#archlinux32 | Logs for 2019-05-05

Back
[00:06:12] -!- rcf has joined #archlinux32
[00:09:37] <buildmaster> i686/qt5-tools are broken (says nlopc43).
[00:23:41] -!- thePiGrepper has quit [Ping timeout: 255 seconds]
[00:49:22] <buildmaster> pentium4/qt5-doc is broken (says nlopc43).
[01:45:20] -!- thePiGrepper has joined #archlinux32
[01:51:45] <buildmaster> pentium4/kcoreaddons are broken (says eurobuild3).
[01:57:28] -!- thePiGrepper has quit [Ping timeout: 250 seconds]
[02:00:30] <buildmaster> pentium4/karchive is broken (says rechenknecht).
[02:05:49] <buildmaster> pentium4/kcodecs are broken (says rechenknecht).
[02:07:41] <buildmaster> pentium4/ki18n is broken (says eurobuild3).
[02:08:08] <buildmaster> pentium4/kwidgetsaddons are broken (says nlopc43).
[02:09:02] <buildmaster> pentium4/kwindowsystem is broken (says nlopc43).
[02:10:05] <buildmaster> pentium4/kdbusaddons are broken (says nlopc43).
[02:11:10] <buildmaster> pentium4/kconfig is broken (says rechenknecht).
[02:13:11] -!- buildmaster has quit [Remote host closed the connection]
[02:13:23] -!- buildmaster has joined #archlinux32
[02:13:26] <buildmaster> pentium4/attica is broken (says nlopc43).
[02:14:26] <buildmaster> pentium4/sonnet is broken (says nlopc43).
[02:15:25] <buildmaster> pentium4/kguiaddons are broken (says eurobuild3).
[02:15:49] <buildmaster> pentium4/solid is broken (says nlopc43).
[02:16:51] <buildmaster> pentium4/kitemviews are broken (says rechenknecht).
[02:18:51] -!- buildmaster has quit [Remote host closed the connection]
[02:19:04] -!- buildmaster has joined #archlinux32
[02:20:14] <buildmaster> i686/kcoreaddons are broken (says nlopc43).
[02:21:45] <buildmaster> i686/karchive is broken (says nlopc43).
[02:22:12] <buildmaster> pentium4/kidletime is broken (says eurobuild3).
[02:22:54] <buildmaster> i686/kcodecs are broken (says nlopc43).
[02:24:17] <buildmaster> i686/kwidgetsaddons are broken (says nlopc43).
[02:25:25] <buildmaster> i686/kguiaddons are broken (says nlopc43).
[02:26:26] <buildmaster> i686/kwindowsystem is broken (says nlopc43).
[02:27:28] <buildmaster> i686/ki18n is broken (says rechenknecht).
[02:28:00] <buildmaster> i686/kdbusaddons are broken (says nlopc43).
[02:29:09] <buildmaster> i686/kitemviews are broken (says nlopc43).
[02:30:10] <buildmaster> i686/kconfig is broken (says eurobuild3).
[02:30:19] -!- samantaz__ has quit [Ping timeout: 248 seconds]
[02:30:48] <buildmaster> i686/sonnet is broken (says nlopc43).
[02:39:59] <buildmaster> i686/attica is broken (says rechenknecht).
[02:40:55] <buildmaster> i686/solid is broken (says eurobuild3).
[02:42:20] <buildmaster> i686/kidletime is broken (says nlopc43).
[02:44:20] -!- buildmaster has quit [Remote host closed the connection]
[02:44:32] -!- buildmaster has joined #archlinux32
[02:54:05] <buildmaster> pentium4/syntax-highlighting is broken (says nlopc43).
[02:55:20] <buildmaster> pentium4/kitemmodels are broken (says nlopc43).
[02:55:52] <buildmaster> pentium4/threadweaver is broken (says rechenknecht).
[02:57:52] -!- buildmaster has quit [Remote host closed the connection]
[02:58:05] -!- buildmaster has joined #archlinux32
[03:00:00] <buildmaster> i686/syntax-highlighting is broken (says nlopc43).
[03:04:03] <buildmaster> i686/threadweaver is broken (says eurobuild3).
[03:04:48] <buildmaster> i686/kitemmodels are broken (says rechenknecht).
[03:40:09] <buildmaster> pentium4/gsettings-qt is broken (says eurobuild3).
[03:44:52] -!- ubuntuxp has quit [Quit: ERC (IRC client for Emacs 26.2)]
[04:09:11] <buildmaster> pentium4/alure is broken (says nlopc43).
[04:36:27] <buildmaster> pentium4/breeze-gtk is broken (says nlopc43).
[05:08:02] -!- thePiGrepper has joined #archlinux32
[05:52:20] -!- yans has quit [Ping timeout: 250 seconds]
[06:13:07] -!- guys has quit [Ping timeout: 276 seconds]
[06:17:42] -!- buildmaster has quit [Remote host closed the connection]
[06:17:43] -!- titus_livius has quit [Remote host closed the connection]
[06:18:13] -!- guys has joined joined
[06:18:23] -!- buildmaster has joined joined
[06:20:03] <buildmaster> pentium4/kirigami2 is broken (says nlopc43).
[06:20:39] -!- titus_livius has joined #archlinux32
[06:21:14] <buildmaster> pentium4/kjs are broken (says nlopc43).
[06:22:22] <buildmaster> pentium4/bluez-qt is broken (says nlopc43).
[06:23:27] <buildmaster> pentium4/kdnssd is broken (says nlopc43).
[06:24:22] <buildmaster> pentium4/kholidays are broken (says nlopc43).
[06:25:04] <buildmaster> pentium4/kwayland is broken (says eurobuild3).
[06:25:37] <buildmaster> pentium4/kimageformats are broken (says nlopc43).
[06:25:39] <buildmaster> pentium4/kplotting is broken (says rechenknecht).
[06:26:48] <buildmaster> pentium4/prison is broken (says nlopc43).
