#archlinux32 | Logs for 2018-11-09

Back
[00:00:06] <autofsckk> but is there a video player that i could use meanwhile or instead of the other ones?
[00:01:48] <deep42thought> downgrade, downgrade, downgrade :-/
[00:09:43] -!- guys has joined #archlinux32
[00:22:46] -!- guys has quit [Ping timeout: 244 seconds]
[00:26:28] -!- woshty has quit [Ping timeout: 272 seconds]
[00:41:48] -!- deep42thought has quit [Quit: Leaving.]
[01:04:03] <autofsckk> well i downgraded smplayer 3 versions but still does not work :/ is there something else i can do to make it work? or some other video player ?
[01:20:04] -!- DCyrax has quit [Remote host closed the connection]
[02:48:52] <finsternis> autofsckk: you didn't have to downgrade smplayer, it doesn't work since it needs icu 63
[02:49:43] <finsternis> as video player I use mpv
[03:30:57] <buildmaster> i686/libfbclient is broken (says buildknecht).
[03:49:28] <autofsckk> i also have mpv but doesnt work
[04:53:48] -!- thePiGrepper has joined #archlinux32
[05:12:41] -!- guys has joined #archlinux32
[05:17:55] <autofsckk> what should i downgrade so i can use a movie player? please help?
[05:18:34] <autofsckk> in order for the movie player to work i have tu upgrade icu to 63? but firefox and thunderbird wont work? or is there a work around for that? link it maybe?
[05:59:11] <buildmaster> i686/haskell-fclabels are broken (says rechenknecht).
[06:10:38] -!- autofsckk has quit [Quit: leaving]
[07:11:52] -!- thePiGrepper has quit [Ping timeout: 268 seconds]
[07:18:28] -!- thePiGrepper has joined #archlinux32
[07:25:15] -!- DepositePirate has quit [Ping timeout: 256 seconds]
[08:04:19] -!- deep42thought has joined #archlinux32
[08:04:20] <buildmaster> Hi deep42thought!
[08:04:20] <buildmaster> !rq deep42thought
[08:04:21] <phrik> buildmaster: <deep42thought> please don't go insane - I think I can't fix you like I can fix the buildmaster
[08:04:56] <deep42thought> autofsckk: yeah, most programs should be linked against icu 63, but some (fireforx, thunderbird, some media player(?), ...) are not
[08:05:05] <deep42thought> but are instead linked against icu 62
[08:08:14] -!- pulec has joined #archlinux32
[08:20:12] -!- thePiGrepper has quit [Ping timeout: 252 seconds]
[08:28:45] <buildmaster> i686/haskell-dlist is broken (says rechenknecht).
[08:48:25] <buildmaster> i686/libcryptui is broken (says nlopc46).
[08:57:06] -!- abaumann has joined #archlinux32
[08:57:07] <buildmaster> Hi abaumann!
[08:57:07] <buildmaster> !rq abaumann
[08:57:07] <phrik> buildmaster: <abaumann> I gave up with cross-compiliation. So I downgraded a 64-bit installation to i486 and applied some hacks.
[08:57:15] <abaumann> hi deep42thought
[08:58:01] <abaumann> I'm thinking: as we cannot guarantee a fast rebuild after an ABI change in icu or boost, it might be an idea to always have an icu62, boost61 package around, just to keep the old shared libraries alive?
[08:58:40] <abaumann> /usr/include/boost would always point to the newest version, but new software should be built against the new version anyway.
[08:59:09] <deep42thought> that's a good idea
[08:59:17] <abaumann> symlinking only works, if old versions of symbols stick around, the versioning schema in boost and icu is weird..
[08:59:29] <deep42thought> how do we implement this in the cleanest way (e.g. with the least manual interaction)?
[08:59:58] <deep42thought> and we need to add tons of versioned dependencies
[09:00:08] <abaumann> good question. and another problem is: if icu or boost are required by anything on the build chroot, then we have a version conflict.
[09:00:09] <deep42thought> so pacman sees, that it has to keep icu62 around
[09:00:34] <deep42thought> I thought, the point was to have icu installable in different versions?
[09:00:52] <abaumann> yeah. but build scripts usually just head for /usr/include/boost and be done with it.
[09:00:54] <deep42thought> how does linking work if multiple versions of the same library are available?
[09:01:01] <abaumann> explicit only.
[09:01:07] <abaumann> which points back to the build system.
[09:01:23] <abaumann> so.. we either can build old versions on chroots not having ICU or boost on them..
[09:01:28] <abaumann> ..or it gets messy :-)
[09:01:36] <deep42thought> so the include headers do not force any library version?
[09:01:46] <abaumann> not that I'm aware of.
[09:01:55] <abaumann> it's not glibc/dbus :-)
[09:02:31] <deep42thought> we could make a [graveyard] repository, where we move icu when a new version comes out
[09:02:40] <deep42thought> and we rename it to icu62
[09:02:42] <abaumann> in an empty chroot, we can use the standard PKBUILD and just decrease the version number, change the package name and install only the shared libraries .52 or so.
[09:02:49] <deep42thought> and possibly strip it from file conflicting files?
[09:03:13] <deep42thought> I'm against rebuilding
[09:03:15] <abaumann> that's where I'm not sure: we should not be forced to change every PKGBUILD using icu.
[09:03:24] <deep42thought> it opens the possibility, that something goes wrong
[09:03:30] <abaumann> that's true.
[09:03:47] <abaumann> the other alternative is to speed up the rebuild process of firefox, etc.
[09:03:57] <deep42thought> :-D
[09:04:00] <abaumann> so what's blocking that (besides limited amount of hands)? :-)
[09:04:09] <deep42thought> gpgme was blocking libreoffice
[09:04:12] <deep42thought> it failed in check()
[09:04:15] <deep42thought> I just skipped that now
[09:04:23] <abaumann> Yes, those are just the callbacks for python/python2.
[09:04:24] <deep42thought> for firefox/thunderbird: I guess, its OOM
[09:06:23] <abaumann> when we go with icu62: what about making just a package with no dependencies? it just provides /usr/lib/lib*.so.62 and /usr/lib/lib*.so.62.1?
[09:06:44] <abaumann> then packages draw in the new icu (63) package, fail to start and we have to manually install a standalone icu62 package.
[09:06:44] <deep42thought> that part is ok
[09:07:05] <deep42thought> yes, manuall install would work
[09:07:08] <abaumann> which may lie around on a machine forever.. but that's not a big disaster.
[09:07:18] <deep42thought> but it would possibly leave users with a lot of old icu packages at some point
[09:07:34] <deep42thought> it's probably the only simle solution
[09:07:56] <abaumann> yeah. like on debian :-)
[09:08:08] <deep42thought> except if we programatically add libicu.so.62 dependencies
[09:08:08] <abaumann> but I simply cannot see we can keep up with rebuilds..
[09:08:26] <abaumann> but that's again effort.
[09:09:01] <deep42thought> e.g. in package() of _each_ package: scan if it is linked against icu (we do anyway, but at another point), and add a depends=... entry - then makepkg will add the version of the linked libicu...
[09:09:07] <abaumann> There is also the bootstrap method: just build a stub packges with tar and hand-crafter .PKGINFO.
[09:09:37] <abaumann> ui.. my keyboard gets spongy. :-)
[09:09:40] <deep42thought> ^this is how I would create the old-icu package
[09:09:46] <abaumann> ah.. ok.
[09:10:08] <abaumann> not to have to bother to wait for icu/boost to rebuild in special versions (which always takes more than one attempt to get it right)
[09:10:23] <deep42thought> as I said: I'm against rebuild
[09:10:30] <deep42thought> OTOH repackageing is fine
[09:11:11] <abaumann> ah. I really like seamonkey. It runs without boost/icu, etc. :-)
[09:12:39] <deep42thought> hmm, there will be file conflicts between old icu packages
[09:12:47] <abaumann> why?
[09:12:58] <deep42thought> e.g.: which package should own /usr/lib/libicudata.so.63 ?
[09:13:13] <deep42thought> the icu63.1 or the icu63.2 package?
[09:13:19] <abaumann> ah. yeah. there is no package-libs concept in archlinux
[09:13:26] <deep42thought> or do we want to do this only for major versions?
[09:13:45] <abaumann> mmh. there is no such thing as major versions.. that's part of the problem with icu and boost.
[09:14:00] <abaumann> their versioning schema is - let's be polite - non-standard.
[09:14:22] <abaumann> defaecto every version is a new major release version or should be treated as such.
