main
10 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
| f3ef4d917c |
feat(#64): IMP-35 details_popup_escalation u1~u10 + Stage 3 R7 anchor re-pin
Land the production + test surface for the Step 17 cascade POPUP terminal (DETERMINISTIC -> POPUP -> AI_REPAIR -> USER_OVERRIDE) per Stage 2 plan R2. u11 (baseline-red invariance gate) was already landed in |
|||
| c1df656312 |
feat(#65): IMP-36 fit/rotation generalization (u1~u8)
Generalize Phase Z frame partial responsive fit / rotation to four canonical
F13/F14/F20/F8 family partials. Surface = 13 canonical partials; 19
builder-only contracts remain explicitly out of scope.
u1 test_imp17_comment_anchor: re-pin L570->L578 (restructure+IMP-17),
L571->L579 (IMP-29 -> IMP-47B supersession). Stage 1 red baseline gate.
u2 frame_contracts.yaml: add rotation_eligible (P1) + body_fit_pattern2 (P2)
bool axes on 13 partial-backed contracts. P1 True: F13/F14/F20/F8 (4).
P2 True: F23 + P1_set (5). F29 columns[1].body_parser column_plain ->
column_with_transform (P3 parity).
u3 test_imp36_fit_rotation_generalization (NEW, 166 lines): static
parametrized assertions for P1 metadata + CQ presence, P1 opt-out
absence, P2 --max-body-lines + clamp + cqh, P2 opt-out absence, 19
builder-only exclusion.
u4 three_parallel_requirements (F13): introduce f13b-root container-name +
container-type:size + @container (aspect-ratio<1.5) rotation;
add inline --max-body-lines + body line-height clamp/cqh/calc.
u5 three_persona_benefits (F14): f14b-root P1 + P2 cqh/jinja body fit.
Persona colors (#285b4a/#445a2f/#743002) and circle SVG aspect 1/1
preserved.
u6 dx_sw_necessity_three_perspectives (F20): f20b-root P1 + P2 cqh/jinja
body fit under IMP-49 partial-fidelity lock.
u7 info_management_what_how_when (F8): f8b-root P1 + P2 cqh/jinja body fit.
u8 test_imp36_overflow_chain_self_fire (NEW, 299 lines): Selenium self-fire
harness for F13/F14/F20/F8 at aspect 1.78 vs 1.0. Asserts line-height
changes, font-size invariance across all 4 frames (no per-frame exempt),
grid columns rotate 3 -> 1, OVERFLOW_CASCADE_ORDER remains 4-tuple.
Stage 4 verification (HEAD
|
|||
| 1efbf672bd |
feat(#39): IMP-30 first-render invariant + abort bypass (2 paths)
Restore first-render invariant: final.html + Step 20 slide_status MUST be written for every input where Step 0~5 succeed. Two abort paths replaced with provisional/empty-shell synthesis; MDX content preserved, AI-free. - u1 V4Match.provisional + lookup_v4_match_with_fallback(allow_provisional) chain_exhausted -> synthesize rank-1 provisional (opt-in, default-off) - u2 CompositionUnit.provisional propagation (single / parent_merged / parent_merged_inferred constructors) - u3 select_composition_units(allow_provisional_fill=True) last-resort fill + _candidate_state="selected_provisional" - u4 pipeline.py path-(a) abort guard replaced with provisional retry + terminal __empty__ shell (no sys.exit(1)) - u5 zones_data.provisional -> slide_base.html zone--provisional class + data-provisional + needs-adaptation badge (template-only) - u6 compute_slide_status additive provisional_first_render_count/_units (overall enum unchanged per IMP-05 Codex #10 D4) - u7 regression: tests/test_phase_z2_imp30_first_render.py (28 tests) + tests/test_phase_z2_v4_fallback.py (+5 cases) Guardrails verified: MVP1_ALLOWED_STATUSES unchanged, no calculate_fit, no LLM in fallback path, no MDX 03/04/05 hardcoding. Anchor sync (Rule 13): tests/orchestrator_unit/test_imp17_comment_anchor.py re-pinned 564/565 -> 570/571 to track V4Match.provisional shift at src/phase_z2_pipeline.py:179-184. Cross-ref: IMP-05 (#5) §5 defer + Codex #2 first-render invariant. |
|||
| 5d23b747ff |
fix(orchestrator): P5b first-line agent header strict + supplement throttle