[06:28:48] -!- buildmaster has quit [Remote host closed the connection]
[06:29:00] -!- buildmaster has joined #archlinux32
[06:29:13] <buildmaster> pentium4/modemmanager-qt is broken (says eurobuild3).
[06:30:18] <buildmaster> pentium4/networkmanager-qt is broken (says rechenknecht).
[06:47:59] <buildmaster> pentium4/fluid is broken (says nlopc43).
[07:50:38] <buildmaster> pentium4/mixxx is broken (says rechenknecht).
[08:09:39] <buildmaster> pentium4/prometheus-node-exporter is broken (says rechenknecht).
[08:32:55] <buildmaster> pentium4/prometheus are broken (says rechenknecht).
[08:35:24] <buildmaster> pentium4/qca is broken (says rechenknecht).
[09:00:23] <buildmaster> pentium4/java-sqlite-jdbc is broken (says rechenknecht).
[09:09:55] <buildmaster> pentium4/git-lfs are broken (says rechenknecht).
[09:34:26] <buildmaster> pentium4/sn0int is broken (says rechenknecht).
[09:57:23] <buildmaster> pentium4/vulkan-tools are broken (says rechenknecht).
[10:12:04] <buildmaster> pentium4/gqrx is broken (says rechenknecht).
[10:17:19] <buildmaster> pentium4/falkon is broken (says rechenknecht).
[10:19:35] <buildmaster> pentium4/lrs are broken (says eurobuild1).
[10:30:55] <buildmaster> pentium4/grafana is broken (says rechenknecht).
[10:46:19] <buildmaster> pentium4/ispc is broken (says rechenknecht).
[10:50:49] <buildmaster> pentium4/dovecot is broken (says rechenknecht).
[11:04:49] <buildmaster> pentium4/spamassassin is broken (says nlopc43).
[11:27:03] <buildmaster> pentium4/logstash is broken (says rechenknecht).
[11:32:56] <buildmaster> pentium4/remacs are broken (says eurobuild1).
[11:42:42] <buildmaster> pentium4/libsoup is broken (says rechenknecht).
[11:48:45] <buildmaster> pentium4/arrayfire is broken (says rechenknecht).
[11:59:24] -!- Dimtree has quit [Read error: Connection reset by peer]
[12:07:32] <buildmaster> pentium4/notion is broken (says eurobuild1).
[12:41:52] <thePiGrepper> 'imagemagick: /usr/include/ImageMagick-7/Magick++/Montage.h exists in filesystem (owned by libmagick)
[12:41:55] <thePiGrepper> 'imagemagick: /usr/include/ImageMagick-7/Magick++/Montage.h exists in filesystem (owned by libmagick)
[12:41:59] <thePiGrepper> 'imagemagick: /usr/include/ImageMagick-7/Magick++/Montage.h exists in filesystem (owned by libmagick)
[12:42:06] <thePiGrepper> 'imagemagick: /usr/include/ImageMagick-7/Magick++/Montage.h exists in filesystem (owned by libmagick)
[12:42:09] <thePiGrepper> 'imagemagick: /usr/include/ImageMagick-7/Magick++/Montage.h exists in filesystem (owned by libmagick)
[12:42:13] <thePiGrepper> 'imagemagick: /usr/include/ImageMagick-7/Magick++/Montage.h exists in filesystem (owned by libmagick)
[12:45:29] <thePiGrepper> 'imagemagick: /usr/lib/libMagickWand-7.Q16HDRI.so.6 exists in filesystem (owned by libmagick)' several of this happened during upgrade. I believe the upgrade for libmagick in which it change package's name was the previous one. what do I do?
[12:46:18] <thePiGrepper> oops, sorry for the reposting, I thought it wasnt sending at all! :-(
[12:49:04] <thePiGrepper> ok ,I kinda see what happened. apparently imagemagick wasnt removed when libmagick was added(this one was supposed to be the replacement right?)
[12:49:48] <thePiGrepper> why would that happened?
[12:49:55] <thePiGrepper> *happen
[12:55:26] <buildmaster> pentium4/wireguard-lts are broken (says rechenknecht).
[12:56:28] <thePiGrepper> the only reason I could find for pacman not removing imagemagick is because there are a couple of optional dependencies still needing it at least(in my case, asciidoc and neofetch). would their PKGBUILDs need to be changed to 'libmagick' for the package replacement to be done fully?
[13:08:57] <thePiGrepper> ok, I see now what happened. libmagick is not longer needed because imagemagick package is including now the libs as well? so I just need to remove libmagick and everything will get fixed.
[13:14:52] <thePiGrepper> thx, and sorry again for reposting..
[13:23:39] <thePiGrepper> another issue. apparently texlive-bin still needs icu 63... and there isnt anything on testing for it.
[13:31:23] -!- samantaz__ has joined #archlinux32
[13:33:12] -!- samantaz_ has joined #archlinux32
[13:36:17] -!- samantaz__ has quit [Ping timeout: 268 seconds]
[13:53:41] <thePiGrepper> surf is also requiring icu 63
[13:59:33] <thePiGrepper> correction: it's webkitgtk which requires icu 63 apparently. is it to possible to recompile this against icu 64?
[14:12:47] -!- abaumann has joined #archlinux32
[14:12:47] <buildmaster> Hi abaumann!
[14:12:47] <buildmaster> !rq abaumann
[14:12:48] <phrik> buildmaster: <abaumann> reality is more cruel than cruel jokes about reality.. ;-)
[14:12:54] <abaumann> welcome in a post-icu63 world :-)
[14:13:12] <abaumann> I rescheduled texlive and surf..
[14:13:42] <thePiGrepper> lol. thx
[14:14:01] <thePiGrepper> i think it's actually webkitgtk itself though
[14:14:12] <thePiGrepper> I may be wrong
[14:14:45] <abaumann> yeah. I'm trying to get to a more scripted way to get to those broken dependencies.