[09:15:18] <abaumann> so. that's bad. an icu63.so would now be owned by icu, icu62.so by icu62, when icu bumps to 64, then we have a installation conflict.
[09:15:54] <deep42thought> this example would work, actually
[09:16:21] <abaumann> icu would upgrade, then you have to manually install icu63 and uninstall icu62 optionally
[09:16:36] <abaumann> what I don't like: we deviate from upstream.
[09:16:41] <deep42thought> when something depends on icu=63, it would automatically install
[09:16:45] <deep42thought> true
[09:17:00] <deep42thought> maybe we should care more about the root cause: delayed / failing builds
[09:17:10] <abaumann> exactly.
[09:17:25] <abaumann> icu62 sounds like a workaround.
[09:20:28] <deep42thought> maybe, having a script around to create such a legacy package from an archive package is a good idea - then we can advise people to install that from some "external" source?
[09:20:45] <deep42thought> this is minimally invasive and should solve the issue
[09:21:30] <abaumann> good idea.
[09:22:14] <deep42thought> damn, i686/libreoffice-still is not handed out, because it run-depends on gpgme and that is not yet built for i486
[09:22:17] <deep42thought> ^bug
[09:23:46] <deep42thought> ah, no
[09:23:52] * deep42thought talks too fast
[09:25:13] -!- thePiGrepper has joined #archlinux32
[09:33:25] <buildmaster> i686/make is broken (says nlopc43).
[09:33:32] <deep42thought> ^ needs patch cleanup
[09:33:34] <abaumann> make.. cool.
[09:33:59] <abaumann> ah. ok. on it..
[09:34:05] <deep42thought> thx
[09:44:36] -!- thePiGrepper has quit [Ping timeout: 260 seconds]
[10:27:02] -!- woshty has joined #archlinux32
[10:32:21] <abaumann> error: could not open file /var/lib/pacman/sync/bootstrap-staging.db: Unrecognized archive format
[10:32:28] <abaumann> mmh. on which slave is that I wonder..
[10:32:44] <deep42thought> which log?
[10:32:53] <abaumann> zenity.367d04ad6c1de9bc340e77b75832631c6ca79d74.1661411e226a7aa5f98427aa20b0347ac990ca0f.extra.i486.2018-10-21T09:51:16.build-log
[10:34:55] <deep42thought> nlopc46-i486bs1
[10:35:03] <abaumann> ah, ok.
[10:36:48] <deep42thought> my build slaves are now running w/o your bootstrap repo, anyways
[10:36:56] <abaumann> ah. good.
[10:37:00] <deep42thought> (because it caused errors yesterday)
[10:37:06] <deep42thought> some packages won't build that way
[10:37:08] <abaumann> I saw that they are mostly empty anyway.
[10:37:52] <abaumann> Maybe we should keep the one bootstrap (the main directory) and remove the rest?
[10:38:01] <deep42thought> yeah
[10:38:03] <abaumann> Aka. move bootstrap-staging to bootstrap.
[10:38:06] <abaumann> I'll do that.
[10:38:08] <deep42thought> throw all packages into that one db for now
[10:38:12] <deep42thought> :-)
[10:38:16] <abaumann> :-)
[10:45:48] -!- thePiGrepper has joined #archlinux32
[11:00:02] -!- thePiGrepper has quit [Ping timeout: 268 seconds]
[11:10:34] -!- DepositePirate has joined #archlinux32
[11:21:57] -!- DepositePirate has quit [Remote host closed the connection]
[11:22:31] -!- DepositePirate has joined #archlinux32
[12:10:44] <buildmaster> i686/haskell-cairo is broken (says rechenknecht).
[12:32:16] <buildmaster> i686/haskell-glib is broken (says nlopc46).
[12:51:50] <buildmaster> i686/haskell-ghc-typelits-natnormalise is broken (says rechenknecht).
[12:55:31] <buildmaster> i686/haskell-attoparsec is broken (says nlopc43).
[13:01:38] <buildmaster> i686/haskell-hourglass are broken (says nlopc43).
[13:06:32] <buildmaster> i686/haskell-spdx is broken (says nlopc43).
[13:10:19] <buildmaster> i686/haskell-data-dword is broken (says eurobuild3).
