[PR #578] [MERGED] fix(artifacts): resolve parents after archive enumeration #2079

New issue

Closed

opened 2026-10-16 21:14:00 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-16 21:14:00 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/578
Author: @kaoru0822-kitauji
Created: 10/5/2026
Status: ✅ Merged
Merged: 10/5/2026
Merged by: @morluto

Base: main ← Head: fix/artifact-parent-entry-order

📝 Commits (1)

25d83af fix(artifacts): resolve parents after archive enumeration

📊 Changes

4 files changed (+268 additions, -6 deletions)

View changed files

📝 src/application/ArtifactInventory/scanReader.ts (+7 -6)
➕ tests/boundary/filesystem/artifactInventoryEntryOrder.test.ts (+136 -0)
➕ tests/boundary/mcp/artifactEntryOrderMcp.test.ts (+67 -0)
➕ tests/fixtures/artifactEntryOrder.ts (+58 -0)

📄 Description

Summary

Resolve artifact containment after reader enumeration so ZIP members can precede their directory entries without corrupting parent relationships or virtual-directory identities.

No tracking issue: found with a valid three-entry ZIP during boundary testing; no matching open issue or PR was found.

Problem and expected behavior

A ZIP ordered pkg/sub/data.txt, pkg/sub/, pkg/ previously attached all three occurrences to the archive root. Both nonempty directories received the empty-directory identity, and inspect-artifact exposed incorrect contains relationships while reporting complete substep coverage.

Archive member order must not determine the nearest represented directory parent. With the fix, the same bytes produce root → pkg → pkg/sub → data.txt.

Change and scope

Resolve occurrence parents against the complete path index after enumeration, before virtual-directory materialization. Per-entry path registration, hashing, safety limits, archive validation, and reader cleanup remain in place.

Missing directory entries are not synthesized. The existing nearest-directory policy and nested-ASAR containment semantics are unchanged. Separate implicit-path identity behavior is outside this fix.

Contract and boundary impact

Semantic owner and earliest changed stage: application artifact inventory, after reader enumeration and before directory identities/edges are constructed.
CLI and MCP/tool-catalog contract: existing shapes and tools; corrected parent IDs and derived containment relationships.
Provider, bridge, target-format, or platform compatibility: real ZIP, filesystem directory, and ASAR fixtures exercised on Linux; no provider protocol changes.
Evidence, artifact, provenance, or reconstruction contract: affected directory identities now incorporate their actual represented children; observed entry paths, content digests, and source locations remain intact.
Process execution, authorization, cleanup, or containment impact: no new execution or authority; path-validation and archive-safety regressions retained.
Generated metadata, package, or installation impact: none; generated-document checks pass.

Evidence and regression coverage

Six permutations of parent/subdirectory/file entries, with an explicit empty-directory control.
Stable directory identities, sorted occurrence paths, containment edges, and repeat-inventory equality.
Implicit-directory cases retain the nearest existing directory or root without synthesizing entries.
Duplicate file/directory paths in both orders, case-fold collisions, prefix conflicts, and traversal rejection.
Filesystem-directory and real-ASAR compatibility controls; MCP boundary regression.
On unchanged main bb91abd, the two new regression files produce 8 failures and 8 passing controls. The focused fixed matrix including existing archive safety and ASAR freshness tests passes 25/25.
Executed the built CLI and a real stdio MCP client against the child-first ZIP: both return the corrected parent chain. Independent review found no blocking issues, including supplemental cancellation, limits, encrypted-member, symlink, and collision probes.

Evidence review:

Observed, derived, and inferred claims remain distinguishable.
Artifact identity, source provenance, and failed attempts remain preserved.
Unsupported, incomplete, unavailable, or uncertain outcomes remain visible.

Validation performed

Rebased and tested on current main bb91abd, including #575 and #576.
npm run check:pr: static checks passed; 1,855 tests passed, 1 skipped, 40 failed in seven native bridge/lease suites because this sandbox denies Unix-domain socket listening (EPERM). The same seven suites on unchanged bb91abd reproduce 40 failures / 22 passes. This is not a full local test-suite pass.
npm run docs:check: passed separately because the full-test failure short-circuits this step in check:pr.
npm run check:ci: passed.
npm_config_cache=/tmp/rea-npm-cache npm run verify:package: passed, including packaged CLI/MCP checks; cache override was needed for this environment.
npm pack --dry-run --cache /tmp/rea-npm-cache: passed.
Real CLI inspect-artifact <child-first.zip> --format json and stdio MCP open_binary + inspect_artifact: passed.
git diff --check: passed.

Compatibility, safety, and release

No schema changes or migration steps. Corrected containment and virtual-directory IDs are intentional for previously misparented entries.
Real Hopper/Ghidra and native Windows/macOS execution were not run locally; no such support claim is made.
Package/release metadata unchanged.
No archive extraction, path-validation relaxation, dependency changes, or unrelated cleanup.

