Repository navigation
fix(e2e): keep desktop test controls working - #1108
Conversation
Run the input daemon under the user service manager so it survives SSH session cleanup. Hive-Run: projectbluefin#1059 Hive-Plan: issue 1059 approved workflow change Hive-Spec: issue-1059#daemon-lifetime Assisted-by: openai-codex/gpt-6.1-terra via omp Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: Karibhou <Karibhou@users.noreply.github.com>
|
@Karibhou this PR is back in your court ( Review asks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Your relay session received this as a |
|
Removed the closing keyword from the PR description; #1059 remains referenced as partial work. No code change was needed. 🐝 Hive Agent: |
59544ee
|
There was a problem hiding this comment.
Summary: This change moves the small helper that sends clicks and keystrokes to the desktop under test out of the remote login session and into the user's background service manager, so it survives when that login connection expires. I read the change and the surrounding setup step and it does what it says. One low-priority note: the fixed service name chosen here is the same one the upstream project installs, so the day the disk image ships that helper natively, this step would refuse to start.
correctness
.github/workflows/e2e.yml:1311—--unit=gnome-ponytail-daemonclaims a fixed unit name. The build at.github/workflows/e2e.yml:1264passes-Dsystemd_user_unit_dir=/usr/lib/systemd/user, i.e. upstream installs its own user unit. If the image starts shipping the daemon (tracked by PR #618 per the comment at:1248;scripts/guest_preflight.py:67-69already looks for/usr/libexec/gnome-ponytail-daemon),systemd-runfails with "Unit … already exists" and, underset -eat:1296, aborts setup — no fallback. A name that cannot collide with a packaged unit avoids that:
systemd-run --user --quiet --collect --unit=e2e-ponytail-daemon \
Verified by reading: \$HOME escaping resolves correctly (runner double-quote → remote single-quote → inner bash -c), source /tmp/session.env inside the unit gives the daemon the Wayland/D-Bus env, and a systemd-run failure stops the step as the PR body states. Not verified: the logind/ControlMaster kill mechanism itself — no VM run was possible here either, as the body notes.
No findings from: security, intent-alignment, style, docs-currency.
Confidence: 5/5 (safe)
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Needs a decision, not a merge:
Head: verdict approved @59544ee, 6/6 checks green, sentinel alert, soak 14m0s of 30m0s. What to do:
Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
Danathar
left a comment
There was a problem hiding this comment.
Sentinel reviewed (sensitive_path only). e2e.yml now starts gnome-ponytail-daemon under systemd-run --user instead of nohup, so logind no longer kills it when the SSH session closes. This is the fix testsuite#1059 asks for.
…wn (#1112) * fix(e2e): start ponytail under a unit name the image cannot already own #1108 starts gnome-ponytail-daemon with systemd-run --user --unit=gnome-ponytail-daemon. On dakota:testing that fails before any test runs: "Unit gnome-ponytail-daemon.service was already loaded or has a fragment file." Every GUI and lifecycle shard on main now fails at "Install gnome-ponytail-daemon" (run 42124147912: smoke-a, smoke-b, smoke-firefox, lifecycle), which is what keeps v1 held (#1087). Use the unit name testsuite-ponytail-daemon instead, and fall back to the previous nohup start if the user manager refuses the transient unit for any other reason. The bus-name wait that follows is unchanged. Refs #1087, #1108 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SnzFs53Qcer1xDySyxv1Vj Signed-off-by: Danathar <berths-purger-0g@icloud.com> * fix(e2e): drop the nohup fallback for the ponytail daemon The e2e skill doc requires setup to fail when the transient user unit cannot start: a nohup'd daemon in the SSH session is killed by logind when the ControlMaster connection expires, leaving a stale D-Bus owner. Keep the unit rename and let systemd-run failures fail setup again. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SnzFs53Qcer1xDySyxv1Vj Signed-off-by: Danathar <berths-purger-0g@icloud.com> --------- Signed-off-by: Danathar <berths-purger-0g@icloud.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Summary: Desktop tests can lose the service that sends clicks and keystrokes when their remote connection expires. This change keeps that service running for the duration of the tests, so later checks can still interact with the desktop.
Why
Dakota terminates processes in the SSH session when its ControlMaster expires. The input daemon must run in the user service manager instead. The service sources the desktop session environment before starting; setup fails if the service cannot be submitted rather than silently reverting to the vulnerable SSH session.
The workflow skill documents the lifetime requirement. PR #1037 also touches the workflow and skill file and will need its conflicts resolved when integrating.
Validation
bash -n.ruff check tests/ --select E,F,W --ignore E501passed.python3 -m pytest tests/unit/ -q: 2602 passed, 4 subtests passed.python3 scripts/validate_docs.py,python3 scripts/generate_skill_index.py --check, andjust list-stubspassed./dev/kvmis absent.Refs #1059 — the digest-pinned Dakota smoke run and the remaining Power status color image issue still need verification or resolution.
— hive: backend=omp model=openai-codex/gpt-6.1-terra
🐝 Hive Agent:
contributor| SHA:59544ee1