morluto/rea

Fork 0

mirror of https://github.com/morluto/rea.git synced 2026-11-05 06:45:22 +00:00

[PR #5788] [MERGED] fix(native): admit thin arm64_32 Mach-O targets for static inspection #7061

New issue

Closed

opened 2026-11-01 21:45:04 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-11-01 21:45:04 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/5788
Author: @Lew1sWong
Created: 10/21/2026
Status: ✅ Merged
Merged: 10/22/2026
Merged by: @morluto

Base: main ← Head: fix/thin-arm64-32-macho-admission

📝 Commits (2)

a73ae99 fix(native): admit thin arm64_32 Mach-O targets for static inspection
b011081 ci(apple): run the thin arm64_32 CLI regression on macOS

📊 Changes

15 files changed (+213 additions, -20 deletions)

View changed files

📝 .github/workflows/ci.yml (+1 -0)
📝 scripts/ci/scopes.mjs (+1 -0)
📝 src/domain/binaryTarget.test.ts (+1 -0)
📝 src/domain/binaryTarget.ts (+19 -1)
📝 src/domain/binaryTargetTypes.ts (+11 -1)
📝 src/domain/elfFixedHeader.ts (+8 -5)
📝 src/ghidra/GhidraProvider.behavior.test.ts (+19 -1)
📝 src/hopper/HopperAnalysisProfile.ts (+19 -3)
📝 src/hopper/HopperProvider.test.ts (+19 -0)
📝 src/hopper/HopperProvider.ts (+3 -8)
📝 src/ida/IdaProvider.test.ts (+19 -0)
📝 src/ida/IdaProvider.ts (+7 -0)
📝 src/native/MachoArchitectureSelection.ts (+4 -0)
➕ tests/boundary/cli/nativeMachoArm64_32Cli.test.ts (+77 -0)
📝 tests/fixtures/nativeCommands.ts (+5 -1)

📄 Description

Summary

Fixes #5713. A thin arm64_32 Mach-O is now admitted as a target, so list-architectures, inspect-macho and inspect-signature inspect it with Apple's native tools instead of refusing it at target admission.

Problem and expected behavior

parseThinMachO mapped Mach-O CPU types only to x86, x86_64, arm and arm64. CPU_TYPE_ARM64_32 (0x0200000c) fell through to err("unsupported Mach-O architecture"), so all three commands failed with target_unavailable / unsupported_target before any tool ran. The FAT control already showed the inconsistency: the same slice inside a universal file was inventoried by lipo.

Expected, per the list_architectures contract and the issue: the thin object returns ordinary thin-file Evidence. Thin rows keep the existing convention of null CPU code, offset, size and alignment.

Change and scope

Domain. arm64_32 joins BINARY_ARCHITECTURES as a Mach-O-only architecture. This follows the Hexagon precedent from #4819: it is excluded from HeaderArchitecture and gets its own thin mach-o variants in ExecutableMetadata and BinaryTarget. As a result, the ELF, PE, raw-binary and universal paths cannot carry it, and the ELF fixed-header type now says so explicitly.
Native slice policy. MACHO_ARCHITECTURE_POLICIES gains arm64_32. Without it, the newly admitted target reached the 32-bit dispatch-metadata reader and failed as execution_failure: "Mach-O CPU type does not match requested architecture arm64_32". With it, that reader raises its existing unsupported-32-bit error, like other 32-bit Mach-O images.
Deep-analysis providers. Ghidra, Hopper and IDA have no verified arm64_32 lane, so they decline the target with architecture_unsupported:
- Ghidra, through its existing supported-architecture set;
- Hopper, through one isHopperDeclinedArchitecture predicate that now replaces the two duplicated checks;
- IDA, through an explicit refusal.
Kuna already declines it through its verified-architecture set.

Not included:

FAT files. Universal availableArchitectures still lists only the four provider-neutral families, so arm64_32 slices in FAT files remain visible only through lipo inventory, as before.
Duplicated CPU-type maps. These existed before this PR. This PR updates the two that arm64_32 targets reach: machArchitecture admission and the native slice policy. It leaves two unchanged:
- machoTargetSelectionMatchesArchitecture, which only deep-analysis profiles reach;
- Hopper's slice CPU map.
Consolidating them is a separate refactor.

Contribution scope: ordinary fix within existing boundaries; maintainer scope decision not applicable.

