fix: two rc.6 refresh-cycle robustness gaps found in downstream review - #49
Merged
Conversation
A catalog that renames the confirmation field via entry.verified_key was told to stamp a literal verified: field the sweep does not read, leaving the entry stale and its reminder open. The refresh script now exposes the resolved key as an output and the guidance comment uses it. Found by Codex review on the BCHC rc.6 update PR (bchc#30). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD
MAX_ISSUE_PAYLOAD (200) issue bodies serialized into one env entry can pass Linux's ~128 KiB single-entry cap, and the runner then cannot launch the github-script step at all — no issues opened, refreshed, or closed. The sweep now writes the bodies to $SWEEP_ISSUES_FILE under runner.temp and the workflow reads the file; count, slugs and max_new stay step outputs. Found by Codex review on the BCHC rc.6 update PR (bchc#30). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD
There was a problem hiding this comment.
Sorry @crypticpy, you've used your own review budget of 250,000 diff characters for the last 7 days.
You can request another review in 1 day and 4 hours by commenting @sourcery-ai review. Upgrade to get a review now.
Reviewer's GuideHardens the rc.6 refresh cycle by making maintainer guidance honor custom verification keys and transporting potentially large sweep issue bodies through a shared temporary file instead of a step-output environment entry. Tests pin both workflow contracts and the file-based subprocess behavior. Sequence diagram for file-based verification sweep transportsequenceDiagram
participant Sweep as verification_sweep.mjs
participant Temp as runner.temp file
participant Workflow as GitHub Actions workflow
participant GitHub as GitHub Issues API
Sweep->>Temp: writeFileSync(JSON.stringify(issues))
Sweep-->>Workflow: setOutput(count, slugs, max_new)
Workflow->>Temp: readFileSync(SWEEP_ISSUES_FILE)
Workflow->>GitHub: create or refresh issues using wanted bodies
Workflow->>GitHub: close stale refresh issues
Sequence diagram for custom verification-key guidancesequenceDiagram
participant Refresh as refresh_entry_from_issue.mjs
participant Workflow as refresh-entry workflow
participant GitHub as GitHub Issues API
Refresh-->>Workflow: setOutput(verified_key, verifiedKey)
Workflow->>Workflow: read ENTRY_VERIFIED_KEY
Workflow->>GitHub: createComment with resolved verification key
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
crypticpy
added a commit
that referenced
this pull request
Aug 28, 2026
* docs: record the 2026-08-28 gate evidence and the rc.7 requirement Updates the RR-H rows with the drills, rehearsals and checks completed against the rc.6 candidate: the live updater pass and rollback proof (RR-H02), the end-to-end publish rehearsal and takedown (RR-H03, RR-H07), the re-audited operations inventory sync (RR-H05), the in-flight one-business-day soak (RR-H08), and the deployed-candidate link check with its Discussions finding (RR-H09). Records that the PR #49 merge moves the stable cut behind a v1.9.0-rc.7 candidate. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD * docs: qualify the rollback proof and takedown rehearsal as partial The rc.6 rollback evidence is the offline revert only — the live revert-PR path stays open under RR-H07 — and BCHC PR #32 rehearsed an ordinary unpublish, not the protected-data history purge. Say so. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD * docs: sync the automated updater rows to rc.6 and scope RR-H03 honestly The automated-evidence table still said 'Pass for rc.4' while the new summary cited the rc.6 updater run — record the rc.6 rehearsal (fail- closed contract gate, machine-verified BCHC PR #30, lock at c0826d9) in the exact-rehearsal and protected-content rows. RR-H03 drops back to in progress: only the BCHC leg ran; the PHCT-side rehearsal and the notification confirmation are still open. Found by Codex review on this pull request. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Merged
crypticpy
added a commit
that referenced
this pull request
Aug 28, 2026
* docs: cut the v1.9.0-rc.7 release record Version, changelog section and compare links for the two rc.6 refresh-cycle robustness fixes (PR #49), the upgrade preview pointed at rc.7, the executable plan ledger retargeted from rc.6 to rc.7, and the readiness evidence updated: candidate provenance, the completed PHCT leg of the live publish rehearsal (issue #51 -> PR #52 -> live -> removed in PR #53), and the RR-H03 remainder narrowed to notification delivery. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD * docs: record the rc.7 verify run at its exact commit Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes the two real findings Codex raised on the BCHC rc.6 update PR (bchc#30):
The "something changed" guidance named a literal
verified:field. A catalog that renames its confirmation field viaentry.verified_keywould follow the instruction, stamp a field the sweep never reads, and the entry would stay stale with its reminder open. The refresh script now exposes the resolved key as a step output and the guidance uses it.The sweep copied up to 200 issue bodies into a single environment entry. Linux caps one env entry near 128 KiB; past it the runner cannot launch the github-script step, and the monthly job opens, refreshes, and closes nothing. The bodies now travel through
$SWEEP_ISSUES_FILEunderrunner.temp;count/slugs/max_newremain step outputs.Three new tests pin both contracts (subprocess test for the file channel, workflow-text assertions for both). Full Node suite 799 pass, eslint clean.
Hold: do not merge before the stable
v1.9.0cut. The accepted candidate isv1.9.0-rc.6, and the candidate-to-stable diff must contain no behaviour change (docs/polish-and-publish-plan.md §5) — merging this first would force an rc.7. Neither bug affects the BCHC deployment today (default key, 13 entries). This PR is the head of the next release line.🤖 Generated with Claude Code
https://claude.ai/code/session_014sAVayAT2tkQ7p9U2o9CsD
Summary by Sourcery
Harden refresh-cycle automation against renamed verification fields and oversized issue payloads.
Bug Fixes:
Tests: