#archlinux32 | Logs for 2024-05-04

Back
[03:28:45] -!- bill-auger_ has joined #archlinux32
[03:30:09] -!- bill-auger has quit [*.net *.split]
[03:47:51] -!- socksinspace has quit [Ping timeout: 256 seconds]
[03:49:25] -!- socksinspace has joined #archlinux32
[06:27:48] -!- titus_livius has joined #archlinux32
[06:29:53] -!- abaumann has joined #archlinux32
[06:29:53] <buildmaster> Hi abaumann!
[06:29:53] <buildmaster> !rq abaumann
[06:29:54] <phrik> buildmaster: <abaumann> reality is more cruel than cruel jokes about reality.. ;-)
[06:30:02] <abaumann> nice to get puthon 3.12 :-)
[06:30:21] <abaumann> currently most packages on i486 which are broken are broken because of a broken python (what a sentence).
[06:30:34] <abaumann> https://bbs.archlinux32.org
[06:30:35] <phrik> Title: Pacman installing Python packages in the previous version directory / Pacman / Pacman Upgrades / Arch Linux 32 Forum (at bbs.archlinux32.org)
[06:30:45] <abaumann> this indicates that some python packages are still on 3.10.
[06:31:03] <abaumann> I see no other option than to use pyenv/pipx or so to create your own envirnoments.
[06:31:13] <abaumann> Or rebuilding an older version of the python package for 3.11
[06:35:12] -!- buildmaster has quit [Remote host closed the connection]
[06:35:27] <abaumann> ui, my NAS doesn't work anymore:
[06:35:28] <abaumann> /usr/bin/smbd: /usr/lib/libldb.so.2: version `LDB_2.7.2' not found (required by /usr/lib/samba/libldbsamba-samba4.so)
[06:35:36] <abaumann> I have to rebuild ldb and possibly samba
[06:36:08] <abaumann> pacman -Q| grep ldb
[06:36:08] <abaumann> ldb 2:2.6.1-1.0
[06:36:22] <abaumann> yep, and samba is built against 2.7.2
[06:36:27] <abaumann> ldb 2:2.9.0-3
[06:36:34] <abaumann> that's the version from upstream.
[06:37:00] <abaumann> and again, I'm enabling all possible build slaves to speed up the python 3.12 transition process..
[06:39:24] -!- buildmaster has joined #archlinux32
[06:40:36] <abaumann> https://buildmaster-status.archlinux32.org
[06:40:42] <abaumann> ah, most python modules fail to build
[06:44:36] -!- zxrom has joined #archlinux32
[06:50:05] <abaumann> and again: somebody changed the files in /usr/share/webapps/cgit on the buildmaster directly. If this continues, I'll just disable login accounts..
[06:50:33] <abaumann> also, it could be an automatic update mechanism of sorts updating from an old git repo
[06:51:56] <abaumann> whoever is changing that should use the git repo or notify us at least..
[06:55:19] <abaumann> I'll have a look at the changes and what they are doing, but the CSS is clearly failty, the changes to robots.txt look interesting.
[07:02:48] <abaumann> Checking for system pyldb-util.cpython-312-i386-linux-gnu (>=2.9.0 <=2.9.999) : not found
[07:02:57] <abaumann> and ldb cannot be built without python.
[07:06:11] <abaumann> I hate to say it: but the Linux userland is a mess which can only be built with tons of strange build tools, most of them written in weird languages.
[07:06:30] <abaumann> I'm currently enyoing minix and BSDs, they just have a make for the whole basic userland
[07:06:43] <abaumann> Linux goes a strange way from a Unix perspective.
[07:08:31] <abaumann> http://archive.archlinux32.org samba 4.20.0-2.0 is in the archive, but the officially publishied version is 4.18?!
[07:09:56] <abaumann> Checking for system tdb (>=1.4.10) : yes
[07:09:59] <abaumann> ERROR: Python module pytdb of version 1.4.10 not found, and bundling disabled
[07:10:17] <abaumann> sigh
[07:13:07] <abaumann> May 04 09:12:43 eurodata smbd[1101]: /usr/bin/smbd: error while loading shared libraries: libicui18n.so.74: cannot open shared object file: No such file or directory
[07:13:10] <abaumann> yeah.
[07:13:14] <abaumann> this distro falls apart..
[07:13:17] <abaumann> ..brutally.
[07:15:38] <abaumann> samba and smbclient 4.17.5 are the last working versions
[07:21:48] -!- abaumann has quit [Quit: leaving]
[08:20:27] <Foxboron> g48
[08:42:35] <KitsuWhooa> <abaumann> and again: somebody changed the files in /usr/share/webapps/cgit on the buildmaster directly. If this continues, I'll just disable login accounts..
[08:42:40] <KitsuWhooa> Wasn't me
[08:42:58] <KitsuWhooa> Are you sure there's no corruption, o
[08:43:01] <KitsuWhooa> *?