Contract and boundary impact

Semantic owner and earliest changed stage: target admission, thin Mach-O header parsing in src/domain/binaryTarget.ts.
Contribution scope fit and linked maintainer decision: not applicable (bug fix).
CLI and MCP/tool-catalog contract: additive. Schemas that enumerate BINARY_ARCHITECTURES gain arm64_32, for example Evidence subject.architecture. The value appears only for targets that were previously rejected.
Provider, bridge, target-format, or platform compatibility: native macOS tools inspect thin arm64_32 targets. Ghidra, Hopper, IDA and Kuna decline them.
Evidence, artifact, provenance, or reconstruction contract: thin arm64_32 Evidence carries format: "mach-o" and architecture: "arm64_32" with native-tool provenance.
Target selection, authority, upstream behavior, and cleanup details: thin targets stay unfiltered, so no -arch is passed. FAT selection is unchanged.
Generated metadata, package, or installation impact: none committed. npm run docs:check passes.
Read existing files or captured data.

Evidence and regression coverage

Tests added or updated:
- tests/boundary/cli/nativeMachoArm64_32Cli.test.ts (macOS only): compiles a header-free object with xcrun clang -arch arm64_32 -c and checks the public CLI. It is listed in the macOS apple-artifacts job and in SCOPES.apple, so CI runs it.
  - All three commands exit 0 with subject.architecture: "arm64_32".
  - list-architectures lists exactly arm64_32.
  - inspect-native-dispatch-metadata returns the typed refusal capability_unavailable / target_unsupported, retryable: false.
- Parser row in binaryTarget.test.ts, plus refusal tests for Ghidra, Hopper and IDA.
- tests/fixtures/nativeCommands.ts: the arm64 fixture helper's return type is narrowed to what it builds.
Base reproduction:
- The issue's nine-command probe on 7d8e4e433 (macOS arm64, Apple clang): all three thin arm64_32 commands exit 1 with unsupported Mach-O architecture; the arm64 and FAT controls exit 0. After the change, all nine exit 0.
- The new CLI test fails on the unmodified source with that same target_unavailable error.
- With only the slice policy removed, it fails on the execution_failure CPU-mismatch error described above.

User-visible CLI output (list-architectures arm64_32.o --json, abridged):

"subject": { "format": "mach-o", "architecture": "arm64_32" },
"normalized_result": { "architectures": { "items": [
  { "name": "arm64_32", "cpu_type": null, "cpu_subtype": null,
    "file_offset": null, "size": null, "alignment": null } ],
  "total": 1, "exhaustive": true, "limitations": [] } }

Remaining proof gaps:
- No linked watchOS executable was tested; only a compiler-produced object was used.
- Provider refusals are verified with unit tests only; no real Ghidra, Hopper or IDA run.
- MCP was not exercised directly. From the source, it uses the same target admission as the CLI.

For evidence-bearing changes:

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

Node 24.18.0, macOS 27.0 arm64, Apple CommandLineTools clang, lipo, otool and codesign. Head b0110810a on 7501911f1.

npm run check:fast — passed (typecheck, lint 0 errors)
npm run knip — passed
npm run docs:check — passed
npm run test:focused -- on every test under src/domain, src/hopper, src/ida, src/ghidra, src/native and tests/boundary/providers/native, src/application/binaryDiagnostics, both native Mach-O CLI tests, and the four CI planning tests — 454 files, 9072 tests passed, 2 skipped
scripts/ci/plan.mjs, simulated as a pull_request from 7501911f1 to this head — apple=true
Full npm run check:pr — not run

Compatibility, safety, and release

Breaking changes or migration steps: none. The new enum value appears only for targets that were rejected before.
Real Hopper/Ghidra, browser, or OS coverage: real macOS Apple tools only; no deep-analysis provider installed.
Package or release metadata impact: none.
Security, privacy, process, or containment review: static file inspection only. No target execution.

Author readiness

Human owner: @Lew1sWong

I reviewed the final diff and relevant existing code, understand the change and its edge cases, and take responsibility for the complete PR, including AI-assisted work.
I checked the evidence above, confirmed it supports the current change, and recorded meaningful validation gaps.
I declare this PR ready for initial maintainer review.

Before merge

This PR addresses a concrete problem or an agreed enhancement.
Material scope and authority questions are resolved, and any required maintainer scope decision covers the final diff.
The PR has one focused outcome and the title follows type(scope): outcome.
Related issue is linked, or the reason for not linking one is stated above.
Tests cover changed observable behavior and meaningful failure paths.
Owning docs, contracts, and generated metadata are updated where needed.
User-visible CLI/MCP changes include representative output.
I checked the final diff for secrets, unrelated cleanup, and unsupported claims.

