#archlinux32 | Logs for 2024-04-08

Back
[00:52:01] -!- GNUtoo has joined #archlinux32
[03:04:30] -!- ssserpent has quit [Ping timeout: 255 seconds]
[03:06:15] -!- ssserpent has joined #archlinux32
[04:00:31] -!- GNUtoo has quit [Ping timeout: 256 seconds]
[07:17:07] -!- ssserpent has quit [Quit: WeeChat 4.2.1]
[07:18:35] -!- ssserpent has joined #archlinux32
[10:53:42] <KitsuWhooa> Is it not possible to have a package in build-support with the same name as the normal repos?
[11:46:18] <KitsuWhooa> abaumann: I believe euronuc ran out of memory trying to build clang
[11:46:41] <KitsuWhooa> I'll temporarily disable it so that my builders can build it
[12:50:55] -!- AtleoS has quit [Ping timeout: 272 seconds]
[14:02:15] -!- abaumann has joined #archlinux32
[14:02:16] <buildmaster> Hi abaumann!
[14:02:16] <buildmaster> !rq abaumann
[14:02:16] <phrik> buildmaster: <abaumann> how did this ever build?
[14:02:38] <abaumann> KitsuWhooa: oups, yes. euronuc was on -j8 and not enough swap :-)
[14:05:45] <KitsuWhooa> no worries
[14:05:49] <KitsuWhooa> I got llvm built on my machine
[14:06:05] <KitsuWhooa> currently trying to bootstrap glib2
[14:06:16] <abaumann> oh. good.
[14:06:31] <KitsuWhooa> glib2 depends on gobject-introspect which depends on glib2 :p
[14:06:34] <abaumann> I wanted to add (seen deep42thoughts remark) that you do most of the work right now. :-)
[14:06:43] <abaumann> I'm just doing some emergency stuff (sometimes)..
[14:06:55] <KitsuWhooa> and answering my questions and fixing things I don't know how to fix :p
[14:06:59] <KitsuWhooa> that's pretty valuable I'd say
[14:07:17] <abaumann> gobject-introspect: mmh. that one is the one with binary code in it and a hard way to bootstrap IIRC
[14:07:25] <KitsuWhooa> 17:05:49 <KitsuWhooa> I got llvm built on my machine <-- *clang, sorry
[14:07:31] <abaumann> good.
[14:07:33] <KitsuWhooa> it should actually be relatively easy
[14:07:56] <abaumann> I reconfigured eurnuc now, should cope better with high load/memory situations - I hope
[14:07:57] <KitsuWhooa> you disable introspection in both, and then do another build pass with introspection
[14:08:00] <KitsuWhooa> thanks!
[14:08:03] <abaumann> ah.
[14:08:14] <abaumann> yes. You still have an old working gobject-introspection around.
[14:08:19] <KitsuWhooa> it's for python 3.10
[14:08:22] <abaumann> I had none when bootstrapping.
[14:08:30] <KitsuWhooa> I'd say it's the same as none :p
[14:08:36] <abaumann> :-)
[14:09:01] <KitsuWhooa> but yes, the idea is to make -nogir packages in build-support, and then make the proper packages depend on those
[14:09:03] <KitsuWhooa> that should fix it
[14:09:16] <abaumann> ah, yes. that should work.
[14:09:51] <KitsuWhooa> Speaking of bootstrapping
[14:09:56] <KitsuWhooa> any idea how one would bootstrap mingw-w64?
[14:10:03] <abaumann> uh.
[14:10:15] <abaumann> never done that.
[14:10:28] <KitsuWhooa> I guess it was deep42thought who did it :p
[14:10:29] <abaumann> but shouldn't the thing build itself from the host?
[14:10:38] <KitsuWhooa> didn't seem like it
[14:10:48] <KitsuWhooa> I didn't look too much into it, but I'd like to get it working eventually
[14:10:55] <abaumann> mmh. no. sorry, never felt the urge to build windows software that way ;-)
[14:11:22] <KitsuWhooa> fair enough
[14:15:04] <abaumann> mmh. Do we still have this systemd/cryptsetup/mkinitcpio messup. Oh dear!
[14:15:21] <KitsuWhooa> we do?
[14:15:29] <abaumann> In testing.
[14:15:32] <KitsuWhooa> oh
[14:15:37] <abaumann> maybe just a matter of pushing the right packages.
[14:15:44] <KitsuWhooa> I never touch testing, so I don't know
[14:16:20] <abaumann> I have my 6 test guinea pigs updating now, stable is fine on each subarchitecture, testing insists on deleting systemd on upgrade :-)
[14:16:28] <KitsuWhooa> oh dear
[14:16:30] <KitsuWhooa> yeah that's bad
[14:17:55] <KitsuWhooa> I'm having trouble getting pacman to replace glib2 with glib2-nogir automatically on build-support
[14:18:01] <KitsuWhooa> I tried conflicts, and it just aborted
[14:18:10] <KitsuWhooa> as in, conflicts=(glib2)
[14:18:18] <KitsuWhooa> and glib2 is installed by default so I need it to replace it
[14:19:06] <abaumann> mmh. something very low-level depends on glib2?
