Fix failing core ptest in systemd and libguestfs

[kgodara912]

Summary

Fix failing %check (ptest) stages for systemd and libguestfs in the Azure Linux build chroot. No upstream version changes — release-bump + targeted in-spec test fixes only.

Changes

Package	Fix
systemd	Skip tests in %check that require capabilities not available in the build chroot
libguestfs	Switch make quickcheck to the direct backend with force_tcg; extend supermin appliance with sqlite*

Details

systemd

Constructs an explicit test list and excludes tests that fail in a chroot but pass in a real environment:

• test-fd-util, test-mount-util, test-mountpoint-util, test-path-util, test-time-util, test-calendarspec, test-date — need mount-namespace privileges (MS_SLAVE remount fails with EINVAL in the chroot).
• test-udev — invokes systemd-detect-virt, which is not on PATH yet at this stage.
• test-rpm-macros — environmental mismatch with the build chroot.

libguestfs

• Force the direct libguestfs backend with force_tcg in %check so make quickcheck no longer requires a running libvirtd or /var/run/libvirt/libvirt-sock, and does not need /dev/kvm.
• Add sqlite, sqlite-libs, sqlite-devel to the supermin appliance via --with-extra-packages so guestfsd (which depends on libsqlite3 through librpm) can start inside the appliance.
• Make the QEMU log dump on failure non-fatal when the log directory is missing.

Verification

• Both Release: bumps to the packages
• No upstream version changes
• No toolchain manifest update needed (neither systemd nor libguestfs appears in any toolchain / pkggen_core manifest; systemd-bootstrap-250.3-19 in the manifests is a separate SPECS/systemd-bootstrap/systemd-bootstrap.spec, not affected)

Related

#17749 ("Fix failing core ptest across packages")

[kgodara912]

Buddy build

[kgodara912]

Buddy build for systemd after the recent changes in systemd.

[Copilot]

Pull request overview

This PR updates Azure Linux RPM specs to fix failing %check (ptest) stages for systemd and libguestfs in the build chroot by adjusting test execution behavior and bumping Release values to produce rebuilt packages.

Changes:

• systemd: bump Release and run a filtered meson test list to skip chroot-incompatible unit tests; add versioned reciprocal Conflicts for standalone-* subpackages.
• libguestfs: bump Release, switch %check to the direct backend with force_tcg, and add extra packages to the supermin appliance; make QEMU log dump non-fatal when missing.
• systemd-boot-signed: bump Release to match the unsigned systemd release.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
SPECS/systemd/systemd.spec	Release bump; adjust %check to skip known chroot-incompatible tests; add reciprocal conflicts for standalone subpackages.
SPECS/libguestfs/libguestfs.spec	Release bump; tweak %check backend settings; extend supermin appliance package set; make log dump non-fatal.
SPECS-SIGNED/systemd-boot-signed/systemd-boot-signed.spec	Release bump to keep signed sd-boot package aligned with systemd release.

[Kanishk-Bansal]

LGTM,

in libguestfs forcing the direct backend with force_tcg is a nice minimal approach, since we don't have libvirtd / /var/run/libvirt/libvirt-sock (or /dev/kvm) in the build chroot.

in systemd the pakcage conflict refactor is a good add on which is unrelated to ptest fix, we are basically skipping the test although these passes in non chroot env, but due to env constraints we are skipping it here.

[mfrw]

🥇

[kgodara912]

Buddy build for this is successful. The PR check failures are also not there for systemd and libguestfs. All the changes are related to tests. The PR ensures that any future genuine failures for these packages, we check properly instead of assuming that it was always failing. @PawelWMS, could you please check and if looks good, could you please approve it?