Report the closed surface's own ref from surface.close

[teamleaderleo]

cmux close-surface --surface surface:N answers OK surface:<N+1-ish> instead of OK surface:N. The new ref points at the surface that was just closed, so it never shows up anywhere and looks like a transient replacement panel. That is one of the two symptoms in #12524.

Cause: closing a tab runs cleanupSurfaceState, which removes the surface's ref from the handle registry. surfaceClose then built the reply with ref(.surface, id), which mints a brand new ref for the dead UUID (and leaves that entry in the registry for good).

Fix: look up the target's existing ref (new non-minting ControlHandleRegistry.existingRef) before closing and use it in the reply. If the closed surface differs from the requested one, or no ref existed, the reply keeps the old behavior.

This does not explain the other, more serious symptom in #12524 (the wrong surface being closed), so the issue stays open.

Tests: ControlCommandCoordinatorSurfaceTests/surfaceClosePayloadReportsTheClosedSurfaceRef (first commit fails without the fix: the reply carries surface:3; second commit is the fix).

Refs #12524

🤖 Generated with Claude Code

---

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

---

Summary by cubic

Fixes surface.close replying with a freshly minted ref (OK surface:<N+1>) instead of the closed surface's own ref.

• Captures the target's ref via new non-minting ControlHandleRegistry.existingRef before teardown removes it, and uses it in the reply.
• Covers all targeting modes: surface_id, terminal_id, and tab_id.
• Falls back to the old minting behavior if no ref existed or the closed surface differs from the requested one.
• This addresses one symptom of close-surface (CLI): killed a different surface than requested; new behavior spawns an empty replacement panel #12524; the other (wrong surface being closed) remains open.

^{Written for commit 71ef081. Summary will update on new commits.}

[coderabbitai]

Warning

Review limit reached

Next included review available in 37 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used all 10 included reviews currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Repository: manaflow-ai/cmux/.coderabbit.yaml

Review profile: ASSERTIVE

Plan: Advanced

Run ID: ca7235d5-70cc-4fed-bcb9-983574d72c00

📥 Commits

Reviewing files that changed from the base of the PR and between a75ab64 and 71ef081.

📒 Files selected for processing (4)

• Packages/macOS/CmuxControlSocket/Sources/CmuxControlSocket/Coordinator/Surface/ControlCommandCoordinator+Surface.swift
• Packages/macOS/CmuxControlSocket/Sources/CmuxControlSocket/Wire/ControlHandleRegistry.swift
• Packages/macOS/CmuxControlSocket/Tests/CmuxControlSocketTests/ControlCommandCoordinatorSurfaceTests.swift
• Packages/macOS/CmuxControlSocket/Tests/CmuxControlSocketTests/FakeSurfaceControlCommandContext.swift

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[github-actions]

Merge receipt for 71ef081f3f, merged 2026-09-25 19:51:30 UTC

• Not verified at merge: ci-status (not reported), Web complexity (in progress), web-validation (not reported)
• Skipped by policy: web-build, web-tests
• Full suite: runs on main after merge.

Labeled merged-unverified: if main breaks near this merge, look here first.