Skip to content

Commit 4ef6c7f

Browse files
committed
chore(scaffolding): 2026-07-02 inventory quick wins
- scripts/migrate_debt_to_yaml.py DELETED (one-shot ADR-0072 migration; source .dev/debt.md long gone, zero inbound refs). - bench/record.sh DELETED (depended on the retired quick.sh / quick_baseline.txt — could only die at startup); bench/history.yaml header rewritten to the on-demand reality. - bench/benchmarks 26_json_parse -> 39_json_parse (number collision with 26_gc_alloc_rate). - test/run_all.sh comment rot: pre-ADR-0107 "Mac per-commit" usage line, stale ~248-step counts x2, retired "Phase 5+ they become blocking", dead "Future suites" block (test_clj already wired as test_clj_tier_a; diff_runner/tier_check retired with the phase queue, ADR-0142). - gate_continue_remind.sh matcher narrowed to actual gate INVOCATIONS (was bare *run_all.sh* substring — 7 false injections in one session). - SessionStart perf-campaign essence injection now gated on .dev/.perf_campaign_active (same pause switch perf_campaign_remind.sh honours — a PAUSED campaign no longer costs every session's context). Smell-audited: 1: audit-evidenced deletions/corrections only; no behavior change outside hooks
1 parent e1db60a commit 4ef6c7f

13 files changed

Lines changed: 21 additions & 310 deletions

File tree

‎.claude/settings.json‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -63,7 +63,7 @@
6363
"hooks": [
6464
{
6565
"type": "command",
66-
"command": "printf '=== language policy ===\\nReply to the user in Japanese (画面表示は日本語). Code, identifiers, file paths, commit messages, and English docs stay in English. See .claude/output_styles/japanese.md.\\n\\n' && printf '=== gate cadence (dev efficiency — read BEFORE the first commit) ===\\nDo NOT run the full gate per commit (it is slow and kills iteration). Per commit: bash test/run_all.sh --smoke <changed-e2e-step> (~tens of sec; ADR-0107 — the smoke tier authorizes shared-code commits too, up to 5 before a forced full gate) and do not block. Batch the full bash test/run_all.sh --serial-e2e at the ceiling / phase boundary / pre-release (backgroundable as a look-ahead). If a full gate fails on ONLY a non-code check (e.g. check_debt_id_refs), fix it + re-smoke — do NOT re-run the whole full gate. SSOT: .claude/rules/gate_cadence.md + memory smoke_first_batch_full_gate.\\n\\n' && test -f $CLAUDE_PROJECT_DIR/.dev/handover.md && printf '=== .dev/handover.md ===\\n' && cat $CLAUDE_PROJECT_DIR/.dev/handover.md && printf '\\n=== git log -3 ===\\n' && git -C $CLAUDE_PROJECT_DIR log -3 --decorate --oneline; test -f $CLAUDE_PROJECT_DIR/.dev/perf_campaign_essence.md && printf '\\n=== perf-campaign essence (§9.2.S — relentless lookahead, machine-injected) ===\\n' && cat $CLAUDE_PROJECT_DIR/.dev/perf_campaign_essence.md || true"
66+
"command": "printf '=== language policy ===\\nReply to the user in Japanese (画面表示は日本語). Code, identifiers, file paths, commit messages, and English docs stay in English. See .claude/output_styles/japanese.md.\\n\\n' && printf '=== gate cadence (dev efficiency — read BEFORE the first commit) ===\\nDo NOT run the full gate per commit (it is slow and kills iteration). Per commit: bash test/run_all.sh --smoke <changed-e2e-step> (~tens of sec; ADR-0107 — the smoke tier authorizes shared-code commits too, up to 5 before a forced full gate) and do not block. Batch the full bash test/run_all.sh --serial-e2e at the ceiling / phase boundary / pre-release (backgroundable as a look-ahead). If a full gate fails on ONLY a non-code check (e.g. check_debt_id_refs), fix it + re-smoke — do NOT re-run the whole full gate. SSOT: .claude/rules/gate_cadence.md + memory smoke_first_batch_full_gate.\\n\\n' && test -f $CLAUDE_PROJECT_DIR/.dev/handover.md && printf '=== .dev/handover.md ===\\n' && cat $CLAUDE_PROJECT_DIR/.dev/handover.md && printf '\\n=== git log -3 ===\\n' && git -C $CLAUDE_PROJECT_DIR log -3 --decorate --oneline; test -f $CLAUDE_PROJECT_DIR/.dev/.perf_campaign_active && test -f $CLAUDE_PROJECT_DIR/.dev/perf_campaign_essence.md && printf '\\n=== perf-campaign essence (§9.2.S — relentless lookahead, machine-injected) ===\\n' && cat $CLAUDE_PROJECT_DIR/.dev/perf_campaign_essence.md || true"
6767
}
6868
]
6969
}
File renamed without changes.

‎bench/history.yaml‎

Lines changed: 7 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -4,19 +4,18 @@
44
# (originally drafted at slot 0027 at row 8.2; renumbered to 0044
55
# at the Phase 8 → 9 audit to resolve a collision with the NaN-box
66
# ADR).
7-
# Row 8.3's 1.2x regression gate consumes only `lock: true` entries
8-
# matching (machine.id, backend); CI's `bench/quick.sh` appends raw
9-
# samples to `quick_baseline.txt`, NOT here.
10-
#
11-
# Curated via `bench/record.sh --id=<id> --reason=<reason>`
12-
# (helper lands at row 8.3 cycle 1). Manual edits OK while the
13-
# helper is pending — keep entries sorted by date.
7+
# Historical lock-points. The quick.sh auto-baseline + the record.sh
8+
# helper were RETIRED (2026-06-11 gate change; record.sh removed
9+
# 2026-07-02 — it depended on the retired quick.sh/quick_baseline.txt
10+
# and could no longer run). Perf is measured on demand via
11+
# bench/compare_langs.sh / bench/run_bench.sh; append entries manually
12+
# if a new lock-point is worth recording — keep entries sorted by date.
1413

1514
schema_version: 1
1615
env:
1716
cpu_arch: aarch64 # primary dev host narrow info
1817
os: Darwin
19-
tool: bash + EPOCHREALTIME # quick.sh micro-timer
18+
tool: bash + EPOCHREALTIME # micro-timer (was quick.sh's)
2019
entries:
2120
- id: "8A.0"
2221
date: "2026-05-27"

‎bench/record.sh‎

Lines changed: 0 additions & 111 deletions
This file was deleted.

‎scripts/check_debt_id_refs.sh‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -44,7 +44,7 @@ search_paths=(src .dev .claude scripts test data/feature_deps.yaml data/placemen
4444
# ID in immutable history to resolve to a live row fights that immutability,
4545
# and ADRs do not drive the Step-0.5 trigger system (debt.yaml rows do). The
4646
# check guards LIVE trigger sites (src + live docs), not history.
47-
exclude='\.dev/debt\.ya?ml|\.dev/decisions/|tech_debt_consolidation\.md|audit-lens|check_debt_id_refs\.sh|migrate_debt_to_yaml'
47+
exclude='\.dev/debt\.ya?ml|\.dev/decisions/|tech_debt_consolidation\.md|audit-lens|check_debt_id_refs\.sh'
4848

4949
# 1. Phantom placeholder IDs — never a real row.
5050
phantom=$(rg -n --no-heading 'D-NEW[A-Z0-9-]*' "${search_paths[@]}" 2>/dev/null \

0 commit comments

Comments
 (0)