[13:14:33] -!- woshty has quit [Read error: Connection reset by peer]
[13:21:19] <deep42thought> abaumann: please don't remove dependencies $CARCH conditional like so: https://git.archlinux32.org
[13:21:20] <phrik> Title:archlinux32/packages: Package customizations and pure-i686 packages - Archlinux32 Gitea (at git.archlinux32.org)
[13:21:33] <deep42thought> there is makedepends_i686 and alike for that
[13:26:25] <abaumann> ah, right. now there is. :-)
[13:26:36] <abaumann> this might be a copy-paste from the bootstrapping phase.
[13:26:53] <deep42thought> yeah, I'm cleaning up the other ones, too
[13:26:59] <abaumann> thanks
[13:27:09] <deep42thought> but first I want to verify, that it indeed breaks the buildmaster's logic ;-)
[14:13:45] -!- guys has quit [Ping timeout: 252 seconds]
[14:21:48] -!- abaumann has quit [Quit: leaving]
[14:28:36] <deep42thought> abaumann: where do you get all those to-be-rescheduled packages from?
[14:28:51] <deep42thought> The list looks quite large - is there something wrong with the usual logic?
[14:47:08] -!- abaumann has joined #archlinux32
[14:47:09] <buildmaster> Hi abaumann!
[14:47:09] <buildmaster> !rq abaumann
[14:47:09] <phrik> buildmaster: <abaumann> ok. it fails, but it works..
[14:47:25] <abaumann> deep42thought: aeh.. the list got longer than intended.. but it's finished now.
[14:47:42] <deep42thought> ah, good to know :-)
[14:47:45] <abaumann> those was a try to reschedule packages which broke because of broken pacman databases.
[14:47:46] <deep42thought> where did it come from?
[14:47:58] <deep42thought> ah, no need to reschedule those
[14:48:00] <abaumann> I have a arch32logs script analyzing the logfiles
[14:48:10] <deep42thought> they will stay on the build list
[14:48:20] <abaumann> I get things like:
[14:48:21] <abaumann> Total execution format errors during build (binary blobs of wrong architecture?): 1 (vulkan-tools)
[14:48:24] <abaumann> Illegal instructions during build: 9 (intel-gmmlib io libkkc-data libkkc libkolabxml nodejs-lts-dubnium nodejs xalan-c zam-plugins)
[14:48:30] <deep42thought> ah, ok
[14:48:39] <deep42thought> yeah, those should be rescheduled :-D
[14:48:41] <abaumann> most are broken dependencies.. those make no sense to reschedule.
[14:48:57] <abaumann> they where because of the i686 uname setarch thingy, right?
[14:49:03] <deep42thought> you can use the -a switch to only reschedule i486 packages if you need to
[14:49:09] <deep42thought> I think so
[14:49:26] <deep42thought> (but -a shouldn't really accellerate the queries)
[14:49:48] <abaumann> ah. yes. that would be better.
[14:50:03] <abaumann> I have to add architecture to my script..
[14:50:34] <abaumann> ah. only 3 i686 ones, the rest is i486.
[14:50:54] <deep42thought> what's the issue with the i686 ones?
[14:51:14] <abaumann> arbitrary failures: libgit2 has a offline test failure
[14:51:41] <abaumann> libnotify now needs gtk3, oj!
[14:52:10] <abaumann> ah. all i486 ones.
[14:52:13] <abaumann> wrong grep :-)
[14:52:36] <abaumann> grep 686 matches to md5s :-)
[14:52:49] <deep42thought> lol
[15:08:13] -!- DCyrax has joined #archlinux32
[15:30:09] -!- guys has joined #archlinux32
[15:37:28] <deep42thought> finally: all typos fixed - now only the packages need to be built :-D
[15:45:09] <buildmaster> i686/haskell-wl-pprint-annotated is broken (says eurobuild3).
[16:19:45] -!- autofsckk has joined #archlinux32
[16:20:10] -!- deep42thought has quit [Quit: Leaving.]
[16:21:18] <autofsckk> hello, this icu problem is killing me :/ i cant open any browser or thunderbird with version 63, and if i downgrade to 62 i cant use any movie player :/ jajajaja well chromium still dont work with either version :/
[16:46:21] <autofsckk> if i try to run chromium i get this error /usr/lib/chromium/chromium: error while loading shared libraries: libicui18n.so.62: cannot open shared object file: No such file or directory :( is there a dirty fix for it?