[14:14:53] <abaumann> hunting them down by hand is not fun
[14:15:34] <thePiGrepper> do you happen to know if webkitgtk takes too long to build? I might have started something that wont end soon on my n270 old atom system lol
[14:16:02] <abaumann> if it's just a gtk wrapper to webkit, then it is fast.
[14:16:28] <abaumann> OTOH modern software development embedds everything and doesn't use system libraries, so webkitgtk might build its own webkit.
[14:16:33] <thePiGrepper> I happened to remember that webkit itself is huge to build
[14:16:37] <abaumann> yep.
[14:17:10] <thePiGrepper> it took hours to build webkit in my main system (ryzen 7..)
[14:17:18] <thePiGrepper> that's just absurd
[14:17:48] <abaumann> yep. now you know, why there is currently a backlog of 5000 packages on the build slaves. :-)
[14:18:06] <abaumann> adding new architecture also didn't help speeding things up. ;-)
[14:18:45] <thePiGrepper> what are the specs of the slaves?
[14:19:44] <abaumann> I can just tell you something about mine: tyzoid-srv0-vm486 is a virtual machine on tyzoids server, eurobuild1 and eurobuild3 are some 10 year old 64bit machines.
[14:19:51] <abaumann> though they have both Intel server boards in them
[14:20:05] <abaumann> 8 GB eurobuild, 2 GB RAM eurobuild3
[14:20:14] <abaumann> *eurobuild3 first, eurobuild1 second
[14:21:06] <thePiGrepper> I see.
[14:22:22] <thePiGrepper> that backlog look a bit more than a pressing issue. it kinda only gets longer
[14:23:36] <thePiGrepper> cmake says that 'something' is at 13%, hmm, so I dont know. Im kinda curious now
[14:24:36] <thePiGrepper> btw, something weird happened with cmake. CMakeLists.txt asked for sse2
[14:24:54] <abaumann> on pentium4 or i686?
[14:25:20] <abaumann> I think, when adding -march=pentium4 cmake starts to see SSE2..
[14:25:30] <thePiGrepper> Im using i686, however I might have just used asp32 checkout and entered trunk instead, so probably my fault
[14:26:13] <thePiGrepper> I copied makepkg.conf to the current dir and changed it to atom, both -march and -mtune
[14:26:17] <abaumann> maybe SSE2 is the default now, so it has to be removed for i686 builds.. that's something I want to look in when problems are solved..
[14:26:21] <thePiGrepper> so it should have worked either way right?
[14:26:59] <abaumann> how are you building?
[14:27:04] <thePiGrepper> abaumann: probably. you'd need to check for how many packages make an explicit requirement for it
[14:27:28] <abaumann> with makepkg on the host the settings in /etc/makepkg.conf are relevant.
[14:27:52] <abaumann> If you use archbuild, then the corresping makepkg*.conf in /usr/local/share (wherever they get installed) are taken.
[14:27:56] <thePiGrepper> makepkg from 'asp32 checkout webkit2gtk' at trunk(changed arch to i686)
[14:28:10] <abaumann> ah. ok
[14:28:12] <thePiGrepper> yeah but I added --config ./makepkg.conf
[14:28:16] <abaumann> a ok.
[14:28:21] <abaumann> very safe :-)
[14:29:15] <thePiGrepper> that's the weird part, the cmake still said that I didnt have sse2, which isnt what the makepkg.conf said(I modified it), neither what the lscpu said
[14:29:17] <abaumann> I don't like this over-optimization of things. Usually it should be sufficient to use i486 binaries, then for important libraries like glibc, SDL, Qt, gtk use the best optimizations you can get.
[14:29:31] <abaumann> There is an issue I'm not quite sure, how the ABI changes, but actually, it should not.
[14:30:01] <abaumann> mmh.
[14:30:11] <abaumann> depends how cmake is checking for SSE2.. hang on a sec..
[14:30:37] <abaumann> *abaumann is checking out webkit2gtk and has a look at some funcky cmake code..
[14:30:45] <thePiGrepper> abaumann: I agree. optimize *every* final application is too much of a hassle for just a bit of optimization
[14:31:38] <thePiGrepper> there's this line 'if (NOT SSE2_SUPPORT_FOUND)'
[14:31:38] <abaumann> OHOH it's sometimes hard to tweak just part of a system, running everything under the same optimize flags maybe gives more stable results..
[14:31:53] <abaumann> I'm not a Gentoo-guy wanting to optimize the last bit of silicon.. :-)
[14:32:12] <abaumann> ..I'm more the "let's think and write good software with good algorithms" guy
[14:32:42] <thePiGrepper> abaumann: exactly, and even if you were, not even gentoo makes that the default. everyone can pick their PKGBUILD and set whatever flag they want and rebuild if they like to
[14:33:02] <abaumann> of course.
[14:33:09] <thePiGrepper> the issue is with it being the upstream's default as well
[14:33:33] <thePiGrepper> and arch having the rule of trying to not mod too much the upstream config
[14:34:09] <abaumann> yeah. lately I got the feeling that you have to change more and more.
[14:34:31] <abaumann> cherry pick git patches because of sloppy releasing up-up-stream
[14:34:34] <thePiGrepper> in this case for instance, what could you do without making an issue on webkit2gtk upstream? make a sed line on CMakeList.txt to comment those 4,5 lines?
[14:34:37] <thePiGrepper> lol
[14:35:01] <abaumann> most likely, yes.
[14:35:08] <thePiGrepper> did you check the sse2 finding method?
[14:35:20] <abaumann> I managed to check out the package now.. :-)
[14:35:22] <thePiGrepper> what do you think? Im not used to cmake
[14:35:34] <abaumann> * abaumann has some severe disk space issues on his eeepc..
[14:35:58] <thePiGrepper> isnt it nice to work on a netbook?
[14:36:08] <abaumann> sure.
[14:36:29] <abaumann> but 32GB on the SD card feels a little bit tight at the moment.