[09:05:07] <KitsuWhooa> abaumann: Yeah, I haven't had a chance to get 3.12 up and running yet
[09:05:42] <KitsuWhooa> I got python-bootstrap built at least, so now I need to get the basic python packages up
[09:05:58] <KitsuWhooa> last night all builders were building ruby so I had to waitr
[09:05:59] <KitsuWhooa> *wait
[09:07:53] <KitsuWhooa> if buildmaster doesn't schedule the packages I want it to, I'm going to just dump a whole bunch of basic python from upstream into build-support-manual
[09:12:07] <KitsuWhooa> ...honestly, what if we just had a repo mirror for build support that synced all `any` packages from upstream
[09:12:35] <KitsuWhooa> it is getting impossible to keep up with the few resources we have
[09:14:03] <KitsuWhooa> hahaha > Please, whoever changes /usr/share/webapps/cgit, please talk to us. :-)
[09:14:21] <KitsuWhooa> there has to be a way to find out what touches that...
[09:25:10] <KitsuWhooa> yeah okay, more python packages should build now
[09:30:01] <KitsuWhooa> maybe we should add a nocheck build support straw
[09:35:20] <KitsuWhooa> er, I misused the word straw but w/e
[09:55:07] <KitsuWhooa> abaumann: When you get the chance, can you git pull the builder repo on your builders?
[09:58:48] -!- abaumann has joined #archlinux32
[09:58:49] <buildmaster> Hi abaumann!
[09:58:49] <buildmaster> !rq abaumann
[09:58:50] <phrik> buildmaster: <abaumann> the world would be a better place if less would be programmed in C, or at least by people knowing C :-)
[09:59:19] <abaumann> Foxboron: I should have written "..this fork is falling apart brutally." :-)
[09:59:26] <abaumann> And I don't understand 'g48'..
[10:00:11] <abaumann> KitsuWhooa: corruption, always in the very same directory? hardly. I'm more suspecting some conflict between the git version and the cgit package..
[10:00:19] <KitsuWhooa> ah, maybe
[10:00:38] <abaumann> cgit 1.2.3.r54.g00ecfaa-11
[10:00:48] <abaumann> yes, and gitolite@git.archlinux32.org:cgit.git
[10:00:57] <abaumann> I must build a package and the install it :-)
[10:01:25] <abaumann> and see, what the differences are about, most likely vanilla cgit is adapted to the arch webpage style.
[10:01:34] <abaumann> ah, yes. I vaguely remember now :-)
[10:01:58] <abaumann> and cgit should really be a cgit32 or so and be a proper fork. I'll do that..
[10:02:40] <KitsuWhooa> in the meantime, I'm going to try to figure out how to make schedule-for-rebuild ignore missing pacakges instead of exiting
[10:02:50] <KitsuWhooa> I told it to reschedule the entirety of python and it failed at the third package
[10:02:56] <abaumann> oh.
[10:03:08] <KitsuWhooa> pykf5 85faaa480c1185d93da06e92b9eadd1bea53337c 0000000000000000000000000000000000000000 5f64ec224ade003ee382d49d1eb1a99706dca7d7 extra
[10:03:08] <KitsuWhooa> "make_source_info" did not return a "pkgbase" - eh, what?
[10:03:08] <KitsuWhooa> >curl: (22) The requested URL returned error: 500<
[10:03:09] <abaumann> but I saw quite some python packages building
[10:03:10] <KitsuWhooa> it no longer exists upstream
[10:03:26] <abaumann> the problem is in the shell code there.
[10:03:43] <KitsuWhooa> https://gitlab.archlinux.org
[10:03:44] <phrik> Title: Sign in · GitLab (at gitlab.archlinux.org)
[10:03:48] <abaumann> it simply does no error handling. I tried once to add it around make_pkgsrc_info. It's a little bit tricky.
[10:03:54] <KitsuWhooa> ah
[10:06:38] <abaumann> and there is the fallback-case for svngit of the old packages/community
[10:06:56] <KitsuWhooa> ah, that never worked for me
[10:07:05] <abaumann> pkginfo.php could returl 404, if a package doesn't exist.
[10:07:17] <abaumann> then the makepkg_src_info_from_cache could maybe also return some code
[10:07:24] <abaumann> and then the caller could check for the error
[10:07:48] <abaumann> or return something which at least fails fast on the build slave, like a errornous package
[10:08:03] <KitsuWhooa> mhm
[10:08:04] <abaumann> not producing a real package. but this might mess up the database a little bit. :-)
[10:08:22] <KitsuWhooa> we need to prune old packages at some point
[10:08:36] <KitsuWhooa> why do we still have haskell in the repos
[10:08:43] <KitsuWhooa> there's no chance in hell it still runs
[10:08:45] * abaumann shrugts
[10:08:49] <abaumann> *shrugs
[10:09:06] <abaumann> it's much worse, there are also haskell build straws and more..