[16:53:57] <autofsckk> i also get this error with libreoffice /usr/lib/libreoffice/program/soffice.bin: error while loading shared libraries: libicuuc.so.62: cannot open shared object file: No such file or directory
[16:56:36] -!- thePiGrepper has joined #archlinux32
[17:29:07] <abaumann> autofsckk: can you add my repo as last repo in /etc/pacman.conf: [aba-aur]
[17:29:07] <abaumann> Server = http://archlinux32.andreasbaumann.cc
[17:29:16] <abaumann> then do a pacman -Syyu to get the new icu 63
[17:29:23] <abaumann> then do a pacman -S icu62
[17:29:34] <abaumann> this installs a stub library with just the icu62 libraries.
[17:30:35] <abaumann> deep42thought: it's maybe time to add an official addon repo like the alarm guys..
[17:34:45] <autofsckk> firefox is working now
[17:34:53] <abaumann> :-)
[17:35:03] <abaumann> I'll do the same for boost167-libs
[17:35:30] <autofsckk> thunderbird working too, thanks
[17:36:03] <abaumann> it's temporary, so icu62 can be removed if all packages depending on icu63 have been rebuilt.
[17:36:09] <buildmaster> i686/deepin-session-ui is broken (says buildknecht).
[17:36:36] <autofsckk> chromium also working
[17:36:48] <abaumann> thanks for testing :-)
[17:37:03] <autofsckk> thanks a lot for fixing it
[17:37:25] <abaumann> np
[17:37:57] <buildmaster> i686/startdde is broken (says nlopc43).
[17:37:59] <abaumann> vlc works, but it worked before.
[17:38:04] <autofsckk> i tried yesterday with downgrading icu but today when i tried to start X i couldnt jajajaja so i had tu update all again, but still i couldnt use smplayer nor vlc yesterday so now it seems that everything is working
[17:38:34] <autofsckk> the other thing was pypanel i thinkg i had to install and configure xfce, let me see if it works now
[17:38:55] <buildmaster> i686/virtualbox-modules-arch is broken (says buildknecht).
[17:39:11] <abaumann> downgrading works sometimes, sometimes not. so does symlinking shared library sometimes work sometimes not. :-)
[17:39:31] <autofsckk> i also tried symlinking but didnt work
[17:39:35] <autofsckk> what is icu?
[17:39:40] <autofsckk> well give me a second
[17:39:41] -!- woshty has joined #archlinux32
[17:39:45] <autofsckk> ill reboot and back in a moment
[17:39:46] -!- autofsckk has quit [Quit: leaving]
[17:40:26] <abaumann> "International Components for Unicode", the library handling Unicode on Linux systems, and in Java.
[17:41:30] <abaumann> mmmh. boost has an unversioned shared library: /usr/lib/mpi.so
[17:41:46] <abaumann> so this would produce a name clash, I have to remove the library from the boost167-libs stub.
[17:44:03] -!- autofsckk has joined #archlinux32
[17:44:19] <autofsckk> well lxpanel still death jajajaja :/ so sad i dont like xfce panel
[17:45:13] <abaumann> mmh. Works for me, when I start it manually.
[17:46:12] <autofsckk> not for me
[17:46:26] <autofsckk> let me paste the output somewhere
[17:46:48] <autofsckk> but it does not say too much
[17:46:58] <autofsckk> https://bpaste.net
[17:46:59] <phrik> Title:show at bpaste (at bpaste.net)
[17:47:20] <abaumann> floating point error.. mmh.
[17:47:27] <autofsckk> the last part says something like floating point exception maybe?
[17:47:27] <abaumann> Excepci├│n de coma flotante (`core' generado)
[17:47:49] <autofsckk> i forgot how to get the default(english) output
[17:47:51] <abaumann> I don't speak Spanish, but I would bet, it does say "floating point exception" :-)
[17:47:57] <abaumann> no problemo
[17:47:59] <autofsckk> yes something like that
[17:48:08] <autofsckk> jajaja cool
[17:48:15] <abaumann> what's you CPU? I doubt you have no floating point unit.
[17:48:26] <abaumann> *your
[17:48:49] <autofsckk> a very rare one, let me see because i forgot how to see wich one is it
[17:48:57] <abaumann> cat /proc/cpuinfo
[17:49:29] <autofsckk> model name: Intel(R) Core(TM) Duo CPU T2450 @ 2.00GHz
[17:50:12] <autofsckk> let me c&p everything
[17:50:35] <autofsckk> https://bpaste.net
[17:50:36] <phrik> Title:show at bpaste (at bpaste.net)
[17:50:38] <abaumann> so, libreoffice-fresh also works, I added a boost167-libs package.