[14:36:35] * thePiGrepper writes this message on a latitude 2100 ;-)
[14:36:56] <abaumann> FindSSE2.cmake
[14:37:20] <thePiGrepper> wow, Im just surprised that the sd card is supporting that capacity at all lol
[14:37:28] <thePiGrepper> *sd card reader
[14:37:33] <abaumann> ah. it tries to build and run some code with __m128d
[14:37:37] <thePiGrepper> let me see
[14:37:53] <abaumann> you can always do set(SSE2_SUPPORT_FOUND TRUE)
[14:38:07] <abaumann> in the main CMakeLists.txt
[14:38:09] <abaumann> if (WTF_CPU_X86)
[14:38:10] <abaumann> include(FindSSE2)
[14:38:10] <abaumann> if (NOT SSE2_SUPPORT_FOUND)
[14:38:10] <abaumann> message(FATAL_ERROR "SSE2 support is required to compile WebKit")
[14:38:12] <abaumann> endif ()
[14:38:14] <abaumann> endif ()
[14:38:28] <abaumann> elseif (LOWERCASE_CMAKE_SYSTEM_PROCESSOR MATCHES "(x64|x86_64|amd64)") set(WTF_CPU_X86_64 1)
[14:38:31] <abaumann> elseif (LOWERCASE_CMAKE_SYSTEM_PROCESSOR MATCHES "(i[3-6]86|x86)")
[14:38:33] <abaumann> oh sweet :-)
[14:38:49] <abaumann> cmake is such a broken make system, so everybody starts to write his own compiler sniffers.
[14:39:38] <abaumann> so if on IA32, it tries to sniff for SSE2 by compiling a piece of software.
[14:39:43] <thePiGrepper> it's kinda weird that so many ppl say it's the best one, does that mean that everyother makesystem is bad af? lol
[14:39:54] <thePiGrepper> what's that __m128d?
[14:40:48] <abaumann> a compiler intrinsic for a 128-bit MMX register.
[14:41:07] <abaumann> that's what SSE2 adds and this makes code really fast.
[14:41:33] <abaumann> it should NEVER be used outside of the compiler, as it is used by the compiler to generate efficient code.
[14:41:53] <abaumann> So expect it to be used soon everywhere instead of int, long ;-)
[14:42:57] <abaumann> gcc -m32 -march=pentium4 -o test test.c
[14:42:57] <abaumann> In file included from /usr/include/features.h:474,
[14:42:57] <abaumann> from /usr/include/bits/libc-header-start.h:33,
[14:42:57] <abaumann> from /usr/include/stdlib.h:25,
[14:42:57] <abaumann> from /usr/lib/gcc/x86_64-pc-linux-gnu/8.3.0/include/mm_malloc.h:27,
[14:43:00] <abaumann> from /usr/lib/gcc/x86_64-pc-linux-gnu/8.3.0/include/xmmintrin.h:34,
[14:43:03] <abaumann> from /usr/lib/gcc/x86_64-pc-linux-gnu/8.3.0/include/emmintrin.h:31,
[14:43:06] <abaumann> from test.c:1:
[14:43:08] <abaumann> /usr/include/gnu/stubs.h:7:11: fatal error: gnu/stubs-32.h: No such file or directory
[14:43:11] <abaumann> # include <gnu/stubs-32.h>
[14:43:14] <abaumann> ^~~~~~~~~~~~~~~~
[14:43:16] <abaumann> compilation terminated.
[14:43:30] <abaumann> aha. so, if you extract the test as test.c and then try to compile it with the 64-bit gcc with proper march, the compilation fails..
[14:43:56] <abaumann> this test doesn't work
[14:44:02] <abaumann> and it doesn't cross-compile
[14:44:04] <abaumann> *sigh8
[14:45:13] <abaumann> mmh. this gives me some severe pondering, how we will build this package pentium4/sse2 optimized..
[14:45:28] <abaumann> we cannot hack all the funny build systems
[14:46:27] <abaumann> -m32 support not working in gcc, so I need the multiarch one?
[14:47:24] <abaumann> aha. yes.
[14:47:38] <abaumann> gcc -march=pentium4 -o test test.c on 32-bit Archlinux32 with 32-bit gcc works.
[14:47:56] <abaumann> so, on 64-bit we have to install the multilib gcc..
[14:54:32] <thePiGrepper> well that last part was already known right? isnt it always that way? if you want to build for 32-bit on a 64-bit system
[14:55:39] <abaumann> yes. it's only an issue when using the makepkg straw in the build slave, otherwise a 32-bit gcc is installed in to chroot of the slave
[14:55:40] <thePiGrepper> something is funny though, I just commented out that part of the cmakelist.txt and, at least until now(41%), it's working just fine. so then, it's a problem with the detection function?
[14:56:22] <abaumann> you can copy the test into a test.c from this FindSSE2 function and try to compile it by hand and play with the march flags.
[14:56:26] <thePiGrepper> it works only on a 64bit gcc maybe?
[14:56:40] <thePiGrepper> I will do that
[14:57:46] <abaumann> on 64-bit gcc (no multilib) I got an gnu-32-stub.h error of sorts, which means, I don't have the 32-bit part of gcc. It compiled find on my 32-bit eeepc Archlinux32 (i686)
[15:01:07] <thePiGrepper> when I do a simple 'gcc test.c -o test' it returns some errors like '/usr/lib/gcc/i686-pc-linux-gnu/8.3.0/include/emmintrin.h:171:1: error: inlining failed in call to always_inline ‘_mm_storeu_pd’: target specific option mismatch
[15:01:11] <thePiGrepper> '
[15:02:15] <thePiGrepper> but I believe this warning has more to do with the issue: warning: SSE vector return without SSE enabled changes the ABI [-Wpsabi]
[15:02:48] <thePiGrepper> do you know what that is?
[15:03:28] <thePiGrepper> oh, ok I understand now
[15:03:56] <thePiGrepper> the default gcc flags are not tuned for this, because it's the i686 build, is that it?