[10:09:13] <KitsuWhooa> well...
[10:09:24] <KitsuWhooa> haskell_without_check is just "without_check"
[10:09:35] <KitsuWhooa> there's a sed in place :p
[10:09:42] <KitsuWhooa> while it'
[10:09:50] <KitsuWhooa> it's called haskell in the code, it's simply without_check
[10:09:58] <KitsuWhooa> at least that's how I understand it
[10:10:33] <KitsuWhooa> https://git.archlinux32.org
[10:10:34] <phrik> Title: Add without_check build_support step - builder - Archlinux32 build system (at git.archlinux32.org)
[10:10:36] <KitsuWhooa> which is why I added that
[10:10:41] <KitsuWhooa> to hopefully help with python bootstrapping
[10:11:02] * abaumann finds only references to gun models when he googles 'G48'
[10:11:12] * abaumann shrugs again
[10:11:23] <KitsuWhooa> maybe it's the code word to get us disowned :p
[10:11:28] <KitsuWhooa> /j
[10:11:39] <KitsuWhooa> Anyway, the ghc package is from 2021
[10:12:24] <abaumann> The problem is like this: if Archlinux32 falls apart, then it sheds a bad light on Archlinux. Though the quality issues are completely different ones and not primarily caused by upstream. Linux just gets more and more complicated and messy, that's all.
[10:12:32] <abaumann> And we don't have the resources to keep up.
[10:13:13] <KitsuWhooa> I agree, but also we are a fork
[10:13:42] <KitsuWhooa> maybe we should rebrand to NotArchlinux32 :p
[10:13:46] <abaumann> yep. But we use the trademarked name.
[10:13:46] <KitsuWhooa> yeah I don't know honestly
[10:14:01] <abaumann> And people still tend to confuse Archlinux, ArchlinuxARM, ArchlinuxPPC, Archlinux32, etc.
[10:14:10] <abaumann> and report to wrong forums
[10:14:39] <KitsuWhooa> yup
[10:15:08] <abaumann> I was hoping, that Debian dropping 32-bit Intel fast would cause some popularity boost for the remaining 32-bit dirstros around..
[10:15:16] <abaumann> ..so far, I don't see much.
[10:15:46] <abaumann> https://www.theregister.com
[10:15:46] <phrik> Title: Debian preps ground to drop 32-bit x86 as a separate edition • The Register (at www.theregister.com)
[10:15:48] <KitsuWhooa> technically they aren't dropping it
[10:15:58] <KitsuWhooa> as in, the term "drop" is a bit misleading
[10:15:59] <abaumann> No ISOS, no kernel. Just 32-bit userland.
[10:16:16] <abaumann> You cannot install an old 32-bit only machine with it.
[10:16:35] <abaumann> You can run some old 32-bit software, also known as steam/wine/proton.
[10:16:42] <KitsuWhooa> ah, hm
[10:17:01] <abaumann> The article neither mentions Archlinux32, nor Gentoo, nor T2SDE, of course.
[10:17:25] <KitsuWhooa> I feel like Gentoo isn't really viable on such slow machines
[10:17:29] <KitsuWhooa> but yeah I guess I'm wrong
[10:17:37] <abaumann> ah, they mention Gentoo.
[10:17:45] <KitsuWhooa> and they do, yes
[10:17:53] <abaumann> "openSUSE Tumbleweed, Alpine Linux, Mageia, Gentoo, and Void Linux"
[10:18:11] <abaumann> Gentoo goes more binary lately.
[10:18:29] <abaumann> and you always had the binary stages for bootstrapping.
[10:19:22] <KitsuWhooa> oh there's a flag "ignore_mysql_generate_package_metadata_errors"
[10:19:33] <KitsuWhooa> is that what -f does...
[10:19:35] <KitsuWhooa> hm
[10:20:19] <KitsuWhooa> yup, that's what -f does
[10:20:23] <KitsuWhooa> my job is done for today /s
[10:20:49] <KitsuWhooa> we still have pyqt4?!
[10:21:03] <KitsuWhooa> why does the db keep track of it
[10:21:34] <KitsuWhooa> anyway, I'm rescheduling all of python that I can. We'll see what happens
[10:21:41] * abaumann just burned a finger on hot broccholi..
[10:21:49] <abaumann> yep.
[10:21:51] <KitsuWhooa> ouch
[10:22:29] <abaumann> I'll keep the lots of slaves running, then it should at least fail fast..
[10:22:47] <KitsuWhooa> thanks
[10:22:53] <KitsuWhooa> did you git pull the builder repo on them
[10:22:53] <KitsuWhooa> ?
[10:23:01] <abaumann> yes.