[17:51:32] <abaumann> mmmh. fpu, mmx, sse, sse2, allthere
[17:52:04] <buildmaster> i686/deepin-daemon is broken (says eurobuild3).
[17:52:06] <autofsckk> well not my libreoffice, i dont remember what version i have
[17:52:13] <autofsckk> i had to use openoffice
[17:52:19] <abaumann> pacman -Q | grep libreoffice
[17:52:27] <autofsckk> still
[17:52:40] <autofsckk> you say the fresh version works?
[17:52:45] <abaumann> yeas.
[17:52:56] <abaumann> well. it crashes on shutdown..
[17:52:57] <abaumann> but..
[17:53:04] <abaumann> ..yes it starts.
[17:53:23] <autofsckk> ok changing it to fresh then
[17:53:28] <abaumann> let me test libreoffice-still..
[17:53:33] <autofsckk> do you have any idea on how to fix the lxpanel ?
[17:54:26] <abaumann> Not really. Without a core dump it gets really hard to debug.
[17:54:43] <autofsckk> because i really like it, i dont like xfce and i dont want to spend time on finding something similar to lxpanel, i tried mmmm dont remember but it does not have menu button so i can find apps and that stuff
[17:54:53] <abaumann> maybe when I find what the specialties of this T2450 are, reading some Intel docu right now :-)
[17:55:03] <abaumann> ..libreoffice-still also works
[17:55:57] <autofsckk> so what does icu do?
[17:56:00] <autofsckk> libraries for what?
[17:56:13] <abaumann> every time you handle a Unicde character.
[17:56:31] <abaumann> for instance do a tolowercase on Chinese characters.
[17:56:36] <autofsckk> still not working
[17:56:37] <autofsckk> Warning: failed to read path from javaldx
[17:56:38] <autofsckk> /usr/lib/libreoffice/program/soffice.bin: error while loading shared libraries: libboost_locale.so.1.67.0: cannot open shared object file: No such file or directory
[17:56:44] <abaumann> ah.
[17:56:49] <abaumann> install boost167-libs from my repo :-)
[17:56:53] <buildmaster> i686/virtualbox is broken (says nlopc46).
[17:57:23] <abaumann> *abaumann feels elibrokeit flying over him and warning him about user binary blobs and personal repos..
[17:57:45] <autofsckk> abaumann: well it seems to not be there, pacman says it doesnt exist
[17:57:52] <elibrokeit> uhhh
[17:57:56] <abaumann> pacman -Syy
[17:57:58] <elibrokeit> why not just fix boost
[17:58:13] <elibrokeit> and dependent packages
[17:58:30] <abaumann> elibrokeit: we have a huge backlog. icu is not broken, but our build process is stuck in the middle.
[17:58:40] <abaumann> half of the packages need old icu, the other half new icu
[17:59:04] <elibrokeit> and you accidentally moved the wrong things?
[17:59:44] <abaumann> it's a trade off, actually. the dependency system doesn't allow any moves.. then people complain about a new feature.. then we move.. then things break :-)
[18:00:42] <abaumann> http://archlinux32.andreasbaumann.cc
[18:00:43] <phrik> Title:#archlinux32 | Logs for 2018-11-09 (at archlinux32.andreasbaumann.cc)
[18:00:56] <abaumann> we debated some options. :-)
[18:01:23] <abaumann> autofsckk: any luck with libreoffice?
[18:03:33] <buildmaster> i686/libgsignon-glib is broken (says buildknecht).
[18:04:20] <autofsckk> abaumann: give me a secondd please
[18:04:34] <abaumann> I even give you some minutes.. :-)
[18:09:07] <autofsckk> im back, havent read the logs
[18:09:21] <autofsckk> let me see and ill tell you if i can make it work
[18:16:37] <buildmaster> i686/frei0r-plugins are broken (says nlopc46).
[18:18:52] <buildmaster> i686/ptlib is broken (says buildknecht).
[18:19:40] <buildmaster> i686/ganv is broken (says nlopc43).
[18:22:05] <autofsckk> abaumann: so i got a little lost, very technical for me really, but i think it was good for me so i can try to understand what is the icu problem
[18:22:27] <autofsckk> but i dont see how can i make libreoffice run with that info :(
[18:25:29] <abaumann> did you install boost167-libs?