Bug discovered during #24 IMP-24 K6 Stage 2 (2026-05-20): - Codex r1, r2, r3 started with '=== IMPLEMENTATION_UNITS ===' on first line (not '[Codex #N] ...'), so detect_agent (P0-1 strict, first-line only) returned None. - For non-audit issues, the P5 supplement guard was audit-only gated → silent loop until Codex r4 happened to use correct format. 4 rounds wasted. Verified that #21 Stage 4 had the same latent silent loop pattern ('## [Codex #1]' first line) — orchestrator looped through ~10 Claude rounds before random recovery. P5b fix addresses this long-standing bug. Patch (defensive parser-contract hardening; does not assume single root cause): 1. RULES global gets explicit "FIRST non-empty line MUST be [Claude #N] / [Codex #N]" rule that OVERRIDES any stage-specific "body MUST contain" constraint. 2. COMPACT_PLAN_RULE wording clarified: "body" begins AFTER the first-line agent header. The 'body MUST contain ONLY' set no longer accidentally permits '=== IMPLEMENTATION_UNITS ===' on line 1. 3. is_codex None supplement guard: - audit-only gate REMOVED → fires for all issues (#24 latent loop fixed) - Throttle: max 2 supplements per stage; on 3rd violation, orchestrator hard-stops the issue with explicit "user action required" message and exits run_stage cleanly - Supplement message names both Claude AND Codex (Claude's first-line violation also breaks downstream via Codex mimicry) - Body-head 80 chars logged on detection failure (debugging aid) 4. Regression tests (+5 cases in test_orchestrator_core.py): - TestDetectAgent: '=== IMPLEMENTATION_UNITS ===' first line → None - TestDetectAgent: [Codex #N] first line + units after → 'codex' OK - TestDetectAgent: '## ', '📌 **', '**' prefix all → None - TestRulesAndCompactPlanFirstLineContract: RULES wording has FIRST/OVERRIDES - TestRulesAndCompactPlanFirstLineContract: COMPACT_PLAN_RULE has carve-out Cosmetic side effect (accepted): Claude's '📌 **[Claude #N] ...**' or '## [Codex #N] ...' decoration prefixes will fail detect_agent. Agents will drop decorations from line 1; line 2+ can still use them. Out of scope (NOT included to keep regression risk low): - detect_agent function logic UNCHANGED (P0-1 strict preserved) - consensus parser UNCHANGED - stage loop structure UNCHANGED - git/Gitea retrieval logic UNCHANGED - audit-only mode P4/P4a guards UNCHANGED - pre-post comment validation (future axis, larger refactor) Total: 131/131 pytest pass (126 prior + 5 new). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
|||
| 134f52d3d3 |
feat(#58): L3 dormant trigger guard -- DORMANT-TRIGGERS.yaml + checker + orchestrator hook
P5-1 docs/architecture/DORMANT-TRIGGERS.yaml -- 5 entries (IMP-16/17/18/19 active + IMP-20 followup-linked #55). P5-2 scripts/check_dormant_triggers.py -- standalone, reads registry, scans tree + diff, writes .orchestrator/dormant_alerts.json, exit 0 always. P5-3 orchestrator.py -- _check_dormant_triggers() helper + Stage 4->5 informational alert branch (skips audit-only, never blocks). P5-4 tests/orchestrator_unit/test_dormant_triggers.py -- 30 cases (yaml schema, registry contents, checker matching, false-positive guards, manual-evidence skip, orchestrator branch, audit bypass, governance ref). P5-5 PROJECT-INTENT-AND-GOVERNANCE.md -- single anti-patterns row referencing the L3 registry as binding contract surface. Tests: pytest -q tests = 337 passed (baseline 307 + 30 new). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
|||
| 9389b8425b |
fix(orchestrator): P5 audit-anchor-first-line regression guard