[10:23:03] <KitsuWhooa> awesome
[10:23:11] <abaumann> and I reinstalled the newest devtools32 everywhere
[10:23:17] <abaumann> and updated some signing keys too :-)
[10:23:35] <KitsuWhooa> speaking of isos
[10:23:38] <KitsuWhooa> we need to get a new iso out at some point
[10:23:49] <KitsuWhooa> maybe once python 3.12 is built
[10:23:59] <abaumann> definitely.
[10:24:07] <abaumann> but building them currently fails.
[10:24:24] <KitsuWhooa> Well I have no idea how one does it, so
[10:24:45] <abaumann> Not complicated, there are some repos for it.
[10:24:51] <abaumann> I can do them easily.
[10:24:55] <abaumann> semi-automatic.
[10:25:03] -!- AtleoS has quit [Ping timeout: 255 seconds]
[10:25:05] <abaumann> but currently some essential packages are missing
[10:25:24] <abaumann> and archbuild is written in.. guess.. Python :-)
[10:25:42] -!- AtleoS has joined #archlinux32
[10:29:47] <KitsuWhooa> mhm
[10:32:00] -!- drathir_tor has quit [Read error: Connection reset by peer]
[10:35:25] * abaumann goes from eating brocoli to planting aloe veras now
[10:37:26] -!- drathir_tor has joined #archlinux32
[11:24:54] -!- zxrom has quit [Quit: Leaving]
[11:41:34] <KitsuWhooa> sounds fun
[11:41:39] <KitsuWhooa> meanwhile, just built the first jaraco package
[12:01:05] -!- AtleoS has quit [Ping timeout: 252 seconds]
[13:25:22] <KitsuWhooa> buildmaster is refusing to build other important packages and it's pissing me off
[13:25:46] <KitsuWhooa> also uh
[13:25:52] <KitsuWhooa> Why is eurobuild3-i486 building pentium4/llvm14
[13:26:54] <abaumann> <oeh. this is ineresting..
[13:36:03] <KitsuWhooa> Yeah I don't understand why it's not building the python stuff
[13:57:19] <abaumann> maybe because the main pentium4/python is building currrently?
[13:57:26] <KitsuWhooa> ah
[13:57:27] <KitsuWhooa> that'd do it
[13:57:56] <abaumann> the pentium4/llvm14 on a 486 vm is a bug.
[13:58:32] <abaumann> eurobuild6-8-i486 builds only outdated stuff? ok. something to check for me..
[14:04:44] <abaumann> ah. stuck..
[14:07:18] <abaumann> urgh eurobuild6-7-i486 takes 6 cpus or so, for a 486.. well. :-)
[14:23:28] <KitsuWhooa> 17:22:36 <trotz> 2024/05/04 16:21 WARN buildmaster Current Users USERS WARNING - 21 users currently logged in
[14:23:29] <KitsuWhooa> 21?
[14:23:34] <KitsuWhooa> ooooh
[14:23:35] <KitsuWhooa> it's the builders
[14:26:16] <abaumann> yea, originally the plan was not to run builders on the buildmaster..
[14:26:22] <abaumann> I increase the nagios value
[14:26:39] <KitsuWhooa> we're building any/python-iniconfig
[14:26:42] <KitsuWhooa> so yeah we're getting somewhere
[14:26:49] <KitsuWhooa> and any/python-jaraco.functools
[14:28:10] <abaumann> good
[15:19:58] <KitsuWhooa> So, `schedule-for-rebuild -j -w -l -p '^pattern$'` will build the last version of all the packages found by the pattern
[15:20:01] <KitsuWhooa> hopefully this makes life easier
[15:20:21] <abaumann> cool
[15:20:24] <abaumann> :-)
[15:30:07] <KitsuWhooa> let's see if meson builds
[15:30:21] <abaumann> uiui, feeling curagious.. ;-)
[15:32:15] <KitsuWhooa> https://archlinux32.org
[15:32:17] <phrik> Title: Arch Linux 32 - meson 1.4.0-5.0 (any) (at archlinux32.org)
[15:32:18] <KitsuWhooa> we do in fact have a meson
[15:32:32] <abaumann> \o/
[15:33:23] <KitsuWhooa> unsure why buildmaster won't build hatchling though
[15:33:35] <KitsuWhooa> all of its dependencies should be met
[15:35:06] <KitsuWhooa> oh, pluggy isn't built yet
[16:13:36] -!- abaumann has quit [Quit: leaving]
[16:35:24] -!- AtleoS has joined #archlinux32
[16:58:58] -!- zxrom has joined #archlinux32
[18:12:09] -!- zxrom_ has joined #archlinux32
[18:13:49] -!- zxrom has quit [Ping timeout: 268 seconds]
[18:36:38] zxrom_ is now known as zxrom
[20:58:00] -!- zxrom has quit [Quit: Leaving]