Lennart Poettering
Lennart Poettering
Sorry, but this is simply not the way this is designed. The idea is that some system service (i.e. display manager) first acquires a session on some seat, and then...
Use "gdm", gnome's display manager. You are trying to use from the shell interfaces that are intended for use by display managers. This cannot work, it's simply not intended for...
(I am going to add an integration test for the pid1 stuff still. Adding integration tests for the vmspawn side of things is a bit more involved, we have nothing...
I added some integration tests now. It doesn't cover the vmspawn side of things (we have zero tests there currently, because of nested VM being a miss on CIs, this...
the subgroup killing stuff has since been merged as PR #38102 already. Since your review points were mostly about that I prepped a follow-up PR for those fixes, see #38136
Force pushed a new rebased version, on top of the now merged #38102, and that hopefully will pass CI. PTAL
force pushed new version, addressing all points pretty much, except of the "pidfdid" naming, see comment about that.
We nowadays have strv_rebreak_lines() which this could just reuse
Force pushed a new version that also fixes https://github.com/systemd/systemd/issues/34416. Let's see if this passes CI
Please have another look!