[14:19:19] <KitsuWhooa> I have no idea
[14:19:29] <KitsuWhooa> I do mkarchroot with base-devel and glib2 is there
[14:19:55] <KitsuWhooa> maybe I'm doing something wrong
[14:20:31] <abaumann> glib2 is not part of base-devel
[14:20:40] <abaumann> maybe a indirect dependency
[14:20:54] <abaumann> I would try to chroot into the generated chroot and remove glib2 and see what complains. :-)
[14:21:18] <KitsuWhooa> Ooooh
[14:21:24] <KitsuWhooa> I think I know why
[14:21:32] <KitsuWhooa> it is probably being dragged in as another dependency
[14:21:36] <KitsuWhooa> when trying to build the package
[14:21:42] <abaumann> I think so..
[14:21:51] <KitsuWhooa> oh no
[14:22:04] <abaumann> don' tell mit it's dbus or systemd ;-)
[14:22:09] <KitsuWhooa> :: removing glib2 breaks dependency 'glib2' required by libsecret
[14:22:15] <abaumann> *don't tell me
[14:22:15] <KitsuWhooa> :: removing glib2 breaks dependency 'glib2' required by pinentry
[14:22:21] <KitsuWhooa> why are those installed by default
[14:22:29] <KitsuWhooa> mkarchroot -C /pacman-i686-staging-build-support-manual.conf /chroottest/i686bs/root base-devel
[14:22:32] <KitsuWhooa> this is how I'm making the chroot
[14:23:08] <KitsuWhooa> then I immediately chroot into it, and it has glib2
[14:23:34] <abaumann> yeah, pacman needs gpgme, which needs gnupg, which needs pinentry which needs libsecret and voila -glib2
[14:23:36] <KitsuWhooa> I guess the real question is, can I get buildmaster to pick up on glib2 from build support without renaming it?
[14:24:02] <abaumann> it should take the first glib2 from the first repo it finds.
[14:24:18] <abaumann> so if you name it the same way and put it int build-support-manual (or build-support) that should work
[14:24:41] <KitsuWhooa> so you mean building it manually
[14:25:12] <KitsuWhooa> just to make sure we're on the same page
[14:25:14] <KitsuWhooa> I did this https://git.archlinux32.org
[14:25:15] <abaumann> if you cannot build it in build-support with the same name (I remember some issues we had?)
[14:25:15] <phrik> Title: packages - Archlinux32 package modifications (at git.archlinux32.org)
[14:25:22] <KitsuWhooa> and then get-package-updates ignored it compeltely
[14:25:25] <KitsuWhooa> it didn't even show up in the list
[14:25:46] <abaumann> mmh.
[14:25:54] <KitsuWhooa> and then I did this https://git.archlinux32.org
[14:25:54] <phrik> Title: packages - Archlinux32 package modifications (at git.archlinux32.org)
[14:25:56] <KitsuWhooa> and it picked it up
[14:26:04] <KitsuWhooa> well, that and https://git.archlinux32.org
[14:26:05] <phrik> Title: packages - Archlinux32 package modifications (at git.archlinux32.org)
[14:26:14] <abaumann> strange, strange
[14:26:23] <KitsuWhooa> I know :p
[14:27:22] <KitsuWhooa> only other thing I can think of is to have it install things in a different directory
[14:27:26] <KitsuWhooa> but that sounds really annoying
[14:27:52] <abaumann> uh, yes.
[14:28:03] <abaumann> I would probably go the forced manual way.
[14:28:13] <abaumann> Apr 08 16:27:46 euronuc.lan build-packages[3169596]: Build master is not sane.
[14:28:16] <abaumann> oupsi :-)
[14:28:20] <KitsuWhooa> I'll look into it
[14:28:24] <abaumann> thanks :-)
[14:28:31] <KitsuWhooa> > >> Ctrl-C -- query killed.
[14:28:31] <KitsuWhooa> ah
[14:28:32] <KitsuWhooa> that's me
[14:28:40] <KitsuWhooa> I keep pressing Ctrl C without thinking
[14:29:04] <abaumann> no problem. you can just remove the sql thingies
[14:29:07] <KitsuWhooa> yup
[14:29:10] <abaumann> and do a sanity test again
[14:29:11] <KitsuWhooa> I learned now
[14:29:15] <abaumann> happens to me all the time :-)
[14:29:22] <KitsuWhooa> good to hear I'm not the only one :p
[14:29:25] <abaumann> quite robust - the buildmaster
[14:29:31] <KitsuWhooa> honestly
[14:29:37] <KitsuWhooa> Once I started getting things going
[14:29:40] <KitsuWhooa> it started behaving
[14:29:52] <KitsuWhooa> still have to manually pick which git revision to rebuild though
[14:29:58] <KitsuWhooa> right, sane again
[14:30:08] <abaumann> yes, there is quite some things not ok.