🤖 Generated with Claude Code

_{🔄 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/5788 **Author:** [@Lew1sWong](https://github.com/Lew1sWong) **Created:** 10/21/2026 **Status:** ✅ Merged **Merged:** 10/22/2026 **Merged by:** [@morluto](https://github.com/morluto) **Base:** `main` ← **Head:** `fix/thin-arm64-32-macho-admission` --- ### 📝 Commits (2) - [`a73ae99`](https://github.com/morluto/rea/commit/a73ae99dee289aadc64ddbe1215e439868068db5) fix(native): admit thin arm64_32 Mach-O targets for static inspection - [`b011081`](https://github.com/morluto/rea/commit/b0110810ae4f775a31fc3d873d8c593819379162) ci(apple): run the thin arm64_32 CLI regression on macOS ### 📊 Changes **15 files changed** (+213 additions, -20 deletions) <details> <summary>View changed files</summary> 📝 `.github/workflows/ci.yml` (+1 -0) 📝 `scripts/ci/scopes.mjs` (+1 -0) 📝 `src/domain/binaryTarget.test.ts` (+1 -0) 📝 `src/domain/binaryTarget.ts` (+19 -1) 📝 `src/domain/binaryTargetTypes.ts` (+11 -1) 📝 `src/domain/elfFixedHeader.ts` (+8 -5) 📝 `src/ghidra/GhidraProvider.behavior.test.ts` (+19 -1) 📝 `src/hopper/HopperAnalysisProfile.ts` (+19 -3) 📝 `src/hopper/HopperProvider.test.ts` (+19 -0) 📝 `src/hopper/HopperProvider.ts` (+3 -8) 📝 `src/ida/IdaProvider.test.ts` (+19 -0) 📝 `src/ida/IdaProvider.ts` (+7 -0) 📝 `src/native/MachoArchitectureSelection.ts` (+4 -0) ➕ `tests/boundary/cli/nativeMachoArm64_32Cli.test.ts` (+77 -0) 📝 `tests/fixtures/nativeCommands.ts` (+5 -1) </details> ### 📄 Description ## Summary Fixes #5713. A thin `arm64_32` Mach-O is now admitted as a target, so `list-architectures`, `inspect-macho` and `inspect-signature` inspect it with Apple's native tools instead of refusing it at target admission. ## Problem and expected behavior `parseThinMachO` mapped Mach-O CPU types only to `x86`, `x86_64`, `arm` and `arm64`. `CPU_TYPE_ARM64_32` (`0x0200000c`) fell through to `err("unsupported Mach-O architecture")`, so all three commands failed with `target_unavailable` / `unsupported_target` before any tool ran. The FAT control already showed the inconsistency: the same slice inside a universal file was inventoried by `lipo`. Expected, per the `list_architectures` contract and the issue: the thin object returns ordinary thin-file Evidence. Thin rows keep the existing convention of null CPU code, offset, size and alignment. ## Change and scope - **Domain.** `arm64_32` joins `BINARY_ARCHITECTURES` as a Mach-O-only architecture. This follows the Hexagon precedent from #4819: it is excluded from `HeaderArchitecture` and gets its own thin `mach-o` variants in `ExecutableMetadata` and `BinaryTarget`. As a result, the ELF, PE, raw-binary and universal paths cannot carry it, and the ELF fixed-header type now says so explicitly. - **Native slice policy.** `MACHO_ARCHITECTURE_POLICIES` gains `arm64_32`. Without it, the newly admitted target reached the 32-bit dispatch-metadata reader and failed as `execution_failure`: "Mach-O CPU type does not match requested architecture arm64_32". With it, that reader raises its existing unsupported-32-bit error, like other 32-bit Mach-O images. - **Deep-analysis providers.** Ghidra, Hopper and IDA have no verified `arm64_32` lane, so they decline the target with `architecture_unsupported`: - Ghidra, through its existing supported-architecture set; - Hopper, through one `isHopperDeclinedArchitecture` predicate that now replaces the two duplicated checks; - IDA, through an explicit refusal. Kuna already declines it through its verified-architecture set. Not included: - **FAT files.** Universal `availableArchitectures` still lists only the four provider-neutral families, so `arm64_32` slices in FAT files remain visible only through `lipo` inventory, as before. - **Duplicated CPU-type maps.** These existed before this PR. This PR updates the two that `arm64_32` targets reach: `machArchitecture` admission and the native slice policy. It leaves two unchanged: - `machoTargetSelectionMatchesArchitecture`, which only deep-analysis profiles reach; - Hopper's slice CPU map. Consolidating them is a separate refactor. Contribution scope: ordinary fix within existing boundaries; maintainer scope decision not applicable. ## Contract and boundary impact - Semantic owner and earliest changed stage: target admission, thin Mach-O header parsing in `src/domain/binaryTarget.ts`. - Contribution scope fit and linked maintainer decision: not applicable (bug fix). - CLI and MCP/tool-catalog contract: additive. Schemas that enumerate `BINARY_ARCHITECTURES` gain `arm64_32`, for example Evidence `subject.architecture`. The value appears only for targets that were previously rejected. - Provider, bridge, target-format, or platform compatibility: native macOS tools inspect thin `arm64_32` targets. Ghidra, Hopper, IDA and Kuna decline them. - Evidence, artifact, provenance, or reconstruction contract: thin `arm64_32` Evidence