feat: add ontology-backed parallel reclaim planning - #334
Conversation
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Team Run ID: 📒 Files selected for processing (11)
Included review availability: Your plan provides up to 1 included review per hour; 0 remain after this review. 📝 WalkthroughWalkthrough디스크 재생 온톨로지와 파싱 필드가 추가되었습니다. Colima, Podman, Git 클론, 임시 파일, 압축 및 클라우드 eviction의 계획·승인 경로가 확장되었습니다. Tauri 명령, CLI, TypeScript API, 안전 처리와 릴리스 검증이 연결되었습니다. Changes재생 계약과 후보 메타데이터
재생 구현
클라우드와 연결
CLI와 검증
Estimated code review effort: 5 (Critical) | ~120 minutes Merge Risk: 🟠 High · up to This PR adds destructive reclaim planning and execution across several storage providers, but the current head still contains unresolved risks including a permission-check bypass, a race that can delete an unrelated trash destination, incorrect approval/execution behavior, possible hangs, and inaccurate reclaim sizing. These issues can cause data loss, security-boundary weakening, failed cleanups, or misleading plans, so the PR is not ready to merge without fixes or explicit acceptance. Sequence Diagram(s)sequenceDiagram
participant Frontend
participant TauriCommands
participant ReclaimModule
participant FilesystemOrProvider
Frontend->>TauriCommands: inspect 또는 execute 호출
TauriCommands->>ReclaimModule: 경로·프로필·승인 정보 전달
ReclaimModule->>FilesystemOrProvider: 증거 수집 및 상태 확인
ReclaimModule->>FilesystemOrProvider: 승인된 재생 명령 실행
FilesystemOrProvider-->>ReclaimModule: 실행 결과와 사후 상태 반환
ReclaimModule-->>TauriCommands: 계획 또는 실행 결과 반환
TauriCommands-->>Frontend: JSON 결과 반환
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
Full details: Docstring CoverageExplanation Docstring coverage is 57.51% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 386 functions across 62 files. (1 skipped: 1 unsupported.)
✨ Finishing Touches 💡 1📝 Generate docstrings 💡
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Superseded by the current head 39cef1f. The blocking release-runner mismatch from reviewed head e8f67d5 is no longer present: the current release workflow explicitly uses windows-2022, matching verify-release-artifacts.sh's release-disksage-windows-2022-* contract. The OneDrive defensive thread referenced by that review is also resolved on the current PR state. Dismissing this stale changes-requested review does not constitute approval; current-head checks/reviews remain authoritative.
Purpose
Extend DiskSage reclaim planning with ontology-bound evidence and bounded cross-platform adapters while preserving deterministic safety authority. This lane owns the ontology-backed planning/adapter delta above #314; #314 remains the native-temp foundation, #264 remains the release/Test foundation, and #315 owns the canonical product-gap projection.
Current authority — 2026-09-05 KST
main:0e90f9cebadbd7f59606baaec4ca1d2f178c899a;3e33229b4157a084a1985a1c7e952f98a64f25e3;cc9c4f7400813c913cee79c5f317c9581a41acde;ef9f0d3819ed765e5de6742005da09e33479c1e3;.github/workflows/test.ymlremains part of the feat: add ontology-backed parallel reclaim planning #334 semantic delta. Current fix: verify Windows release artifact namespace #264 release/Test contracts are inherited rather than reimplemented, anddocs/product-technical-gap-baseline.mdremains delegated to docs: establish canonical DiskSage product requirements #315;Product and DDD boundary
The lane binds reclaim evidence to DiskSage ontology classes and adds bounded planning for Colima, Podman volumes, OneDrive temporary files, stale Git clones, temporary artifacts, transparent compression, provider-local eviction, and platform process/filesystem evidence. Ontology or model output never becomes mutation authority: deterministic filesystem identity, active-use/materialization evidence, approval/fingerprint checks, and reversible recovery boundaries remain authoritative.
#338 separately owns ordered Test path-filter semantics. This branch owns only the additional platform jobs required by its adapter boundary and must continue to consume #338/#264 contracts rather than silently absorbing them.
Buyer-visible safety gap
Fail-closed planning and rejection tests are necessary but do not by themselves prove buyer-safe successful mutation. Each executable reclaim path requires realistic filesystem/process acceptance binding the same reviewed object through final mutation, rejecting symlink/reparse/hardlink/ancestor replacement races as applicable, producing durable journal/recovery evidence, and demonstrating undo/recovery without unrelated-data loss.
Required before integration
Keep Draft until #264 and #314 are integrated or equivalently present on protected lineage, one unchanged #334 exact head satisfies every applicable native/central Test/Release/Security/SAST/CodeQL/OSV/Scorecard/review/ruleset gate, and destructive/recovery acceptance is real rather than synthetic or rejection-only. No self-approval, force-push, destructive rebase, gate weakening, permanent user-file deletion, or administrative bypass.