#archlinux32 | Logs for 2024-04-19

Back
[00:13:10] -!- bdju has joined #archlinux32
[01:03:14] -!- abouvier has quit [Read error: Connection reset by peer]
[01:07:13] -!- AtleoS has joined #archlinux32
[01:19:05] -!- abouvier has joined #archlinux32
[02:03:35] -!- bill-auger has quit [Ping timeout: 264 seconds]
[02:05:06] -!- bill-auger has joined #archlinux32
[03:05:22] -!- ssserpent has joined #archlinux32
[06:57:59] -!- abaumann has joined #archlinux32
[06:57:59] <buildmaster> Hi abaumann!
[06:57:59] <buildmaster> !rq abaumann
[06:58:00] <phrik> buildmaster: <abaumann> *abaumann does some hackeroo on his rsync-woodo-script..
[06:59:58] <abaumann> pacman -S dbus-broker-units
[06:59:58] <abaumann> error: target not found: dbus-broker-units
[07:00:04] <abaumann> on pentium4 and i686
[07:00:15] <abaumann> this should be in conflict with dbus-daemon-units and provide dbus-units
[07:01:23] <abaumann> and rebuilding systemd, dbus or dbus-broker has a conflict on dbus.service file
[07:01:26] <abaumann> on i486
[07:01:38] <abaumann> I might have to hack some packages and put them into build-support..
[07:09:21] <abaumann> dbus-broker and dbus-broker-units were not pushed to stable, ok.
[07:22:31] <abaumann> switching back from dbus-broker-units to dbus-daemon-units fails in some insetall hooks (failed to restart user@slice) and reboot cannot communicate to the dbus daemon
[07:22:34] <abaumann> I wonder..
[07:23:46] <abaumann> also I have the suspicion I have now both dbus versions on the system, the dbus-daemon version is the activated one..
[07:24:49] <abaumann> default installation of systemd in a chroot, relies on the natural order by package name, so dbus-broker wins of dbus-daemon. Explicit priorisation in pacman would be better..
[07:25:02] <abaumann> *wins over
[07:33:03] <abaumann> (6/7) upgrading systemd [##################################] 100%
[07:33:06] <abaumann> warning: directory permissions differ on /usr/share/polkit-1/rules.d/
[07:33:09] <abaumann> filesystem: 750 package: 755
[07:53:45] <abaumann> mmh. testing vms hang when running udev hooks, so it might be a dbus/systemd issue
[07:57:13] <abaumann> GLIBC_2.38 not found, aha. it _is_ a glibc mismatch
[08:00:52] <abaumann> but libc.so.6 says it is a 2.38, mmh
[08:02:37] <abaumann> ah. no.
[08:02:51] <abaumann> I booted from the ISO and this one is 2.36, stupid me. :-)
[08:15:55] <abaumann> uiui. bin is a link to usr/bin
[08:16:07] <abaumann> lib not and contains lib/cmake and lib/qml
[08:16:27] <abaumann> usr/lib contains kind of an important file /lib/ld-linux.so.2
[08:16:55] <abaumann> so, something cmakisk or qtist graeatly destroys the file structure
[08:17:37] <abaumann> cmake contains KQtQuickCharts and KF6BreezeIcons
[08:19:22] <abaumann> mv lib/cmake/* usr/lib/cmake/.
[08:19:22] <abaumann> mv lib/qml /usr/lib/.
[08:19:22] <abaumann> rmdir lib/cmake
[08:19:22] <abaumann> rmdir lib
[08:19:22] <abaumann> ln -s usr/lib lib
[08:19:32] <abaumann> but, this should not get to stable.. really not.
[08:20:42] <abaumann> so the curplrit is some KDE things.
[08:21:09] <abaumann> fixed and booting still hangs, arch-chroot from ISO works. mmh.
[08:24:11] <abaumann> ah. Rebuilding Dynamic Linker Cache. Yes, that will take a while if /lib is now poiting to /usr/lib again
[08:26:37] <abaumann> looking good. and we have an emergency instruction..
[08:28:08] <abaumann> mmh. this could be a case for some sanity hooks in pacman:
[08:28:26] <abaumann> check if nothing vioales things set up in the 'filesystem' package.
[08:28:43] <abaumann> permission changes in polkit. removing essential symlinks or/and files, etc.
[08:36:15] <abaumann> flatpak: symbol lookup error: /usr/lib/libappstream-glib.so.8: undefined symbol: g_once_init_leave_pointer
[08:36:18] <abaumann> on login.
[08:36:19] <abaumann> cool
[08:37:11] <abaumann> my test machines are funny: the have everything installed I ever tested or fixed, so they contain everything and the kitchen sink. :-)
[08:46:14] <abaumann> kqtquickcharts: /lib exists in filesystem (owned by filesystem)
[08:46:14] <abaumann> kqtquickcharts: /lib/cmake/KQtQuickCharts/KQtQuickChartsConfig.cmake exists in filesystem
[08:46:17] <abaumann> kqtquickcharts: /lib/cmake/KQtQuickCharts/KQtQuickChartsVersion.cmake exists in filesystem
[08:46:20] <abaumann> oh. now it complains?
[08:48:07] <abaumann> mmh, really weird, so this should never have happened on testing..
[10:00:35] <abaumann> so, I think things start to work correctly now around the dbus/dbus-broker thingy on all subarchs.
[10:08:07] -!- abaumann has quit [Quit: leaving]
[13:31:11] -!- ssserpent has quit [Quit: WeeChat 4.2.2]
[13:32:30] -!- ssserpent has joined #archlinux32
[14:30:45] <KitsuWhooa> abaumann: Since you mentioned it, I've been thinking. It should be relatively easy to add a fake /proc/cpuinfo and uname in the chroots
[14:30:53] <KitsuWhooa> both using bind mounts, and a simple shell script
[14:31:38] <KitsuWhooa> the fake /proc/cpuinfo would be generated based on the number of cores in the host, but with a i486, a pentium3 and a pentium4
[14:31:56] <KitsuWhooa> and uname will just be a shell script that prints predetermined responses taken from actual machines
[14:33:14] <KitsuWhooa> so, make chroot, generate fake cpuinfo, copy fake uname, bind mount both -> build -> unmount both
[14:33:45] <KitsuWhooa> won't help with programs looking at CPUID but it should be good enough for most cases
[14:36:24] <KitsuWhooa> maybe a fake cpuid command, too
[14:56:59] <KitsuWhooa> and /sys/devices/system/cpu/modalias
[14:57:19] <KillerWasp> ..
[14:58:21] <KillerWasp> hi, sorry by the dots
[15:22:05] <KitsuWhooa> it's okay
[15:22:13] <KitsuWhooa> abaumann: glibc is still failing on i486
[19:20:01] -!- ssserpent has quit [Quit: WeeChat 4.2.2]
[20:25:29] -!- AtleoS has quit [Ping timeout: 268 seconds]
[20:26:02] -!- AtleoS has joined #archlinux32
[21:13:12] -!- drathir_tor has quit [Remote host closed the connection]
[21:18:45] -!- drathir_tor has joined #archlinux32