Review checklist

One focused outcome with a Conventional Commit title.
Reason for no linked issue stated.
Regression and meaningful failure-path coverage included.
Existing docs/contracts remain accurate; generated metadata checks pass.
Representative CLI/MCP behavior executed.
Final diff checked for secrets, unrelated changes, and unsupported claims.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/morluto/rea/pull/578 **Author:** [@kaoru0822-kitauji](https://github.com/kaoru0822-kitauji) **Created:** 10/5/2026 **Status:** ✅ Merged **Merged:** 10/5/2026 **Merged by:** [@morluto](https://github.com/morluto) **Base:** `main` ← **Head:** `fix/artifact-parent-entry-order` --- ### 📝 Commits (1) - [`25d83af`](https://github.com/morluto/rea/commit/25d83afc0a606c8d086eac3088de1f9735030128) fix(artifacts): resolve parents after archive enumeration ### 📊 Changes **4 files changed** (+268 additions, -6 deletions) <details> <summary>View changed files</summary> 📝 `src/application/ArtifactInventory/scanReader.ts` (+7 -6) ➕ `tests/boundary/filesystem/artifactInventoryEntryOrder.test.ts` (+136 -0) ➕ `tests/boundary/mcp/artifactEntryOrderMcp.test.ts` (+67 -0) ➕ `tests/fixtures/artifactEntryOrder.ts` (+58 -0) </details> ### 📄 Description ## Summary Resolve artifact containment after reader enumeration so ZIP members can precede their directory entries without corrupting parent relationships or virtual-directory identities. No tracking issue: found with a valid three-entry ZIP during boundary testing; no matching open issue or PR was found. ## Problem and expected behavior A ZIP ordered `pkg/sub/data.txt`, `pkg/sub/`, `pkg/` previously attached all three occurrences to the archive root. Both nonempty directories received the empty-directory identity, and `inspect-artifact` exposed incorrect `contains` relationships while reporting complete substep coverage. Archive member order must not determine the nearest represented directory parent. With the fix, the same bytes produce `root → pkg → pkg/sub → data.txt`. ## Change and scope Resolve occurrence parents against the complete path index after enumeration, before virtual-directory materialization. Per-entry path registration, hashing, safety limits, archive validation, and reader cleanup remain in place. Missing directory entries are not synthesized. The existing nearest-directory policy and nested-ASAR containment semantics are unchanged. Separate implicit-path identity behavior is outside this fix. ## Contract and boundary impact - Semantic owner and earliest changed stage: application artifact inventory, after reader enumeration and before directory identities/edges are constructed. - CLI and MCP/tool-catalog contract: existing shapes and tools; corrected parent IDs and derived containment relationships. - Provider, bridge, target-format, or platform compatibility: real ZIP, filesystem directory, and ASAR fixtures exercised on Linux; no provider protocol changes. - Evidence, artifact, provenance, or reconstruction contract: affected directory identities now incorporate their actual represented children; observed entry paths, content digests, and source locations remain intact. - Process execution, authorization, cleanup, or containment impact: no new execution or authority; path-validation and archive-safety regressions retained. - Generated metadata, package, or installation impact: none; generated-document checks pass. ## Evidence and regression coverage - Six permutations of parent/subdirectory/file entries, with an explicit empty-directory control. - Stable directory identities, sorted occurrence paths, containment edges, and repeat-inventory equality. - Implicit-directory cases retain the nearest existing directory or root without synthesizing entries. - Duplicate file/directory paths in both orders, case-fold collisions, prefix conflicts, and traversal rejection. - Filesystem-directory and real-ASAR compatibility controls; MCP boundary regression. - On unchanged main `bb91abd`, the two new regression files produce **8 failures and 8 passing controls**. The focused fixed matrix including existing archive safety and ASAR freshness tests passes **25/25**. - Executed the built CLI and a real stdio MCP client against the child-first ZIP: both return the corrected parent chain. Independent review found no blocking issues, including supplemental cancellation, limits, encrypted-member, symlink, and collision probes. Evidence review: - [x] Observed, derived, and inferred claims remain distinguishable. - [x] Artifact identity, source provenance, and failed attempts remain preserved. - [x] Unsupported, incomplete, unavailable, or uncertain outcomes remain visible. ## Validation performed - Rebased and tested on current main `bb91abd`, including #575 and #576. - `npm run check:pr`: static checks passed; **1,855 tests passed, 1 skipped, 40 failed in seven native bridge/lease suites** because this sandbox denies Unix-domain socket listening (`EPERM`). The same seven suites on unchanged `bb91abd` reproduce **40 failures / 22 passes**. This is not a full local test-suite pass. - `npm run docs:check`: passed separately because the full-test failure short-circuits this step in `check:pr`. - `npm run check:ci`: passed. - `npm_config_cache=/tmp/rea-npm-cache npm run verify:package`: passed, including packaged CLI/MCP checks; cache override was needed for this environment. - `npm pack --dry-run --cache /tmp/rea-npm-cache`: passed. - Real CLI `inspect-artifact <child-first.zip> --format json` and stdio MCP `open_binary` + `inspect_artifact`: passed. - `git diff --check`: passed. ## Compatibility, safety, and release - No schema changes or migration steps. Corrected containment and virtual-directory IDs are intentional for previously misparented entries. - Real Hopper/Ghidra and native Windows/macOS execution were not run locally; no such support claim is made. - Package/release metadata unchanged. - No archive extraction, path-validation relaxation, dependency changes, or unrelated cleanup. ## Review checklist - [x] One focused outcome with a Conventional Commit title. - [x] Reason for no linked issue stated. - [x] Regression and meaningful failure-path coverage included. - [x] Existing docs/contracts remain accurate; generated metadata checks pass. - [x] Representative CLI/MCP behavior executed. - [x] Final diff checked for secrets, unrelated changes, and unsupported claims. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-16 21:14:00 +00:00