[15:04:14] <thePiGrepper> march=pentium4 works here ofc
[15:04:43] <thePiGrepper> as well as -march=atom obviously
[15:06:53] <abaumann> yep it should.
[15:07:07] <thePiGrepper> then, to summarize, the check always fails because of the gcc version and default cflags on the i686 build right?
[15:07:16] <abaumann> -march=pentium4 should be enough. There is also a -msse2 flag, but this one croaks when used with the wrong architecture.
[15:07:27] <abaumann> yep.
[15:07:29] <abaumann> I think so.
[15:07:54] <abaumann> I'm currently testing it on a build slave with pentium4 optimization..
[15:08:03] <thePiGrepper> yeah, it makes sense to me as well.
[15:08:20] <thePiGrepper> oh, there it should work without modification, if we're right ofc
[15:09:45] <thePiGrepper> btw, are you using your eeepc on the pentium4 build of arch32, or the i686 one only?
[15:09:58] <abaumann> I'm still on i686.
[15:10:06] <thePiGrepper> yeah, Im kinda asking the old dogfooding question I guess lol
[15:10:08] <abaumann> I didn't upgrade yet. But my eeepc 701 can do SSE2 :-)
[15:10:22] <thePiGrepper> I want to upgrade, it makes sense. but too soon right?
[15:10:38] <thePiGrepper> it's that damn egg and chicken problem lol
[15:10:49] <abaumann> hard to say. the icu64//icu63 equally hit both architectures, I think
[15:11:32] <thePiGrepper> well the build infrastructure is the same basically, that makes sense. but how about the number of packages issues on pentium4, being the new one and all
[15:12:05] <abaumann> https://archlinux32.org
[15:12:08] <phrik> Title: Arch Linux 32 - List of Package Builds (at archlinux32.org)
[15:12:08] <thePiGrepper> is it way higher than i686? like it was for i486 at the beginning(I havent checked lately)
[15:12:14] <thePiGrepper> thx
[15:12:17] <abaumann> would suggest that pentium4 has much more errors than i686, but..
[15:12:31] <abaumann> ..pentium4 is also built with a higher probability..
[15:12:50] <thePiGrepper> higher probability??
[15:12:51] <abaumann> pentium4 is a brutal rewrite of i686 packages.
[15:12:56] <thePiGrepper> oh
[15:12:57] <abaumann> i486 was a complete bootstrap
[15:13:08] <abaumann> yeah. not on purpose.. by accident.. :-)
[15:13:23] <abaumann> i686: SSE2 support is required to compile WebKit
[15:13:30] <abaumann> log.pentium4:-- Performing Test HAVE_SSE2_EXTENSIONS - Success
[15:13:35] <abaumann> yes. this looks about right.
[15:13:40] <thePiGrepper> well, logically thinking, it would make sense for pentium4 to be easier in principle than i686, because it's closer to upstream, right?
[15:14:14] <abaumann> not because of being closer to upstream.. but because SSE2 is a requirement since all early 64-bit CPUs
[15:14:31] <abaumann> so less funny micro-optimizations in the code can break the build.
[15:14:40] <thePiGrepper> nice. it worked as expected. that also means the cause is probably that one
[15:14:51] <abaumann> like webkit2gtk. Not sure, but enabling non-SSE2 builds may mean some work..
[15:14:57] <thePiGrepper> well, ofc, I actually meant that.
[15:15:01] <abaumann> :-)
[15:15:50] <abaumann> https://buildmaster.archlinux32.org
[15:15:56] <abaumann> mmh. cmake fails to build on pentium4?
[15:15:58] <abaumann> really?
[15:16:41] <thePiGrepper> now that we have pentium4, it seems less of a good idea to have an atom-optimized build I guess..
[15:16:48] <abaumann> yep..
[15:17:01] <abaumann> but we will not add an 'atom' architecture. ;-)
[15:17:14] <thePiGrepper> it would still be nice because the match=atom optimizes for the architecture as well
[15:17:38] <abaumann> I should check once some Gentoo or similar benchmarks.
[15:17:47] <abaumann> how much faster is it really..
[15:17:58] <thePiGrepper> really? that was decided explicitly? I didnt know. at the very least there are a lot of netbook users here
[15:18:11] <thePiGrepper> probably some benchmarking is in order, yes
[15:18:23] <thePiGrepper> I would like for a benchmark that doesnt take forever on old hardware
[15:18:30] <abaumann> exacly.
[15:18:54] <abaumann> this project has some top priorities: a) be as close to Archlinux 64-bit as possible, b) have current software.
[15:19:15] <abaumann> So if those goals are violated by too many feature creep or architectures to support, then we fail.
[15:28:35] <thePiGrepper> I see. one question though. i686 is the old 32-bit arch as much as I know, i486 was needed for older architectured which didnt support .. I dont remember what right now but there were several users for that one, including yourself ofc. pentium4 is for ppl who have sse2 and want to be able to use it basically. the question is how many ofthose ppl are NOT ppl using atom systems, I know that making
[15:28:41] <thePiGrepper> an atom-optimized build is bad because it *excludes* users, whereas making a pentium4 one gives an upgrade to more. in that sense, yeah you're right.. hmm , apparently there wasnt a question after all. ;-)
[15:30:47] <thePiGrepper> it it's and eitheir or situation between pentium4 and atom, ofc pentium4 wins. it would be nice to have an atom one, but the current infrastructure probably will suffer from one more architecture
[15:30:55] <thePiGrepper> that much I understand..
[15:32:55] <abaumann> basically, yes. :-)
[15:38:10] -!- alexandre9099 has quit [Quit: ZNC 1.7.3 - https://znc.in]
[15:40:49] -!- alexandre9099 has joined #archlinux32
[15:40:56] <buildmaster> pentium4/ghc is broken (says rechenknecht).
[15:44:38] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[15:49:16] <elibrokeit> the reason gentoo "supports" atom optimization is because they give you a lot of PKGBUILDs and tell you to compile your own distro :)
[15:50:01] <elibrokeit> seems like an excessive amount of work for a distro to build though -- users can recompile performance-critical packages from the ABS...
