Cole Robinson (066aa232) at 28 Feb 14:08
Moving libvirt-ci to bash-completion-devel for rawhide: libvirt-ci!463 (closed)
The .pc file moved to the bash-completion-devel subpackage.
Signed-off-by: Cole Robinson crobinso@redhat.com
Cole Robinson (de8c34c7) at 25 Feb 21:17
facts: rawhide needs bash-completion-devel
Cole Robinson (6c752094) at 20 Feb 15:40
spec: Handle new location of bash-completion .pc file
... and 2 more commits
I was being lazy and relying on CI to tell me if it's an error
Looking at internal builds of bash-completion-1:2.7-4.el8.noarch.rpm from 2018, it does have Provides: pkgconfig(bash-completion) = 2.7
, and scripts/pkgconfigdeps.sh
has been in rpm.git since 2007 so I think we are safe
On Fedora 41, bash-completion's .pc file moved to
bash-completion-devel
.
Using pkgconfig()
lets us handle this without distro version checks
This isn't new RPM magic, Fedora systemd package has been using it since Feb 2020 https://docs.fedoraproject.org/en-US/packaging-guidelines/PkgConfigBuildRequires/
Cole Robinson (15d472cc) at 19 Feb 16:57
spec: Handle new location of bash-completion .pc file
Cole Robinson (c5625d09) at 03 Feb 19:27
(crobinso) kernel-snp-coconut rpm build tweaks
... and 4186 more commits
@berrange suggested this elsewhere, but it would help if kernel+qemu+libvirt packages had differentiated RPM version strings from official repo builds. So rather than qemu-kvm-8.0.0-8.el9s.x86_64
we would have something like qemu-kvm-8.0.0-8.tdx.el9s.x86_64
etc.
Note in virt-manager we have an EOL heuristic: https://github.com/virt-manager/virt-manager/blob/0fb7377abb5fcecfaf439a68dc4dfd8aa1d5bcfd/virtinst/osdict.py#L323
Explicitly marked rolling
distros are never EOL. Distros without eol date are given a 10 year window which is pretty generous. But when we added the heuristic it gave reasonable results.
Maybe osinfo-db could add some kind of comment annotation to mark that eol date is explicitly omitted, and then add a unit test case to check for it or a valid date, or rolling distro, etc
Patches from: https://github.com/intel/qemu-tdx/tree/tdx-qemu-2023.08.15-v8.1-rc0-match-with-kvm-upstream-2023.08.10-v6.5-rc5-workaround
On top of qemu-kvm-8.0.0-13.el9
Scratch build: https://kojihub.stream.rdu2.redhat.com/koji/taskinfo?taskID=2777093
I'm not sure if this should wait for a new kernel build or not? Intel folks said the qemu branch was tested with kernel kvm-upstream-2023.08.10-v6.5-rc5-workaround
Cole Robinson (85c2923b) at 29 Aug 15:24
spec: Provides: qemu-kvm-tdx
Cole Robinson (273ab84d) at 29 Aug 15:23
linux/kvm.h: Bump KVM_EXIT_MEMORY_FAULT & KVM_EXIT_TDX
... and 155 more commits
Cole Robinson (7e7a3155) at 17 Aug 15:20
Add edk2-ovmf-coconutsvsm subpackage
Cole Robinson (d03253d9) at 17 Aug 15:16
Add edk2-ovmf-coconutsvsm subpackage
... and 200 more commits