Bug discovered during #56 INTEGRATION-AUDIT-02 execution (2026-05-20): - Both Claude and Codex put "Audit anchor: ..." as the FIRST line of every Gitea comment per the #56 issue body instruction "cite anchor at start of every stage". - detect_agent (P0-1 strict, first-line only) then returns None for these comments because the first line is "Audit anchor:..." not "[Codex #N]" or "[Claude #N]". - Result: orchestrator's "is_codex" check (line ~1288) flips false → "Codex 응답 미감지 — continuing" → infinite Stage 4 loop. #56 reached Round #14 (>300 comments, ~2 hours wasted token). Fix path (NOT relaxing detect_agent — that would revive the original #45 pre-P0-1 bug where [Claude #N] citations inside Codex bodies caused mis-detection): 1. AUDIT_ONLY_NOTE updated to enforce comment format: - FIRST non-empty line MUST be `[Claude #N] <stage>` or `[Codex #N] <stage>` - Audit anchor / banners / prefaces MUST appear line 2 or later - Concrete CORRECT example included - Explicit warning that violation breaks stage advance 2. is_codex None guard auto-supplements: - When _audit_mode(title) AND detect_agent returns None, orchestrator posts a Gitea supplement comment requesting the correct format - Next round's Claude/Codex see the supplement and correct - Breaks the infinite loop automatically (no manual ctrl-C needed) 3. Regression tests in TestDetectAgent (test_orchestrator_core.py): - test_audit_anchor_preface_breaks_detection: confirms P0-1 strict correctly returns None when anchor is first line - test_audit_anchor_after_header_works: correct format passes Total: 96/96 pytest pass (94 prior + 2 P5 regression). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
|||
| e32f632464 |
fix(orchestrator): P4a baseline-diff guard + Stage 5 commit scope
P4 had two production issues blocking #50 integration audit deployment: 1. Stage 3 guard had no baseline awareness — flagged ALL forbidden-path changes including pre-existing dirty WIP. Empirical: 328 such files already in current working tree (tests/matching/ artifacts etc). #50 would have hit reject loops immediately without Claude doing anything wrong. 2. Stage 5 had no commit-scope guard — if Claude ran `git add -A` and committed user's existing WIP, audit commit would be polluted with unrelated production changes. P4a additions: - _audit_baseline_path / _ensure_audit_baseline / _load_audit_baseline: snapshot working-tree dirty paths at run_issue entry for audit issues. Resumed runs preserve existing baseline (no overwrite). - _check_audit_only_violations(baseline=None): accept baseline set, subtract from violations — only flags NEW forbidden changes introduced after audit start. - _check_audit_commit_scope: verify HEAD commit's file list matches AUDIT_ALLOWED_COMMIT_GLOBS (INTEGRATION-AUDIT-*.md, BACKLOG.md). - run_issue: save baseline on audit-mode entry only — no impact on normal issues. - Stage 5 (commit-push) YES gate: new guard rejects on out-of-scope files with remediation prompt (git reset --soft + force-with-lease). 19 new tests: - baseline subtraction (5): pre-existing removed, None=keep-all, empty-set=catch-all, full-coverage filter, Windows path normalize. - baseline persist (5): roundtrip, no-overwrite on resume, missing fallback, corrupt JSON fallback, non-list fallback. - commit scope detection (7): report-only allowed, backlog allowed, src/ rejected, unrelated docs rejected, git error fail-open, Windows backslash, empty commit pass. - allowed globs sanity (2): every glob has audit marker, all under docs/architecture/. Total: 94/94 pytest pass (75 prior + 19 new). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
|||
| 4289a500b6 |
feat(orchestrator): P3 wrapper input/encoding fix + P4 audit-only mode