[14:30:16] <abaumann> xz, is it now in the state git?
[14:30:22] <KitsuWhooa> I have no idea
[14:30:35] <KitsuWhooa> as for build-support, I think buildmaster treats it as yet another repo with higher priority
[14:30:37] <abaumann> I had to hack a local copy, because upstream was just changing the PKGBUiLD git repo, and not adapting the state sha
[14:30:42] <abaumann> this is quite annoying.
[14:30:46] <KitsuWhooa> I noticed when python-setuptools got build in normal repos, it deleted it from build-support
[14:30:51] <KitsuWhooa> which then broke python in build-support
[14:30:58] <abaumann> oh, yea.
[14:31:05] <KitsuWhooa> so I then had to manually delete the other bootstrap packages
[14:31:10] <KitsuWhooa> which is fine
[14:31:22] <KitsuWhooa> Oh this reminds me
[14:31:32] <abaumann> ah, good, so there is a way out. which is fine :-)
[14:31:47] <KitsuWhooa> every time get-package-updates runs, buildmaster rebuilds the same three packages
[14:31:55] <KitsuWhooa> https://archlinux32.org
[14:31:57] <phrik> Title: Arch Linux 32 - sysfsutils 2.1.1-1.123 (pentium4) (at archlinux32.org)
[14:31:58] <abaumann> mmh.
[14:32:00] <KitsuWhooa> this has been built 123 times so far
[14:32:02] <abaumann> that sound equally wrong.
[14:32:05] <abaumann> *sounds
[14:33:32] <KitsuWhooa> but anyway
[14:35:08] <abaumann> blasphemy: testing arch32 vms on Windows ;-)
[14:35:09] <KitsuWhooa> Yeah I'll inject the package manually, this is really annoying
[14:35:11] <KitsuWhooa> :o
[14:36:29] <abaumann> more blashemy: using Xming on Windows to remote VNC virtual machines from yet another machine..
[14:37:52] <KitsuWhooa> hah
[14:38:26] <abaumann> mmh. stable i486 doesn't allow me to log in on the console.. some pam/system breakage?
[14:38:41] <KitsuWhooa> unfortunately I have not been testing i486
[14:38:45] <KitsuWhooa> so I have no idea what is going on with it
[14:38:58] <abaumann> no worries. it's quite broken anyway at the moment..
[14:42:43] <KitsuWhooa> okay, stupid question
[14:42:47] <KitsuWhooa> how do I sign a package? :p
[14:43:48] <abaumann> gpg --detach-sign --use-agent -u <pgp key> --no-armor <package>
[14:43:53] <abaumann> or something along the lines
[14:43:57] <KitsuWhooa> oh
[14:43:58] <KitsuWhooa> thanks!
[14:44:03] <KitsuWhooa> I didn't realise I had to use gpg manually
[14:44:18] <abaumann> well. pacman does the same, basically
[14:44:26] <abaumann> or rather makepkg
[14:44:49] <KitsuWhooa> mhm, I see
[14:45:45] <KitsuWhooa> yeah that worked
[14:46:39] <abaumann> cool :-)
[14:47:15] <abaumann> mmh. I can login just fine with ssh into a 486 box
[14:47:55] <abaumann> Apr 08 16:47:37 arch32-stable-i486 systemd-coredump[360]: [LNK] Process 358 (login) of user 0 dumped core.
[14:47:58] <abaumann> Stack trace of thread 358: #0 0x00000000b7dbdf55 n/a (libc.so.6 + 0x8cf55) #1 0x00000000b7d6c413 raise (libc.so.6 + 0x3b413) #2 0x00000000b7d53113 abort (libc.so.6 + 0x22113) #3 0x00000000b7c7324b n/a (pam_systemd_home.so + 0x624b) #4 0x00000000b7c9f595 n/a (pam_systemd_home.so + 0x32595) #5 0x00000000b7c82965 n/a (pam_systemd_home.so + 0x15965)
[14:48:03] <abaumann> #6 0x00000000b7c8531a n/a (pam_systemd_home.so + 0x1831a) #7 0x00000000b7f2c511 n/a (libpam.so.0 + 0x2511) #8 0x00000000b7f2d0b2 pam_end (libpam.so.0 + 0x30b2) #9 0x0000000000465ae6 n/a (login + 0x3ae6) #10 0x00000000b7d54aa5 n/a (libc.so.6 + 0x23aa5) #11 0x00000000b7d54b69 __libc_start_main (libc.so.6 + 0x23b69) #12 0x00000000004666b7 n/a (login + 0x46b7) ELF object binary architecture: Intel
[14:48:05] <KitsuWhooa> ouch
[14:48:09] <abaumann> 80386
[14:48:11] <abaumann> ah, of course
[14:48:17] <abaumann> systemd_home.so is one of the worst PAM modules I have seen.
