Repository navigation
Config: no author Role is named until its manifest exists, so no fix Run is dispatched to a missing Role (#675, record 0189) - #717
Conversation
…Run is dispatched to a missing Role (#675) config/defaults.json named the author Role dev-lead, and roles/ holds no dev-lead.json: each fix Run the Kernel decided failed at harness.read with file_unreadable, on every pass up to the Run ceiling, and three such faults trip the automatic Pause. It names none now, so an item needing a fix waits for a person with outcome nothing. tools/check_authority.py refuses an author_role or owner_role that names no file under roles/; a named author Role is the authority table's, as before, and none is allowed. tests/support.py and sim/world.py take the author Role from the authority table, so the fix path and the simulated Factory are held with a Role in place, and a test reads the real configuration. Decision NNNN records it and when it changes back. Pre-approved by the Maintainer 2026-11-08, ruled by the Factory Builder. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ways named, and the record's paths wait (#675, after Muse) Muse: a name with ../ named a file outside roles/ and passed; each is now held to role/1's name. The owner Role may not be none, as the schema and the Kernel already said; only the author may. The record names roles/dev-lead.json as waiting, as a path not yet in the repo is named. Tests and two mutants. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…er Astra) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ber (#675) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the
|
Codex Review Summary: ❌ FailedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
…h its manifest, is accepted, and a test says so (#675, Astra's re-read) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… test says so (#675, Astra's last read) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…t and a mutant say so (#675, Astra's fourth read) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…lder ordered the queue Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|



Closes #675, "author_role is null until roles/dev-lead.json exists, so a fix Run is never dispatched to a missing role". It holds decision record 0189. The protected edits to
config/defaults.jsonandsim/world.pyare pre-approved by the Maintainer 2026-11-08, ruled by the Factory Builder.What changes
config/defaults.jsonauthor_role: null. The Kernel decides no fix Run (kernel/decide.py's_wanted_run), so an item with an open blocking Finding or a failed check has outcomenothingand waits for a person. Before, each fix Run failed atharness.readwithfile_unreadable, on every pass up to the Run ceiling of 10 an item, and three such faults trip the automatic Pausetools/check_authority.pyauthor_roleorowner_rolethat is no Role's name as role/1 spells one (so none reaches outsideroles/), or that has no file underroles/. The owner is always named; the author may be none. A named author Role is the authority table's, as beforetests/support.py,sim/world.pydev-leadexists, and the Kernel's fix path stays held with a Role in place. The live configuration defers each fix to a person until then, which a new test reading the real configuration holdsThe record names when this changes back: the pull request that adds the author Role's manifest, waiting on
roles/dev-lead.json, names it again.Checks
tests/test_decide.py::test_the_configuration_names_no_author_role_until_its_manifest_exists: with the real configuration, an open blocking Finding and a failed check each giveaddressingwith outcomenothingand no admission.tests/test_tools.py::test_a_role_the_configuration_names_has_a_manifest.config/defaults.jsonandtools/check_authority.pyis killed.make checksandcheck_decisionsare clean, apart from the record gap until 0185 to 0187 are onmain. The simulator's and the Kernel's tests pass. CI runs the full verify.Renumbered 0188 → 0189 in 0504b33, as the Factory Builder ordered the queue (the approval-step fix takes 0187, #713 0188).
Reviews
../named a file outsideroles/; 1 minor: the check allowed a null owner Role🤖 Generated with Claude Code