The second cut and the move of next: release/fixtures/cut.v0-1-1.json

Merged
DO
8d ago
+668-1

This records the v0.1.1 cut as release/fixtures/cut.v0-1-1.json, in the same form as #617's cut.v0-1-0.json. It was built from release run 42697380276 and the eval run its gate read, 42694638100, as the Factory Builder asked.

What GitHub answered

  • Release run 42697380276 (release.yml, success) cut factory/v0.1.1 at 7ff090d7 and moved factory/v0-next there. It recorded Deployment 7665842974, whose status is success. The fixture holds the run, its one job and the job's log at the step.
  • Eval run 42694638100 (eval.yml, workflow_dispatch, success) ran on 7ff090d7 and ended at 20:52:53Z, before the cut was asked at 20:54:21Z. The eval workflow lists it as its only run on that commit.
  • Read after the cut (20:57Z):
    • factory/v0.1.1 and next both name 7ff090d7.
    • factory/v0.1.0 still names 499835d5, as the first cut left it.
    • next has two Deployments, newest first: v0.1.1's, then v0.1.0's.
    • GitHub marked v0.1.0's Deployment inactive when the new one was made, and kept its earlier success.

The reads went through the Maintainer's login and were trimmed to cut.v0-1-0.json's fields and headers. tools/redact_trial.py redacted them outside the repo, and only the clean output was read. It took out the names of steps[1] and steps[5], the token's mint and its post step, as it did for v0.1.0. removed says what was taken.

Checks it adds

  • tests/test_release.py::test_the_next_cut_moved_next_on holds:
    • the cut's writes;
    • the eval run, by id, workflow, event, conclusion and commit, ended before the cut, and the one run on its commit;
    • both tags at the cut commit, and v0.1.0 where it was;
    • both Deployments and their statuses.
  • Nine mutants of the fixture, each killed by that test: another eval workflow, run or commit; an eval that ended after the cut; another run listed on the commit; v0.1.0 moved; the newest Deployment not v0.1.1's; its status not success; v0.1.0's Deployment still reading only success.
  • release/AGENTS.md's cut row now names both tests: "Each cut moved next and recorded its Deployment; no Release moved or was cut twice". The file is at its 40-line limit, so the row is reworded rather than added.

No record is needed: no rule changes, and no record waited on this fixture.

Checks

CheckResult
make checksOK
tests.test_releaseOK, 28 tests
Mutants, filtered9 of 9 killed
gitleaksNo leaks
Full make verifyNot run locally, by the Maintainer's standing change of 2026-11-09: CI's verify is the answer
Argon (f04f530), 3 partsClean
Muse (f04f530)1 major, fixed: the eval run's id was not held. It now is, with a mutant
Claude review and security (f04f530)Nothing blocking. Lows, all fixed: v0.1.0's Deployment went inactive, now read and held; no mutant on the eval's commit, now added; the AGENTS row had dropped "cut twice", now restored; which eval the gate read rested on a comment, now held by the workflow's run list. The security read found nothing: no credential, only kept headers and fields, no private name
AstraOnly where Muse gives no read

Labelled factory:hands-off. The enqueue waits for the Factory Builder.

šŸ¤– Generated with Claude Code