Conversation
…estigation + risk-stratified deferral
There was a problem hiding this comment.
Pull request overview
This PR adds a new tick-history shard documenting an investigation into the remaining CI workflows that still invoke deferred hygiene .sh scripts, and records why their .sh→.ts migration is being deferred into smaller follow-up PRs. It fits the repository’s autonomous-loop hygiene trail by preserving a per-tick write record of operational investigation and rollout planning.
Changes:
- Adds a new 2026-05-03 10:51Z tick-history shard under
docs/hygiene-history/ticks/. - Records the three affected workflows and the five deferred shell-script references.
- Documents a proposed low-risk-to-high-risk rollout order for future workflow migrations.
| @@ -0,0 +1 @@ | |||
| | 2026-05-03T10:51:00Z | opus-4-7 / autonomous-loop continuation | a2e2cc3a | **Deferred-5-CI-ref'd-sh investigation: 3 workflows lack setup-bun; flag-compatibility check needed before .ts swap.** Step 0a 12 min approaching threshold — looking for real work continuously per Aaron's tick-15 corrective. Investigation: deferred 5 .sh files (audit-memory-index-duplicates, audit-memory-references, check-archive-header-section33, check-no-conflict-markers, check-tick-history-order) all referenced from 3 CI workflows: gate.yml + memory-index-duplicate-lint.yml + memory-reference-existence-lint.yml. **None have setup-bun step**. Switching .sh → .ts requires (1) add `oven-sh/setup-bun@v2` step before the lint step, (2) verify .ts equivalent supports same CLI flags (e.g., audit-memory-index-duplicates.sh runs with `--enforce` flag — need to verify .ts has same), (3) verify exit code parity, (4) test in CI before removing .sh. **Risk-stratification**: gate.yml is most-load-bearing (3 of 5 .sh refs); a broken edit blocks all PRs. The two memory-*-lint.yml workflows are scoped (path filters: only memory/MEMORY.md changes). Recommended sequence: (a) memory-index-duplicate-lint.yml first (lowest risk; single-tool workflow), (b) memory-reference-existence-lint.yml next, (c) gate.yml last (after first two prove the pattern). Each is its own bounded PR. Per careful-action discipline, none of these are bounded enough for current tick + risk profile too high for end-of-session. **CADENCE-TRACK**: AutoDream still overdue (last 2026-04-23, ~10 days); cadence-rule prohibits same-session run. Defer next-session unchanged. Today's session has produced 50+ PRs; the .sh→.ts conversion is the natural extension that completes Aaron's "fully converted" directive but at workflow-level granularity, not file-level. | All recent PRs (#1366-1375) MERGED; deferred 5 .sh CI-cleanup awaits dedicated bounded PRs (3 workflows, sequential) | This tick teaches **risk-stratified deferral discipline**: when a directive's remaining scope crosses load-bearing boundaries (gate.yml is the all-PR gate), defer to dedicated bounded PRs with sequenced rollout (low-risk → medium → high). Not all "cleanup" is mechanical; some is substantive infrastructure change. Future-Otto: when filing a multi-file cleanup PR, explicitly stratify by reference-type (zero-ref → workflow-ref → load-bearing-workflow-ref) and execute each stratum as separate bounded PR. | | |||
AceHack
added a commit
that referenced
this pull request
May 3, 2026
…st-risk first per #1376 risk-stratification) (#1377) Per Aaron 2026-05-03 "fully converted" directive + B-0156 TS standardization, deferred 5 .sh files in tools/hygiene/ all referenced from CI workflows. #1376 stratified by risk: - LOW: memory-index-duplicate-lint.yml (this PR — single-tool, scoped path filter) - LOW: memory-reference-existence-lint.yml (next) - HIGH: gate.yml (3 of 5 .sh refs; load-bearing all-PR gate; last) This PR (lowest-risk first) tests the pattern: 1. Add `./tools/setup/install.sh` step (provides bun via mise per GOVERNANCE §24, same pattern as gate.yml) 2. Update path filters .sh → .ts 3. Replace `tools/hygiene/audit-memory-index-duplicates.sh --enforce` with `bun tools/hygiene/audit-memory-index-duplicates.ts --enforce` Verified locally: both .sh and .ts produce identical output ("no duplicate memory-index links in memory/MEMORY.md") and exit code (0) for current MEMORY.md state. If this PR's CI passes, apply same pattern to memory-reference-existence-lint.yml next, then gate.yml. Each is its own bounded PR per risk-stratified deferral discipline. The .sh file is NOT removed in this PR — keep both .ts and .sh during the transition; .sh removal happens AFTER all 3 workflows are .ts-based AND verified.
This was referenced May 3, 2026
Merged
AceHack
added a commit
that referenced
this pull request
May 3, 2026
…t tier per #1376) (#1380) The gate.yml workflow ran three lint scripts as .sh: - tools/hygiene/check-tick-history-order.sh - tools/hygiene/check-no-conflict-markers.sh - tools/hygiene/check-archive-header-section33.sh All three have a TS+Bun port in tools/hygiene/<name>.ts with full output / exit-code parity verified locally (179 tick-history rows non-decreasing; 0 conflict-marker violations; 0 §33 archive-header violations). Why this PR closes the .sh→.ts cleanup loop: - #1377 converted memory-index-duplicate-lint.yml (lowest-risk — only fires on memory/MEMORY.md changes) - #1378 converted memory-reference-existence-lint.yml (lowest-risk — only fires on memory/** changes) - This PR converts gate.yml's three lint jobs (highest-risk — fires on every PR via the all-PR gate critical path) - #1379 (rerere ALLOWLIST parity) was a prerequisite — the .ts port of check-no-conflict-markers had to ALLOWLIST the rerere memo to match the .sh version BEFORE running on every PR After this lands and CI verifies green, the 5 deferred .sh files can be removed (audit-memory-index-duplicates.sh, audit-memory-references .sh, check-archive-header-section33.sh, check-no-conflict-markers.sh, check-tick-history-order.sh) — completing the .sh cleanup the maintainer asked for in #1371 ("we should clean up .sh and any .sh where we have the .ts I thought we were fully converted"). Conversion pattern (identical for all three jobs): - Add "Install toolchain via three-way-parity script" step calling ./tools/setup/install.sh — same shape as #1377 + #1378 - Change `run: tools/hygiene/<name>.sh` to `run: bun tools/hygiene/<name>.ts` Three lint jobs run in parallel (one job each), so the bun install cost is bounded by the slowest job's path — does not compound.
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.
Step 0a 12 min approaching threshold. Investigation: 3 workflows (gate.yml + memory-index-duplicate-lint + memory-reference-existence-lint) reference 5 deferred .sh files; none have setup-bun. Switching = add setup-bun step + verify CLI flag/exit-code parity + sequential rollout (low-risk → high). gate.yml is all-PR-gate; defer to dedicated bounded PRs after testing pattern on smaller workflows first.