carries `format: "mach-o"` and `architecture: "arm64_32"` with native-tool provenance. - Target selection, authority, upstream behavior, and cleanup details: thin targets stay unfiltered, so no `-arch` is passed. FAT selection is unchanged. - Generated metadata, package, or installation impact: none committed. `npm run docs:check` passes. - [x] Read existing files or captured data. ## Evidence and regression coverage - Tests added or updated: - `tests/boundary/cli/nativeMachoArm64_32Cli.test.ts` (macOS only): compiles a header-free object with `xcrun clang -arch arm64_32 -c` and checks the public CLI. It is listed in the macOS `apple-artifacts` job and in `SCOPES.apple`, so CI runs it. - All three commands exit 0 with `subject.architecture: "arm64_32"`. - `list-architectures` lists exactly `arm64_32`. - `inspect-native-dispatch-metadata` returns the typed refusal `capability_unavailable` / `target_unsupported`, `retryable: false`. - Parser row in `binaryTarget.test.ts`, plus refusal tests for Ghidra, Hopper and IDA. - `tests/fixtures/nativeCommands.ts`: the arm64 fixture helper's return type is narrowed to what it builds. - Base reproduction: - The issue's nine-command probe on `7d8e4e433` (macOS arm64, Apple clang): all three thin `arm64_32` commands exit 1 with `unsupported Mach-O architecture`; the `arm64` and FAT controls exit 0. After the change, all nine exit 0. - The new CLI test fails on the unmodified source with that same `target_unavailable` error. - With only the slice policy removed, it fails on the `execution_failure` CPU-mismatch error described above. - User-visible CLI output (`list-architectures arm64_32.o --json`, abridged): ```json "subject": { "format": "mach-o", "architecture": "arm64_32" }, "normalized_result": { "architectures": { "items": [ { "name": "arm64_32", "cpu_type": null, "cpu_subtype": null, "file_offset": null, "size": null, "alignment": null } ], "total": 1, "exhaustive": true, "limitations": [] } } ``` - Remaining proof gaps: - No linked watchOS executable was tested; only a compiler-produced object was used. - Provider refusals are verified with unit tests only; no real Ghidra, Hopper or IDA run. - MCP was not exercised directly. From the source, it uses the same target admission as the CLI. For evidence-bearing changes: - [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 Node 24.18.0, macOS 27.0 arm64, Apple CommandLineTools clang, lipo, otool and codesign. Head `b0110810a` on `7501911f1`. - `npm run check:fast` — passed (typecheck, lint 0 errors) - `npm run knip` — passed - `npm run docs:check` — passed - `npm run test:focused --` on every test under `src/domain`, `src/hopper`, `src/ida`, `src/ghidra`, `src/native` and `tests/boundary/providers/native`, `src/application/binaryDiagnostics`, both native Mach-O CLI tests, and the four CI planning tests — 454 files, 9072 tests passed, 2 skipped - `scripts/ci/plan.mjs`, simulated as a `pull_request` from `7501911f1` to this head — `apple=true` - Full `npm run check:pr` — not run ## Compatibility, safety, and release - Breaking changes or migration steps: none. The new enum value appears only for targets that were rejected before. - Real Hopper/Ghidra, browser, or OS coverage: real macOS Apple tools only; no deep-analysis provider installed. - Package or release metadata impact: none. - Security, privacy, process, or containment review: static file inspection only. No target execution. ## Author readiness Human owner: @Lew1sWong - [x] I reviewed the final diff and relevant existing code, understand the change and its edge cases, and take responsibility for the complete PR, including AI-assisted work. - [x] I checked the evidence above, confirmed it supports the current change, and recorded meaningful validation gaps. - [x] I declare this PR ready for initial maintainer review. ## Before merge - [x] This PR addresses a concrete problem or an agreed enhancement. - [x] Material scope and authority questions are resolved, and any required maintainer scope decision covers the final diff. - [x] The PR has one focused outcome and the title follows `type(scope): outcome`. - [x] Related issue is linked, or the reason for not linking one is stated above. - [x] Tests cover changed observable behavior and meaningful failure paths. - [x] Owning docs, contracts, and generated metadata are updated where needed. - [x] User-visible CLI/MCP changes include representative output. - [x] I checked the final diff for secrets, unrelated cleanup, and unsupported claims. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-11-01 21:45:04 +00:00

closed this issue
added the
pull-request
label

bsakaguchi referenced this issue

2026-11-03 13:39:07 +00:00

[PR #7061] [MERGED] docs(i18n): add Japanese translation #7614

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

morluto/rea#7061

No description provided.

Rows
Columns