[18:28:37] <buildmaster> i686/movit is broken (says buildknecht).
[18:32:12] <abaumann> autofsckk: it's very simple, libreoffice-still and libreoffice-fresh both need boost167-libs and icu62.
[18:34:12] <autofsckk> no i havent
[18:34:17] <autofsckk> but i didnt find the way to install it
[18:34:35] <abaumann> pacman -Syy boost167-libs
[18:34:46] <abaumann> it's also in by aba-aur repo..
[18:38:40] <autofsckk> cool is working now
[18:38:45] <abaumann> :-)
[18:39:07] -!- deep42thought has joined #archlinux32
[18:39:07] <buildmaster> Hi deep42thought!
[18:39:07] <buildmaster> !rq deep42thought
[18:39:09] <phrik> buildmaster: <deep42thought> good night, python - cu tomorrow python2!
[18:39:27] <abaumann> hi deep42thought
[18:39:42] <deep42thought> abaumann: I don't think an [archlinux32] repo makes sense - we should rather try to avoid the current situation ;-)
[18:39:44] <deep42thought> Hi abaumann!
[18:40:25] <abaumann> yeah. but we also habe packages, which got dropped upstream, like pcmciautils, consequently they should go into an archlinux32 repo.
[18:40:45] <abaumann> consequenctly == iff following my alarm logic. :-)
[18:40:58] <deep42thought> yeah, good point
[18:41:13] <deep42thought> but honestly, I don't consider one repository for 3 packages very useful ;-)
[18:41:23] <deep42thought> but I'll think about it
[18:41:37] <deep42thought> because those packages are also built in a different way than the "upstream" packages
[18:41:43] <abaumann> for now my machine is running again.. and the one of autofsckk.
[18:41:57] <deep42thought> calibre is rebuilt in staging already
[18:42:04] <abaumann> ah. ok.
[18:42:14] <deep42thought> I'll move it forward - it won't be worse than the one, we already have ;-)
[18:42:20] <abaumann> true. :-)
[18:44:43] <deep42thought> fatal error: libxul.so: mmap: failed to allocate 1793597048 bytes for output file: Cannot allocate memory
[18:44:47] <deep42thought> thunderbird build
[18:44:52] <abaumann> yeah. this one we know.
[18:44:59] <deep42thought> how did we solve this?
[18:45:14] <abaumann> building on a machine with tweaks..
[18:45:29] <abaumann> ..no systemd-spawn, enough memory.. IIRC
[18:46:52] <deep42thought> ah, that was the reason, why the dependencies were not 100% recorded in the db for firefox and thunderbird
[18:47:05] <abaumann> yeah.
[18:47:29] <deep42thought> we should probably hack build-packages to allow building with "makepkg" directly?
[18:47:40] <abaumann> ah. that's a good idea.
[18:49:53] <buildmaster> i686/geary is broken (says nlopc43).
[18:55:38] <deep42thought> bin/build-packages -s :without_systemd_nspawn:
[18:56:06] <deep42thought> abaumann: can you try that^ on your super-box which is capable of building thunderbird?
[18:56:12] <abaumann> super-box :-)
[18:56:21] <abaumann> yes.
[18:56:27] <deep42thought> -p thunderbird
[18:56:32] <deep42thought> of course ;-)
[18:56:49] <deep42thought> well, your box is better than everything I have
[18:56:56] <deep42thought> it can build thunderbird :-D
[18:57:07] <abaumann> let's hope.
[18:59:43] <abaumann> "x86_64" != "i686"
[18:59:54] <abaumann> mmh. so, I have to downgrade my superbox. :-)
[19:00:10] <deep42thought> well, you obviously need to run that on the target hardware
[19:00:18] <deep42thought> or you do "setarch i686" before
[19:00:34] * deep42thought is glad, he implemented that safety check
[19:01:26] <abaumann> :-)
[19:02:47] <abaumann> building package "squeak-vm" straw :without_systemd_nspawn:) ... failed.
[19:02:49] <abaumann> huh?
