[teamleaderleo]

Media tours can now run several at once on one Mac, each capturing only its own app's window. That lets the runner hook drop its one-tour-per-Mac limit.

What changes:

• The tour no longer activates its app, and launches it with CMUX_NEXT_NO_ACTIVATE=1. It drives the app through the app's own control socket and screenshots it with debug.window_snapshot, so being in front never mattered. Concurrent tours would otherwise trade focus and the key window back and forth.

• The fallback captures the window, not the screen. A build without debug.window_snapshot used to be captured with screencapture -R, a rectangle of the display, which would include whatever covers the window, such as another tour's app. It now captures that window alone: screencapture -l <id>, or the screen-capture helper's window <id>. The window server id comes from the window list by the app's pid. With no id there is no shot, never the display. The manifest records capture_window. This route leaves out child windows (sheets, popovers) that debug.window_snapshot includes; it is only for old bases without that verb.

• A tour holds its tag (scripts/ci/media_tag_lock.py, new step "Hold this tour's tag"). One tag's warm DerivedData, sockets and daemon session take one tour at a time; other tags never wait. The concurrency group already keeps two runs of a pull request apart, but not these cases:

  • a cancelled run that is still stopping;
  • a dispatch;
  • a merge tour beside a dispatch, which share the feat-cmux-next tag.

  The lock lets go when the job's Runner.Worker exits, or after 3 h. A second tour of a tag waits up to 60 min: longer than a warm merge tour holds the shared tag (54 min seen), while leaving the waiter 45+ min of its 120-min job for a cold build and the tour. A longer wait would only end in a job-timeout cancel instead of the "still held" error. When the pull request head predates the lock, the step takes the lock script and the lock-aware prune from the run's commit, so unrebased PRs still run and their prune never deletes a held tag.

• The DerivedData prune skips a held tag and holds a tag's lock itself until its dir is gone, so a tour never builds into DerivedData that is being deleted, or has its DerivedData deleted by another pull request's tour.

What still needs the console, and stays on the console runner, is unchanged here: app-host XCTest (testmanagerd), E2E video recording and frame-pacing benches.

Validation

• tests/test_cmux_next_tour.py and the new tests/test_media_tag_lock.py (linux-guard lane), red then green:
  • At 3b6b740 (tests only) against the old tour.py and with no lock script, test_the_tour_never_activates_the_app and test_without_a_window_snapshot_the_tour_uses_a_screen_route fail, and so do 3 of the 4 lock tests.
  • At ff2f66f all 4 lock tests pass, and so do the tour's activation, screen-route and app-window tests.
  • At 1286a4a (tests only) the prune test fails: the old prune has no --lock-dir. At 7468a34 it passes, and so do the lock and tour tests.
• actionlint .github/workflows/cmux-next-media.yml is clean.
• At 071bcb0 (test only) the wait test fails against the 7200 s wait (7200 not less than or equal to 4500); at bd085a7 it passes. tests/test_media_tag_lock.py was in the linux-guard lane but no workflow ran it; ci-guards.yml now runs it in the release-notary group. All four tour/lock/prune tests pass on Huge Blue at bd085a7.
• Screenshots now show the app's window as not key (CMUX_NEXT_NO_ACTIVATE), so focus styling looks inactive in tour media.
• Missing ExplicitPrecompiledModules/Darwin-*.pcm tour failures (14 of 22 recent failed tours) are not from the prune: every one I checked (jobs 118285853270, 118197207461, 118076888558, 118006980751) followed media_module_cache_stamp.py clear removing ExplicitPrecompiledModules after a binary pin change, while the build's dependency records still point at those .pcm files. The prune only removes whole idle cmux-ci-media-* dirs, and now only under the tag lock. fix(cmux-next media): a pin change also drops Xcode's build description #22181 fixes that clear.
• Base reds on this head, not from this PR: release-notary fails NEXT post-publication dependency contract the same way on feat-cmux-next (job 118292665776), and cmux-next checks fails the same six checks (state roots, concurrency, strings, type scale, script tests, web bundle key) as merged agent pane: permission UI keeps its last state across a chat switch (no blink) #22054 (job 118309312276), being fixed in cmux-next: clear every base red in cmux-next checks (type scale, strings, tree key, state root, concurrency, bundle key) #22117.
• Old PR heads run their own tour.py, which still activates the app and captures a rectangle, until they merge main.
• Not run yet: a tour on a Mac with two tours in flight. The runner-side change that lets them share a Mac follows this PR.

Runtime test: tests/test_media_tag_lock.py, tests/test_ci_prune_media_derived_data.py and tests/test_cmux_next_tour.py in the release-notary/preflight guard jobs at bd085a7

Changelog

none

Review: SHIP at bd085a7 (cmux-review subagent; correctness first)

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

• Configuration used: Repository: manaflow-ai/cmux/.coderabbit.yaml
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: 040b1e8d-ece4-4e33-a1c2-ce3790349775

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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]

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

Review 5bfa69b: 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=pr22046 -f ref=ci/tour-concurrent -f platform=macos -f backend_mode=local

[teamleaderleo]

Bead cx-jf8t2 created for this PR by the cmux coordinator.