P3 hotfix (2026-05-18 — verified during #46 retry attempt): - _run_with_tree_kill: encode input only when Popen is in binary mode. Previously force-encoded str→bytes even with encoding= set, breaking text-mode stdin pipes with: write() argument must be str, not bytes. - run_claude path was the only affected call site. - 3 new C7 regression tests (input+encoding / bytes+binary / auto-encode). - C3/C6 test fixtures hardened with DEVNULL stdio isolation. P4 audit-only mode (2026-05-19, prep for #50 integration audit): - _is_audit_issue: title-based detection for [INTEGRATION-AUDIT*], [AUDIT-ONLY], or "integration audit" phrase. - _audit_mode + --audit-only CLI flag: manual override regardless of title. - AUDIT_ONLY_NOTE injected into context pack across all stages/rounds. - Stage 3 (code-edit) YES gate: deterministic git status check. Changes touching src/**, templates/**, tests/** auto-reject Stage 3 YES and post a supplement-request comment. LLM-independent enforcement. - 26 new audit-mode tests (title detection, CLI override, forbidden prefix detection, allowed paths pass, Windows backslash normalization, quoted paths with spaces, git error fail-open, constants sanity). Total: 75/75 pytest pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
|||
| e10ec36617 |
feat(IMP-17): AI repair fallback infra carve-out — design-only boundary + 3-cond AND gate
u1 — src/phase_z2_pipeline.py:564 route hint comment corrected from non-existent IMP-31 to IMP-17 (carve-out, AI fallback only, normal path 밖). Line 565 IMP-29 frontend override reference untouched. u2 — docs/architecture/IMP-17-CARVE-OUT.md (new) defines: - allowed scope (Step 12 restructure proposal, Step 16/17 retry fallback) - forbidden scope (normal-path AI calls, MDX compression, HTML structure) - 3-condition AND activation gate (User GO ∧ B4 frame_selection evidence ∧ IMP-04 catalog + IMP-05 V4 fallback live) - pattern shape reference (link-only): content_editor.py:21,318 + sse_utils.py:16-50 (Phase Q Archive Candidate, no port) - AI 격리 contract + Kei persona 단절 (permanent) u3 — PHASE-Z-IMPLEMENTATION-ISSUE-BACKLOG.md:68 IMP-17 row gains carve-out doc link + 3-cond AND gate pointer. u4 — PHASE-Q-INSIGHT-TO-22STEP-MAP.md AI repair fallback infra registry row prefixed with IMP-17 + carve-out link; normal_path=no preserved. Anchor test: tests/orchestrator_unit/test_imp17_comment_anchor.py asserts line 564 IMP-17 wording AND line 565 IMP-29 preservation (2 tests pass). Runtime behavior change: 0. Only delta in executable file is one comment line. Normal-path AI invocation count remains 0. Refs: gitea #17 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
|||
| f3bff898fb |
feat(orchestrator): initial orchestrator + subprocess cleanup hardening
Pre-existing P0+P1 fixes (verified via #45 pilot 2026-05-18): - P0-1: detect_agent first-line only (fixes #45 infinite loop) - P0-2: stage_start_count sanity reset on external comment delete - P0-3: 32 pytest cases for parse/detect regressions - P1-4: execution-issue mode prompt (compact scope-tight) - P1-5: Stage 2 COMPACT_PLAN_RULE (size budget, no code snippets) - P1-6: tests:[] orchestrator-level enforcement at Stage 2 YES guard - P1-7: dual-write CRLF/trailing-whitespace normalize P3 subprocess cleanup (PID 2780 orphan grandchild regression): - (pid, create_time) signature tracking — Windows PID reuse safe - _kill_process_tree: parent-alive traversal path - _kill_tracked: parent-dead orphan path - _run_with_tree_kill: 1s monitor thread captures descendants live - atexit + SIGINT safety net via _SPAWNED set - 4 subprocess.run sites switched to wrapper (compaction/exit_report/ run_claude/run_codex) - 12 cleanup pytest cases incl. C6 PID 2780 regression test Selenium boundary unchanged — driver.quit() in phase_z2_pipeline.py and slide_measurer.py already protected by try/finally. Total: 44/44 pytest pass (32 core + 12 cleanup). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
