[teamleaderleo]

Base red: SidebarHoverCardEscapeTests/restingInTheSlotShowsNoCard fails on feat-cmux-next. The error is at SidebarHoverCardEscapeTests.swift:133: Caught error: no timer went to sleep on the manual clock within 2.5 s.

The test rests the pointer in the empty slot beside a workspace row. It expects no hover card to show after the hover delay. The test came from #22473, when that point still started a hover timer: the sidebar margin hit-tested as the row, and the slot logic refused the card when the timer fired. Since then the row's hover stops short of the margin, so the slot starts no card and no timer at all. That is the behaviour the test wants. But the test called clock.waitForSleeper() first, and that throws when no timer sleeps. So it failed before it could check that no card shows.

The fix is in the test only. It waits out the delay only when a card is pending (coordinator.machine.activeTarget != nil), and checks that no card shows either way.

Changelog

none

Testing

Focused command: cmux-ci run --class light --script scripts/ci/package-test-lane.sh --ref <sha> --label ci-steps --label suites --arg suite --arg Packages/macOS/CmuxNext --arg SidebarHoverCardEscapeTests

• Red on base e9590e0 (fleet step 613a8a8c, aws-m4pro-8): restingInTheSlotShowsNoCard fails at :133 with the error above; the other 6 tests pass.
• Green on head 4cba4a7 (fleet step 476e2677, aws-m4pro-9): all 7 tests pass.

There is no separate failing-test commit: the red is the existing test on base.

Observed / Not run

• Observed: the two fleet runs above. In the same base run, GroupHeaderNewWorkspaceTests passes at e9590e0, after c6bb043, so it is not part of this PR.
• Not run: I did not bisect the exact commit that stopped the margin from hit-testing as the row. It lies between cmux NEXT sidebar: an empty slot beside a workspace dismisses its hover card #22473 (032f274) and 6d5ca52; the test fails at cmux-next sidebar: hover follows the pointer, row pills stop short of the edge, the row card stands off its row #22733's parent too.

No capture: test-only change, no UI changes.

🤖 Generated with Claude Code

Agent session: cosmic-gray

---

^{Need help on this PR? Tag @codesmith-bot with what you need.}

---

Summary by cubic

Fixes SidebarHoverCardEscapeTests.restingInTheSlotShowsNoCard on feat-cmux-next, which failed because it called clock.waitForSleeper() when the slot now starts no hover timer at all. The test now waits out the hover delay only when a card is pending (coordinator.machine.activeTarget != nil) and checks that no card shows either way.

Testing

• Red on base e9590e04: the test fails with no timer went to sleep on the manual clock; the other 6 tests pass.
• Green on head 4cba4a75: all 7 tests pass.
• Test-only change, no UI changes.

^{Written for commit 4cba4a7. Summary will update on new commits.}

[teamleaderleo]

Owned by browser-pip on cosmic-gray · session

[teamleaderleo]

Bead cx-w7fvi created for this PR by the cmux coordinator.

[teamleaderleo]

Review (subagent, correctness first): SHIP at 4cba4a7

updateHover sets .pending synchronously (the coordinator drains machine.reduce before it returns), so a margin that hit-tests as the row again takes the timer path and fails if a card shows; an idle machine can never reach .show. Test-only; the red on base and the green on head are fleet steps 613a8a8c and 476e2677.

[github-actions]

cmux status _{(one comment, edited in place; each line is updated by the job that owns it)}

Review 4cba4a7: HQ feed card · gallery before/after runs · checks

App of this head (app.zip, kept 1 day)

gh workflow run reload-build.yml -R manaflow-ai/cmux --ref feat-cmux-next -f tag=pr23550 -f ref=base-red-hover-slot -f platform=macos -f backend_mode=local

Not verified: macOS jobs did not run: compile and app tests were skipped on 4cba4a75a6; ci-status is green without them.

CI passes on 4cba4a75a6 (run 39633772513 attempt 1).
macOS jobs did not run: compile and app tests were skipped: Fast static checks skipped (the macos job needs both to succeed). Re-run CI once they pass, before merging.

Written by scripts/ci/classify_failures.py (ci-failure-attribution.yml); signatures are its SIGNATURES table. A machine verdict is the runner's fault, not this PR's; yours means the failing file is one this PR changes, also red on main that main's latest full suite fails the same way, seen on other PRs that it failed on another pull request's run lately.

Tour of 4cba4a75a6 (run): no media. The tour job failed before the tour ran (build or setup); the run log has the output.

[github-actions]

Shipped in:

• cmux NEXT nightly 1.0.0-nightly.3964654862801 (06af62206e)
• cmux NEXT nightly 1.0.0-nightly.3965027402001 (6bc437994c)
• cmux NEXT nightly 1.0.0-nightly.3965153006601 (a10158f013)