[teamleaderleo]

Review (subagent, correctness first) on f0fe23e, and what 7468a34 does about each finding:

1. Unrebased PRs would fail the new hold step, since their head has no media_tag_lock.py. Fixed: the step fetches the script from the run's commit when it is missing.
2. screencapture -l leaves out child windows (sheets, popovers). Kept, and documented: it is only the last-resort route for bases without debug.window_snapshot, which includes them.
3. The window lookup could pick an off-screen or helper window. Fixed: on-screen windows only, layer 0, matching the app's frame within 1 pt, else the largest.
4. The DerivedData prune could delete a tag another tour is about to build into. Fixed: prune skips any tag whose lock is held (new scenario_held_tag, red at 1286a4a). The hold waits at most 30 min.
5. Test holders outlived their tests. Fixed: they now let go after 4 s.

defaults-snapshot is red from the base, not this PR.

[blacksmith-sh]

Found 1 test failure on Blacksmith runners:

Failure

Test	View Logs
NEXT post-publication dependency contract/NEXT post-publication dependency contract	View Logs

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

[greptile-apps]

Confidence Score: 4/5

[Medium impact] The changes affect shared Mac resources used by the cmux-next media-tour workflow.

Fix the display-shot caller before merging; it currently stops tours that request a display shot.

Findings

1.  Display shots stop tours ▶

Summary

Allows media tours to share one Mac while keeping each tour's captures and build files separate.

• Tours leave focus alone and capture only their app’s window.
• Tours of one tag take turns while different tags run together.

Both earlier unnumbered findings are fixed: the prune keeps its lock through deletion, and older heads receive both lock-aware scripts.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Tour starts] --> B[Acquire tag lock]
  B --> C[Prune idle DerivedData]
  C --> D[Build base and head]
  D --> E[Launch without activation]
  E --> F[Capture app window]
  F --> G[Job ends and releases lock]
  C --> H[Skip tags held by other tours]

Loading

_{Reviews (6) · Last reviewed commit: "Merge branch 'feat-cmux-next' into ci/to..." · Reviewed by Greptile}

[cursor]

Bugbot needs on-demand usage enabled

Bugbot uses usage-based billing for this team and requires on-demand usage to be enabled.

A team admin can enable on-demand usage in the Cursor dashboard.

[teamleaderleo]

Pushed 82b9d03 (merges feat-cmux-next 86efbba):

• d013c33 / 41f9b19 (red first): a tour waits for a held tag up to the job's 120-min timeout, not 30 min. Merge tours ran up to 54 min in the last week.
• c40e202 / a77a158 (red first): an unrebased head takes the lock-aware prune from the run's commit too (greptile's second thread).
• 1b8a89f / e7d0698: the prune holds the tag lock until rmtree returns. Its test was red before the fix.

Reds on 1b8a89f that aren't this PR's:

• nightly-guards / workflow-guard-tests / release-notary ("FAIL: NEXT post-publication dependency contract") also fails on feat-cmux-next 86efbba (job 118292665776).
• cmux-next checks failed on the cmux-tui tree key missing schemas/settings/source/subsessions.json, which 020a352 added on the base. cmux-next: clear every base red in cmux-next checks (type scale, strings, tree key, state root, concurrency, bundle key) #22117 clears it.

Local runs on Huge Blue: test_media_tag_lock.py, test_ci_prune_media_derived_data.py and test_cmux_next_tour.py all pass. actionlint is clean.

[teamleaderleo]

For whoever lands this: #22174 is built on this branch's head (82b9d03) and replaces the per-PR tag with a per-Mac slot claimed through this PR's lock (media_tag_lock.py claim, tags ci-media-slot<N>). It doesn't change hold or the prune's lock handling, so please land this one first and I'll merge feat-cmux-next into #22174 afterwards. Also #22170 fixes the 'module file …Darwin-*.pcm not found' tour failures (14 of 22 today). Those come from #21971's partial clear, not from this PR.

[teamleaderleo]

Re-review (subagent, correctness first) of the commits since the last review: no defects. SHIP at .

[greptile-apps]

Display shots stop tours

Screen.use_display() now requires pid, but display_capture() still calls it without arguments. When a screen capture route is available, a step with display: true raises an unhandled TypeError and stops the tour. The standard core.json tour hits this through its Profile menu step, even though that step is optional.

Update display_capture() to pass the app's PID.

[github-actions]

Shipped in:

• cmux NEXT nightly 1.0.0-nightly.3936775151801 (d7e81d2c47)
• cmux NEXT nightly 1.0.0-nightly.3937100371101 (f4c2e82984)