[teamleaderleo]

Leo: "why are changes to these not in my current nightly". The freshness rule is that each feed's newest published source stays within 30 minutes of its branch's newest landing.

A failed-jobs rerun retains the old picker result. Run 41253243235, attempt 2, queued its DMG image on the Blacksmith alternate after the app build and notarization succeeded. The three secret-free Mac jobs now rejoin the configured owned nightly pool, then the owned side pool, before Blacksmith once on attempt 2 retaining attempt 1's picker. Later reruns retain the existing Blacksmith alternate, so an unavailable fleet can overflow. Fresh picks keep precedence, the fleet switch still applies, and Xcode selection follows the chosen route. The secret-bearing DMG signer is unchanged.

The affected jobs compile nothing and reuse the built app and DMG artifacts; their existing warm-source exemptions still apply. Prior art: #26891 established current-attempt owned selection and matching Xcode pins; the existing legacy verify already has a pooled fallback. #27444 concerns classic's separate picker and is not changed here.

Validation: fleet tests pending. Regression cases execute all three runner and Xcode expressions for stale/fresh picks, attempts 2/3/4, disabled fleet routing, missing nightly/side pools, other branches, and forks.

Performance:

Metric	Before	After	Delta	Harness
Secret-free Mac jobs forced to Blacksmith on a stale-picker rerun with a configured fleet pool	3	0	-3	runner-expression regression cases
Blacksmith minutes/day	Not measured	Not yet measured after deployment	No measured claim	No new build or test work; eligible reruns move to fleet

Agent session: agent session
🤖 Generated with Claude Code

---

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

---

Summary by cubic

Fixes nightly CI reruns so secret-free Mac jobs on nightly-next with fleet routing enabled rejoin the owned fleet pool on attempt 2 instead of queuing on the Blacksmith alternate with attempt 1's stale picker.

On a failed-jobs rerun, attempt 2 now uses the configured owned or side pool runner and the matching PR Xcode pin. Attempts 3+ keep the existing Blacksmith overflow, and fresh runs keep their normal pool selection. Adds regression tests covering stale and fresh picks, attempts 2/3/4, disabled fleet routing, missing pools, other branches, and forks.

^{Written for commit 73b71e5. Summary will update on new commits.}

[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.

[github-actions]

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

Review a338688: 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=pr29847 -f ref=lane/nightly-verify-rerun-fleet -f platform=macos -f backend_mode=local

Shipped in:

• cmux NEXT nightly 1.0.0-nightly.4130052256501 (28aa638015)

[cubic-dev-ai]

1 issue found across 2 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name=".github/workflows/nightly.yml">

<violation number="1" location=".github/workflows/nightly.yml:5048">
P2: The stale-rerun route is a static `||` of two vars, so the side pool is only reached when `CI_NIGHTLY_NEXT_BUILD_RUNNER` is empty. It does not try the nightly pool and then the side pool. If the nightly label is set but has no idle runner, the job queues there until the owned-pool rescue moves it to Blacksmith. Either add an idle check on this path, as the picker does, or correct the PR description so it does not claim a try-then-fallback order.</violation>
</file>

<sub>Reply with feedback, questions, or to request a fix.

View guided diff | Turn on auto-fix | Re-trigger cubic</sub>

[cubic-dev-ai]

P2: The stale-rerun route is a static || of two vars, so the side pool is only reached when CI_NIGHTLY_NEXT_BUILD_RUNNER is empty. It does not try the nightly pool and then the side pool. If the nightly label is set but has no idle runner, the job queues there until the owned-pool rescue moves it to Blacksmith. Either add an idle check on this path, as the picker does, or correct the PR description so it does not claim a try-then-fallback order.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. At .github/workflows/nightly.yml, line 5048:

<comment>The stale-rerun route is a static `||` of two vars, so the side pool is only reached when `CI_NIGHTLY_NEXT_BUILD_RUNNER` is empty. It does not try the nightly pool and then the side pool. If the nightly label is set but has no idle runner, the job queues there until the owned-pool rescue moves it to Blacksmith. Either add an idle check on this path, as the picker does, or correct the PR description so it does not claim a try-then-fallback order.</comment>

<file context>
@@ -5043,7 +5043,9 @@ jobs:
-    runs-on: ${{ github.repository_owner != 'manaflow-ai' && 'macos-26' || needs.pick-nightly-sign-runner.outputs.attempt == github.run_attempt && needs.pick-nightly-sign-runner.outputs.owned || github.run_attempt > 1 && needs.pick-nightly-sign-runner.outputs.alternate || needs.pick-nightly-sign-runner.outputs.runner || vars.MACOS_RUNNER_26 || 'blacksmith-6vcpu-macos-26' }}
+    # A failed-jobs rerun retains an older picker. Rejoin the fleet pool, never
+    # that old runner choice, once; later attempts use the Blacksmith overflow.
+    runs-on: ${{ github.repository_owner != 'manaflow-ai' && 'macos-26' || needs.pick-nightly-sign-runner.outputs.attempt == github.run_attempt && needs.pick-nightly-sign-runner.outputs.owned || github.ref_name == 'nightly-next' && github.run_attempt == 2 && needs.pick-nightly-sign-runner.outputs.attempt == 1 && vars.CI_PR_POOL_OWNED == '1' && (vars.CI_NIGHTLY_NEXT_BUILD_RUNNER || vars.CI_SIDE_LANE_RUNNER) || github.run_attempt > 1 && needs.pick-nightly-sign-runner.outputs.alternate || needs.pick-nightly-sign-runner.outputs.runner || vars.MACOS_RUNNER_26 || 'blacksmith-6vcpu-macos-26' }}
     timeout-minutes: 40
     strategy:
</file context>

[blacksmith-sh]

Found 2 test failures on Blacksmith runners:

Failures

Test	View Logs
test_only_the_nightly_pickers_read_the_next_build_runner (main.NightlyTrustedRoutin g.test_only_the_nightly_pickers_read_the_next_build_runner)/ test_only_the_nightly_pickers_read_the_next_build_runner (main.NightlyTrustedRoutin g.test_only_the_nightly_pickers_read_the_next_build_runner)	View Logs
test_the_repository_highlights_validate (main.DigestTests.test_the_repository_highl ights_validate)/ test_the_repository_highlights_validate (main.DigestTests.test_the_repository_highl ights_validate)	View Logs

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

[teamleaderleo]

Review (subagent opus): SHIP at a338688

On attempt 2 of a nightly-next failed-jobs rerun, verify-signed-nightly, verify-final-dmg-nightly and signer-dmg-image rejoin the fleet nightly or side label with the fleet Xcode pin, instead of the old picker's single-host or alternate runner. Attempt 3 or later, a pool switched off, or another ref keep the earlier routes. The new table test evaluates each branch of runs-on and of CMUX_CI_XCODE_APP.

Follow-ups: the long runs-on expression is now copied verbatim into three jobs; a YAML anchor or a picker output would keep them in step.

[teamleaderleo]

Review (independent Codex gpt-6.1-sol, high, via the review router; full diff from merge base 905d02d to this head): SHIP at a338688.

No P1/P2 findings.

[teamleaderleo]

Review (subagent): SHIP at a338688

[teamleaderleo]

Review (subagent): SHIP at 73b71e5

Reviewed the PR diff against the base-merge second parent. The fleet retry and Xcode selection change is preserved; no code blockers.