[15:50:48] -!- thePiGrepper has joined #archlinux32
[15:54:46] -!- deep42thought has joined #archlinux32
[15:54:47] <buildmaster> Hi deep42thought!
[15:54:47] <buildmaster> !rq deep42thought
[15:54:47] <phrik> buildmaster: <deep42thought> good night, python - cu tomorrow python2!
[15:54:51] <deep42thought> Hi all!
[15:54:57] * deep42thought reads the logs
[16:08:30] <deep42thought> abaumann: I usually parse https://archlinux32.org and feed it into prioritize-build-list and/or db-update (but that can/should be automatable)
[16:08:39] <deep42thought> an I'm against yet another new architecture
[16:08:56] <deep42thought> however, I'm open to changing the attributes of pentium4 - but reluctant ;-)
[16:12:40] <slacka123> deep42thought: Yes, if anything pentum4 could be expanded. 3 arch is already a lot
[16:17:51] -!- thePiGrepper has quit [Ping timeout: 268 seconds]
[16:24:10] -!- yans has joined #archlinux32
[16:25:25] -!- thePiGrepper has joined #archlinux32
[16:25:58] <abaumann> hi deep42thought
[16:26:29] <deep42thought> Hi abaumann!
[16:27:13] <deep42thought> what do you need? prioritizing packages which have broken linking?
[16:27:51] <abaumann> yeah. I wanted to test this by installing a lxc image with all possible archlinux packages
[16:28:08] <deep42thought> the database (should) have it all
[16:28:16] <thePiGrepper> does anyone happens to know of any benchmarks suited for low-spec systems? if I run PTS on my netbook I probably could just stop using it for a season
[16:28:18] <deep42thought> "pentium4/community/strawberry-0.5.2-2.0-pentium4 depends on libcdio.so.18 which is not provided by any package."
[16:28:21] <deep42thought> something like that
[16:33:50] <elibrokeit> thePiGrepper: see backlog while you detached:
[16:33:50] <elibrokeit> <elibrokeit> the reason gentoo "supports" atom optimization is because they give you a lot of PKGBUILDs and tell you to compile your own distro :)
[16:33:50] <elibrokeit> <elibrokeit> seems like an excessive amount of work for a distro to build though -- users can recompile performance-critical packages from the ABS...
[16:36:44] <abaumann> thePiGrepper: thanks, we keep all the logs at https://mirror.archlinux32.org :-)
[16:36:45] <phrik> Title: Index of /irc-logs/ (at mirror.archlinux32.org)
[16:39:55] -!- thePiGrepper has quit [Ping timeout: 248 seconds]
[16:45:13] -!- guys has quit [Quit: A random quit message]
[16:50:09] -!- abaumann has quit [Quit: leaving]
[16:56:08] -!- deep42thought has quit [Quit: Leaving.]
[17:30:18] <buildmaster> pentium4/dvdauthor is broken (says rechenknecht).
[17:31:37] <buildmaster> pentium4/a2ps are broken (says eurobuild3).
[17:32:34] <buildmaster> pentium4/chafa is broken (says eurobuild3).
[17:32:59] <buildmaster> pentium4/cozy-stack is broken (says eurobuild3).
[17:33:43] <buildmaster> pentium4/cups-filters are broken (says eurobuild3).
[17:34:35] <buildmaster> pentium4/converseen is broken (says rechenknecht).
[17:35:02] <buildmaster> pentium4/raylib is broken (says eurobuild3).
[17:35:27] <buildmaster> pentium4/virtualbox-modules-arch is broken (says eurobuild3).
[17:37:27] -!- buildmaster has quit [Remote host closed the connection]
[17:37:39] -!- buildmaster has joined #archlinux32
[18:00:24] -!- Dimtree has joined #archlinux32
[18:01:15] <buildmaster> pentium4/yq is broken (says eurobuild3).
[18:08:12] -!- thePiGrepper has joined #archlinux32
[18:11:09] <thePiGrepper> elibrokeit: well, yeah that makes sense. but that explanation could apply to other choices as well. my point being, that it's more a logistics problem as always, it's too much of a burden to have too many arches.
[18:12:10] <thePiGrepper> abaumann: btw, still compiling webkit2gtk lol (48%)
[18:23:44] <buildmaster> i686/wlc is broken (says eurobuild3).
[18:34:27] <buildmaster> pentium4/amavisd-new is broken (says eurobuild3).
[19:13:55] <buildmaster> i686/arch-audit is broken (says rechenknecht).
[19:24:26] <buildmaster> i686/pyrit is broken (says rechenknecht).
[19:33:01] -!- aliemjay has joined #archlinux32
[19:43:11] -!- aliemjay has quit [Quit: Konversation terminated!]
[19:43:53] -!- aliemjay has joined #archlinux32
[19:49:16] <aliemjay> girls: Hi, I see that qt/kde fail to build because of a dependency loop again!
[19:49:28] <girls> :-(
[19:49:53] <aliemjay> But now the problem is with qt5-tools and qt5-webkit
[19:50:08] <aliemjay> The webkit should be build first
[19:55:35] -!- deep42thought has joined #archlinux32
[19:55:36] <buildmaster> Hi deep42thought!
[19:55:36] <buildmaster> !rq deep42thought
[19:55:36] <phrik> buildmaster: <deep42thought> ah, maybe, borg is borged, too
[19:55:43] <deep42thought> hmm, qt5-webkit is not in a loop ...
[19:56:15] <deep42thought> I'm forcing it anyways ...
[19:56:35] <aliemjay> It is a make Dep. To qt5-tools
[19:56:40] <deep42thought> yes
[19:56:47] <deep42thought> makedepends are non-critical
[19:56:56] <deep42thought> they are satisfied if any version is available ...
