..
[18:15:05] -!- abaumann has quit [Quit: leaving]
[18:15:08] <deep42thought> cu
[18:16:03] <trotz> 2020/05/21 18:15 CRIT szilassi PkgApi JSON HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[18:21:03] <trotz> 2020/05/21 18:20 OK szilassi PkgApi JSON HTTP OK: HTTP/2.0 200 - 1050 bytes in 0.857 second response time
[18:31:32] <buildmaster> i486/jacktrip is broken (says nlopc46-i486bs1): https://archlinux32.org
[19:04:59] <buildmaster> i686/haskell-sandi is broken (says eurobuild6-1): https://archlinux32.org
[19:07:58] <buildmaster> pentium4/haskell-sandi is broken (says eurobuild6-1): https://archlinux32.org
[19:24:45] <buildmaster> i686/haskell-hpack is broken (says eurobuild6-1): https://archlinux32.org
[19:24:59] <buildmaster> i686/haskell-http-conduit is broken (says nlopc46): https://archlinux32.org
[19:26:05] <trotz> 2020/05/21 19:25 CRIT buildmaster Git HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 6 - Couldn't resolve host name
[19:26:05] <trotz> 2020/05/21 19:25 CRIT buildmaster Git Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 6 - Couldn't resolve host name
[19:28:14] <buildmaster> i686/haskell-mustache is broken (says eurobuild6-1): https://archlinux32.org
[19:36:08] <trotz> 2020/05/21 19:35 OK buildmaster Git HTTP OK: HTTP/1.1 200 OK - 9496 bytes in 5.033 second response time
[19:36:08] <trotz> 2020/05/21 19:35 OK buildmaster Git Secure HTTP OK: HTTP/1.1 200 OK - 9257 bytes in 5.034 second response time
[20:06:56] -!- drathir_tor has joined #archlinux32
[20:25:49] -!- isacdaavid has joined #archlinux32
[20:31:10] <trotz> 2020/05/21 20:29 CRIT szilassi PkgApi JSON HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[20:39:34] <nit-picker> deep42thought: your slave rechenknecht builds pentium4/chromium for more than a day, now (2 day(s) 00:38:35)
[20:41:10] <trotz> 2020/05/21 20:40 OK szilassi PkgApi JSON HTTP OK: HTTP/2.0 200 - 1050 bytes in 0.635 second response time
[20:46:38] -!- isacdaavid has quit [Read error: Connection reset by peer]
[21:08:05] <buildmaster> i486/elasticsearch is broken (says eurobuild6-7-i486): https://archlinux32.org
[21:11:44] <buildmaster> i486/rdiff-backup is broken (says nlopc46-i486bs0): https://archlinux32.org
[21:12:30] <buildmaster> i486/fscrypt is broken (says nlopc46-i486bs1): https://archlinux32.org
[21:22:11] <trotz> 2020/05/21 21:21 CRIT buildmaster Git Secure HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[21:34:13] <trotz> 2020/05/21 21:33 CRIT szilassi PkgApi JSON HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[21:42:14] <trotz> 2020/05/21 21:41 OK buildmaster Git Secure HTTP OK: HTTP/1.1 200 OK - 9257 bytes in 0.030 second response time
[21:51:13] <trotz> 2020/05/21 21:50 CRIT buildmaster Git HTTP CRITICAL - Invalid HTTP response received from host on port 80: cURL returned 28 - Timeout was reached
[21:54:13] <trotz> 2020/05/21 21:53 CRIT jeti100 Archive Certificate HTTP CRITICAL - Invalid HTTP response received from host on port 443: cURL returned 28 - Timeout was reached
[21:59:53] -!- deep42thought has quit [Quit: Leaving.]
[22:04:14] <trotz> 2020/05/21 22:03 OK jeti100 Archive Certificate OK - Certificate 'arch.eckner.net' will expire on Tue 11 Aug 2020 03:57:58 PM GMT +0000.
[22:04:14] <trotz> 2020/05/21 22:03 OK szilassi PkgApi JSON HTTP OK: HTTP/2.0 200 - 1050 bytes in 0.616 second response time
[22:11:15] <trotz> 2020/05/21 22:10 OK buildmaster Git HTTP OK: HTTP/1.1 200 OK - 9497 bytes in 0.027 second response time
[22:24:20] <buildmaster> i686/fscrypt is broken (says eurobuild6-1): https://archlinux32.org
[22:27:01] <buildmaster> pentium4/fscrypt is broken (says nlopc46): https://archlinux32.org
[22:29:34] -!- sunshavi has quit [Ping timeout: 240 seconds]
[22:36:38] -!- sunshavi has joined #archlinux32
[23:05:48] <buildmaster> i686/webkit2gtk is broken (says eurobuild6-2): https://archlinux32.org
[23:32:25] <buildmaster> i486/rime-cantonese is broken (says nlopc46-i486bs1): https://archlinux32.org
[23:33:39] <buildmaster> i686/darcs are broken (says nlopc46): https://archlinux32.org
[23:39:16] <buildmaster> i486/sundials are broken (says nlopc46-i486bs1): https://archlinux32.org