Repository navigation
ci(nightly): the sign picker and the visual gate router take the all-nightly reservation first - #31642
Conversation
…nightly reservation first On attempt 1 of a nightly-next run with vars.CI_NIGHTLY_ONLY_RUNNER set, both jobs now start on glaeda-nightly-only-m4pro, as decide and the resolvers already do. Both only run Python and bash pickers and mint the routing App token, which decide already mints there. Blacksmith's 16-vCPU Linux label was queueing them 5-14 min (runs 41373656988, 41374257749), and the sign picker sits on the path to the feed. Both leave BLACKSMITH_LINUX_ONLY_DEBT. The sign picker's last route is now the literal blacksmith-4vcpu-ubuntu-2404 its fromJSON(...)[0] evaluated to. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-f724bb1ff9295c18fee7
|
cmux status (one comment, edited in place; each line is updated by the job that owns it) Review App of this head (app.zip, kept 1 day)
Shipped in:
The feat-cmux-next push run https://github.com/manaflow-ai/cmux/actions/runs/41376579738 at 765d3c4 failed: cmux-next fast tier (cargo), cmux-next fleet Release compile. |
|
Found 4 test failures on Blacksmith runners: Failures
|
|
Review (subagent claude-opus-5-5 medium): SHIP at 5e75d95 The runs-on guard matches Follow-ups (tracked in #31420):
|
…two nightly.yml routing jobs in warm-sources.json) (#31909) - release-notes/next/highlights/attachment-drag-follows-pointer.md (added by 720c301, #31582) had no tryIt and no docs line, so scripts/whats-new/test_digest.py test_the_repository_highlights_validate failed with "needs a tryIt action/deeplink or a docs link". It now links https://cmux.com/docs/agents, as the other composer highlights do. - 34446e1 (#31642) let pick-nightly-sign-runner and route-visual-gate-nightly take glaeda-nightly-only-m4pro, so tests/test_ci_warm_sources.py counts them as mac-runner jobs that are not in .github/warm-sources.json. They are routing controls that compile nothing; they get the same exempt entry as pick-nightly-build-runner. Before: test_digest.py 1 failure, test_ci_warm_sources.py 1 failure (at fc013e8). After: both OK.
![Fix with [code]smith](https://pr-comments-assets.blacksmith.sh/codesmith/fix-with-codesmith-light-3x.png)
Blacksmith's 16-vCPU Linux label is queueing nightly-next's light controls 5-14 min tonight:
pick-nightly-sign-runneris on the path to the feed: signing waits for it.Change: on attempt 1 of a nightly-next run with
vars.CI_NIGHTLY_ONLY_RUNNER == 'glaeda-nightly-only-m4pro', both jobs take the all-nightly reservation first. That is the same guarded first routedecide,resolve-nightly-cmux-tui-clientandfind-nightly-next-recoveryalready use. Otherwise their old route runs.GLAEDA_ROUTE_APP_KEY), whichpick-nightly-build-runneralready mints on these runners (sameci-routing-agentenvironment). No new secret goes to a new runner class.portable-checksentries in.github/runners.jsonmacos_jobs and leaveBLACKSMITH_LINUX_ONLY_DEBT.'blacksmith-4vcpu-ubuntu-2404', which is what itsfromJSON('[...]')[0]evaluated to. The routing-table guard requires a literal Blacksmith Linux fallback.Performance
Tests:
test_nightly_trusted_routing,test_ci_nightly_off_blacksmith,test_runner_routing_table,test_runner_label_policyandtest_ci_linux_single_sizepass.Agent session: agent session
🤖 Generated with Claude Code