Hey, I've talked a bit about it on IRC, but I wanted there to be a real record: Recently, some of my contributions to Parabola haven't... been to Parabola.
# glibc There was a bug in glibc that caused problems in systemd-nspawn that caused problems in devtools that caused problems in libretools. I wrote a patchset for it, submitted it upstream, got it on master, and then it got backported to the 2.26 stable backports branch (we have breakage in Fedora to thank for speedy backporting). And then that filtered back down to Arch glibc! bug upstream https://sourceware.org/bugzilla/show_bug.cgi?id=22145 bug Arch https://bugs.archlinux.org/task/56529 bug Fedora https://bugzilla.redhat.com/show_bug.cgi?id=1468837 patchset master https://sourceware.org/ml/libc-alpha/2017-11/msg00546.html # QEMU The old libre/qemu-static 2.8 was increasingly complaining about unknown syscalls. So I went to upgrade it to the latest qemu, 2.11.0. Now, I discovered a major performance regression in qemu-arm when statically linked. As in 3-second startup time before it even starts emulation. I filed a bug upstream, and wrote a patchset, which is applied to the libre/qemu-user-static package. The patchset has not yet been reviewed upstream. I expect this to attract more attention once one of the major distros upgrades from 2.10. bug https://bugs.launchpad.net/qemu/+bug/1740219 patchset https://lists.nongnu.org/archive/html/qemu-devel/2017-12/msg05237.html The new qemu-user-static package hasn't yet hit i686 yet because I'm waiting for util-linux 2.31 -> 2.31.1 upgrade from archlinux32. -- Happy hacking, ~ Luke Shumaker _______________________________________________ Dev mailing list [email protected] https://lists.parabola.nu/mailman/listinfo/dev