[19:57:17] <deep42thought> the thought is: you do not need the newest git to check out the source - any git will do
[19:57:21] <aliemjay> So that may the problem
[19:57:28] <deep42thought> however, this is wrong in some cases
[19:57:47] <deep42thought> e.g. when git has linking issues, then /any/ git will not do ;-)
[19:57:54] <deep42thought> (just as an example)
[20:02:49] <buildmaster> i686/gnu-apl is broken (says eurobuild3).
[20:04:44] -!- aliemjay has quit [Read error: Connection reset by peer]
[20:06:24] -!- ofara_ has quit [Quit: ofara_]
[20:07:49] -!- ofara_ has joined #archlinux32
[20:11:05] -!- samantaz_ has quit [Remote host closed the connection]
[20:11:55] -!- samantaz_ has joined #archlinux32
[20:23:23] -!- samantaz_ has quit [Ping timeout: 248 seconds]
[20:24:30] -!- samantaz_ has joined #archlinux32
[20:28:09] -!- aliemjay has joined #archlinux32
[20:29:25] <buildmaster> pentium4/xerces-c is broken (says eurobuild3).
[20:31:16] -!- ofara_ has quit [Ping timeout: 246 seconds]
[20:32:50] -!- thePiGrepper has quit [Ping timeout: 255 seconds]
[20:34:05] -!- samantaz_ has quit [Remote host closed the connection]
[20:34:19] -!- ofara_ has joined #archlinux32
[20:34:40] -!- samantaz_ has joined #archlinux32
[20:36:53] -!- thePiGrepper has joined #archlinux32
[20:40:27] -!- ofara_ has quit [Ping timeout: 245 seconds]
[20:42:47] -!- abaumann has joined #archlinux32
[20:42:47] <buildmaster> Hi abaumann!
[20:42:47] <buildmaster> !rq abaumann
[20:42:48] <phrik> buildmaster: <abaumann> I would bet systemd is now also generating pam configuration on the fly
[20:42:57] <deep42thought> good evening, abaumann
[20:43:02] <abaumann> mmh. this is not good. eurobuild1 lost the linux kernel build after hours with 'You do not build anything currently - abort whatever you are doing.'
[20:43:05] <abaumann> hi deep42thought
[20:43:10] <abaumann> the question is why?
[20:43:26] <deep42thought> I pushed an update
[20:43:30] <deep42thought> e.g. new checksums/config
[20:43:33] <abaumann> aha.
[20:43:38] <abaumann> in that case.
[20:43:49] <abaumann> I thought there was something wrong with the new slave. :-)
[20:44:12] <deep42thought> sry for nullifying your work :-/
[20:44:26] <abaumann> it's not my work. slaves are supposed to do brainless work. ;-)
[20:44:32] <deep42thought> !grab abaumann
[20:44:33] <phrik> deep42thought: Tada!
[20:45:09] <abaumann> the machine goes into cleaning mode now anyway.. it makes some weird noices (cpu fan)
[20:45:16] <abaumann> *noises
[20:45:24] <deep42thought> it physically cleans itself?
[20:45:33] <deep42thought> that sounds like a feature that would sell to me
[20:45:35] <abaumann> yes. like a cleaning tape in the tape robot ;-)
[20:46:42] -!- samantaz__ has joined #archlinux32
[20:48:13] <deep42thought> deos it have one of those vacuum cleaner robots inside the case?
[20:48:23] <abaumann> yes, tiny ones. ;-)
[20:48:28] <deep42thought> cool
[20:48:34] <abaumann> :-)
[20:48:47] -!- samantaz_ has quit [Ping timeout: 245 seconds]
[20:49:02] <deep42thought> btw: sry, your scheduling feature is not yet implemented and I will most probably not do it this evening
[20:49:12] <abaumann> np
[20:49:13] <deep42thought> but I noticed, that seed-build-list can /almost/ do what you need
[20:49:21] <deep42thought> it works if the job is not yet on the build list
[20:49:47] <deep42thought> then you can use -a to schedule it at all, -j to insert it with highest priority and -d to do the same for all dependencies
[20:49:56] <deep42thought> e.g. seed-build-list -ajd
[20:50:07] <abaumann> ah cool. this is a powerful tool.
[20:50:21] <abaumann> it has the potential to confuse the buildmaster big times.. :-)
[20:50:24] <deep42thought> yes, but -a does not schedule stuff that is already on the build list
[20:50:30] <abaumann> ..or rebuild 3*10500 packages in the worst case.
[20:50:30] <deep42thought> so it is of no use in your case
[20:50:47] <deep42thought> you can always do a dry-run with -n first ;-)
[20:50:51] <deep42thought> that's what I usually do
[20:51:14] <deep42thought> ... at least for those auto-targets
[20:51:59] <deep42thought> btw: I forced qt5-webkit onto my slave - looks good so far
[20:52:07] <deep42thought> no idea, why it was not scheduled in the first place
[20:52:08] <abaumann> I forced evolution-data-server and webkit2gtk
[20:52:24] <deep42thought> ah, well, I _have_ an idea
[20:52:51] <deep42thought> there are 768 packages "buildable", so the probability to start exactly that is only 1/384 ...
[20:53:39] <buildmaster> abaumann: I'm never confused
[20:53:41] <abaumann> yeah. I usually force stuff, so I can steer it a little bit better.
[20:53:46] <abaumann> lol
[20:54:09] <buildmaster> I might be insane, but never confused ... ;-)
[20:54:21] <abaumann> !grab buildmaster
[20:54:22] <phrik> abaumann: Tada!
[20:55:12] <deep42thought> since the buildmaster is the one who requests quotes of joining devs, we will never automatically see this quote ;-)
[20:55:22] <deep42thought> !rq buildmaster
[20:55:23] <phrik> deep42thought: <buildmaster> I might be insane, but never confused ... ;-)
[21:05:21] <abaumann> I would love to have a pacman options like 'pacman --install-EVERYTHING'...
[21:05:30] <deep42thought> there is
[21:05:34] <abaumann> huh?