[14:48:30] <abaumann> Not that PAM itself is a marvell of software, but..
[14:49:55] <abaumann> I really wonder, I would make the removal of pam_systemd_home.so the default, but that would deviate from upstream, so..
[14:50:15] <KitsuWhooa> probably not the best idea, yeah
[14:50:20] <abaumann> I could of course try to fix the bugs in there, bug the moment I start committing to the systemd project..
[14:50:28] <abaumann> *but
[14:50:55] <abaumann> Can you actually commit to an open-source project and in parallel complain about it? :-)
[14:51:42] <KitsuWhooa> yeah I don't know :p
[14:51:51] <KitsuWhooa> as a package maintainer it's really hard to not try to fix issues
[14:51:57] <KitsuWhooa> but if you do, it will take up so much time it'll not be viable
[14:52:45] <abaumann> I thought more about the moral implications.. ;-)
[14:52:49] <KitsuWhooa> oh :p
[14:55:47] <abaumann> oh. euronuc builds sysfsutils.. ;-)
[14:57:16] <KitsuWhooa> once more :p
[14:57:23] <abaumann> :: mkinitcpio and systemd are in conflict. Remove systemd? [y/N] y
[14:57:23] <abaumann> :: mkinitcpio and cryptsetup are in conflict. Remove cryptsetup? [y/N] y
[14:57:23] <abaumann> :: mkinitcpio and mdadm are in conflict. Remove mdadm? [y/N] y
[14:57:23] <abaumann> :: mkinitcpio and lvm2 are in conflict. Remove lvm2? [y/N] y
[14:57:27] <abaumann> urgh. yeah.
[14:59:01] <KitsuWhooa> I added the packages to the manual repo, now for gobject-introspection
[14:59:45] <abaumann> ok.
[14:59:47] <abaumann> cryptsetup c5510bbd7627ce2e1c6938e12cacd9605c9abe21 0000000000000000000000000000000000000000 98ef9b94872802d649be0e5ba770bbd15da4c024 core
[14:59:50] <abaumann> "make_source_info" did not return a "pkgbase" - eh, what?
[14:59:52] <abaumann> >curl: (22) The requested URL returned error: 500<
[14:59:55] <abaumann> *Sigh*
[15:01:27] <abaumann> AFK for some gaming.. back later. :-)
[15:16:00] <KitsuWhooa> https://archlinux32.org
[15:16:02] <phrik> Title: Arch Linux 32 - gobject-introspection 1.76.1-1.0 (pentium4) (at archlinux32.org)
[15:16:08] <KitsuWhooa> have fun!
[15:17:09] <KitsuWhooa> meant to link to https://archlinux32.org
[15:17:11] <phrik> Title: Arch Linux 32 - gobject-introspection 1.80.1-1.0 (i686) (at archlinux32.org)
[15:28:30] <KitsuWhooa> Run-time dependency gobject-introspection-1.0 found: NO (tried pkgconfig)
[15:28:35] <KitsuWhooa> okay this might be enough for today :p
[15:30:47] <KitsuWhooa> oh actually it works with the glib2 from build support manual
[15:30:48] <KitsuWhooa> nvm
[15:31:03] <KitsuWhooa> oh it failed later
[15:31:05] <KitsuWhooa> eh
[15:36:29] <KitsuWhooa> That was a roller coaster. It built with the clean chroot step
[15:59:19] <KitsuWhooa> genesys/gl124.cpp:790:1: internal compiler error: in find_oldest_value_reg, at regcprop.cc:460
[16:34:29] -!- drathir_tor has quit [Remote host closed the connection]
[16:39:51] -!- drathir_tor has joined #archlinux32
[16:40:49] -!- n0tiz- has joined #archlinux32
[16:42:28] -!- eschwartz has quit [Ping timeout: 268 seconds]
[16:42:35] -!- n0tiz has quit [Ping timeout: 268 seconds]
[16:42:35] -!- sunshavi has quit [Ping timeout: 268 seconds]
[16:55:28] -!- eschwartz has joined #archlinux32
[17:11:18] n0tiz- is now known as n0tiz
[17:12:40] -!- abaumann has quit [Quit: leaving]
[17:47:33] -!- ssserpent has quit [Quit: WeeChat 4.2.1]
[20:12:45] mavicaway is now known as mavica
[20:15:27] -!- AtleoS has joined #archlinux32
[21:33:15] -!- finsternis has joined #archlinux32