[19:02:54] <deep42thought> argh
[19:03:05] <deep42thought> "-p" only gives a preferrence
[19:03:10] <deep42thought> but it does not force anything
[19:03:11] <deep42thought> :-/
[19:03:21] <abaumann> I set the force.eurodbuild3 file on the buildmaster
[19:03:29] <deep42thought> O.o
[19:03:32] <abaumann> forced-package-builds.eurobuild3
[19:03:39] <deep42thought> that should do the trick
[19:18:29] <deep42thought> libreoffice-still is rebuilt :-) ... moving
[19:19:01] <buildmaster> i686/allegro is broken (says nlopc43).
[19:23:33] <abaumann> :-)
[19:23:46] <abaumann> 2018-11-09 19:23:14: building package "thunderbird" (revisions 06fca138ab84545120291bc9ff4357ff8bf1fced c980d934f2ffb085002debe9e4c863b5e5f48934, repository extra, straw :without_systemd_nspawn:) ... failed.
[19:23:54] <deep42thought> :-/
[19:23:57] <abaumann> ..and the logs have gone..
[19:24:05] <deep42thought> have probably been uploaded
[19:24:12] <abaumann> ah.
[19:24:37] <deep42thought> ==> ERROR: Could not resolve all dependencies.
[19:24:54] <deep42thought> how did you build that?
[19:25:07] <abaumann> ./build-packages -s :without_systemd_nspawn: -p thunderbird
[19:25:27] <abaumann> question: does it do makepkg on the host?
[19:25:29] <deep42thought> on which host?
[19:25:34] <deep42thought> yes
[19:25:36] <abaumann> eurobuild3
[19:25:40] <abaumann> oh.
[19:25:44] <deep42thought> this is a x86_64 host, right?
[19:25:45] <abaumann> then it took 64-bit binaries.
[19:25:46] <abaumann> yes.
[19:25:48] <deep42thought> :-D
[19:26:12] <abaumann> mmh. so, I have to set up another build slave purely i686.
[19:26:19] <buildmaster> i686/libgnome-keyring is broken (says buildknecht2).
[19:26:24] <deep42thought> I'll try it in my arch32-test vm
[19:28:01] <abaumann> java7-openjdk on my slave, *sigh*
[19:28:09] <abaumann> I'll leave it there. :-)
[19:28:29] <deep42thought> damn
[19:28:38] <deep42thought> I forgot some flags to makepkg *blush*
[19:28:46] <deep42thought> e.g. "-s"
[19:29:05] <abaumann> oh. this helps a lot with dependencies :-)
[19:29:28] <deep42thought> also --noconfirm is nice ;-)
[19:29:39] <abaumann> *grin*
[19:31:20] <buildmaster> i686/libgxps are broken (says nlopc46).
[19:47:51] <buildmaster> i686/java7-openjdk is broken (says eurobuild3).
[19:49:23] <deep42thought> abaumann: looks, like your build slave is done with java ;-)
[19:54:21] <abaumann> that was fast.
[19:54:32] <abaumann> and doing firefox.
[19:54:33] <abaumann> ok.
[20:48:52] <deep42thought> abaumann: I probably found the problem, why the force file was ignored
[20:49:10] <deep42thought> when the buildmaster thinks, you're already building another package, it will give you that assignment once more
[20:49:13] <deep42thought> ^bug
[20:50:31] <deep42thought> well
[20:51:11] <deep42thought> it's the question: if forced-package-builds.$slave is set and the slave builds some package, but requests a package (again): should we hand out the old one (as we do normally) or a new one?
[20:53:10] <deep42thought> if this happens again, you can run 'return-assignment ABORT' on the buildmaster via ssh from your slave and the buildmaster will remove the current build assignment from your slave
[20:53:20] <deep42thought> maybe, I should add a shortcut for that :-)
[20:55:04] <abaumann> well.. as you explain it.. it all makes sense. :-)
[20:57:50] <deep42thought> that's why teaching is part of my job description
[21:03:40] -!- abaumann has quit [Remote host closed the connection]
[21:35:21] <deep42thought> lol, I just googled the current failure of qt5-webengine and landed in our forum :-D
[21:39:23] <buildmaster> i686/dwarffortress are broken (says nlopc43).
[21:39:43] -!- deep42thought has quit [Quit: Leaving.]
[22:20:07] <buildmaster> i686/qt5-webengine is broken (says buildknecht2).
[23:36:46] -!- isacdaavid has joined #archlinux32
[23:53:08] <buildmaster> i686/extremetuxracer is broken (says nlopc43).