[21:05:59] <deep42thought> pacman -Sqs | pacman -S -
[21:06:13] <abaumann> ah. thanks. yes.
[21:06:23] <deep42thought> you might get conflicts, though
[21:06:34] <abaumann> yes, that's exactly what I want to see.
[21:06:55] <abaumann> this is an excellent quality assurance tool
[21:08:08] <deep42thought> I think, running pacman to install all packages on archlinux32 is still faster than running a regular "apt-get upgrade" on my alix
[21:08:44] <abaumann> there are less packages in Arch than in Debian, to be fair.
[21:08:57] <deep42thought> but I do not have all of them installed ;-)
[21:09:24] <abaumann> especially not on the alix. ;-)
[21:09:44] <abaumann> a big storage array next to an alix box doesn't look cool..
[21:10:06] <deep42thought> attached via a looooong usb cable ;-)
[21:11:57] <elibrokeit> Well, there are packages that conflict/provide each other. ;) So that would definitely not work without filtering.
[21:12:05] <abaumann> extra/perl-mail-spf 2.9.0-4.1 Perl module that provides SPF support
[21:12:05] <abaumann> community/perl-mail-spf-query 1.999.1-9.1 Perl module that provides SPF support
[21:12:08] <abaumann> mmh.
[21:12:10] <elibrokeit> But yeah, Debian is sloooooow
[21:12:13] <deep42thought> elibrokeit: doesn't pacman ask about those?
[21:12:45] <deep42thought> or is this only for updates?
[21:12:47] <abaumann> yeah, but only one by one at the end.
[21:12:51] <deep42thought> :-/
[21:13:23] <elibrokeit> That would still be a lotta questions... And I'm trying to remember if it asks when you try to install two conflicting packages, or only when you install one that conflicts with a pre-existing one.
[21:14:21] <abaumann> libreoffice-fresh and libreoffice-still, well, my approach has a flaw, I think. :-)
[21:14:23] <elibrokeit> I have a Debian chroot I created for weird testing purposes. Going to it and running an update/upgrade takes several hours to unpack packages and install them
[21:14:38] <elibrokeit> Slooooooooowly
[21:15:32] <elibrokeit> It's amazing because pacman -Syu is so blindingly fast...
[21:15:39] <deep42thought> yes
[21:15:50] <deep42thought> I was always wondering, what debian uses the extra cycles for
[21:16:07] <deep42thought> but then I noticed, it offers quite elaborate solutions to fix conflicts
[21:16:50] <abaumann> for i in $(pacman -Sqs); do pacman -S --needed --noconfirm $i; done
[21:16:57] <abaumann> ok, dummie approach..
[21:17:33] <deep42thought> for i in $(pacman -Sqs | shuf); do pacman -S --needed --noconfirm $i; done
[21:18:00] <abaumann> not reproducable :-)
[21:18:06] <deep42thought> yeah
[21:18:13] <deep42thought> ok, yours is better :-)
[21:18:23] -!- thePiGrepper has quit [Ping timeout: 246 seconds]
[21:18:53] <abaumann> mmh. executing all the pacman hooks on each package is not really to be fast..
[21:19:43] <abaumann> reboot
[21:19:43] <abaumann> System has not been booted with systemd as init system (PID 1). Can't operate.
[21:19:46] <abaumann> Failed to connect to bus: Host is down
[21:19:49] <abaumann> Failed to talk to init daemon.
[21:19:52] <abaumann> on the 64-bit host hosting my slave.
[21:20:01] <abaumann> something is definitely fishy with dbus/systemd
[21:21:08] <abaumann> *abaumann just realized that the tvheaded backend is also on this machine, so his tv went blank in the middle of the Snooker world championship finals.
[21:21:52] <deep42thought> oh, who's in the final?
[21:21:59] <abaumann> and of couse. kodi just hangs now..
[21:22:05] <abaumann> Higgins, Trump
[21:23:25] <elibrokeit> abaumann: sure, because hooks were designed to prevent duplicate work. ;)
[21:24:28] <abaumann> kodi: line 219: 296 Segmentation fault (core dumped)
[21:24:41] <abaumann> the you of watching tv on bleeding edge operating systems :-)
[21:25:22] <abaumann> *joy
[21:25:52] <abaumann> ERROR: SQL: [Epg12.db] SQLite error SQLITE_CORRUPT (database disk
[21:25:52] <abaumann> image is malformed)
[21:26:14] <abaumann> if analog tv would have been engineered nowadays with all this software nonsense, it would explode every second weekend..
[21:27:56] <deep42thought> !grab abaumann
[21:27:57] <phrik> deep42thought: Tada!
[21:28:18] <deep42thought> I can't find a livestream that runs without flash either, so I'm just going to bed now ...
[21:28:24] <abaumann> yeah.
[21:28:30] <abaumann> I'm going to my old analog tv..
[21:28:35] <abaumann> this one works in 2 seconds..
[21:28:37] <abaumann> cu
[21:28:40] <deep42thought> cu
[21:28:49] -!- deep42thought has quit [Quit: Leaving.]
[21:28:51] -!- abaumann has quit [Remote host closed the connection]
[21:37:08] -!- ubuntuxp has joined #archlinux32
[22:16:15] -!- ofara_ has joined #archlinux32
[22:22:40] <buildmaster> i686/evolution-data-server is broken (says eurobuild3).
[23:18:40] -!- samantaz__ has quit [Remote host closed the connection]
[23:19:30] -!- samantaz__ has joined #archlinux32
[23:20:08] -!- samantaz__ has quit [Max SendQ exceeded]
[23:21:45] -!- samantaz__ has joined #archlinux32
[23:47:46] <buildmaster> pentium4/evolution-data-server is broken (says eurobuild3).
[23:51:17] -!- thePiGrepper has joined #archlinux32
[23:58:56] <buildmaster> i686/webkit2gtk is broken (says eurobuild3).
[23:59:10] -!- thePiGrepper has quit [Ping timeout: 246 seconds]