[PR #2241] [MERGED] fix(managed): decode the CIL no. prefix #5108

New issue

Closed

opened 2026-10-31 01:04:01 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 01:04:01 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/2241
Author: @Angelopgit
Created: 10/12/2026
Status: ✅ Merged
Merged: 10/13/2026
Merged by: @morluto

Base: main ← Head: fix/managed-no-prefix

📝 Commits (1)

48ffdd9 fix(managed): decode the CIL no. prefix

📊 Changes

2 files changed (+27 additions, -0 deletions)

View changed files

📝 src/dotnet/ManagedCilDecoding.test.ts (+25 -0)
📝 src/dotnet/ManagedMemberInstructionDecoder.ts (+2 -0)

📄 Description

Summary

inspect_managed_members reports a method as malformed when its IL uses the no. prefix. The CIL opcode table has five of the six ECMA-335 prefix opcodes and omits no. (FE 19 <unsigned int8>, Partition III §2.2). No issue is linked: I searched open and closed issues and PRs for 0xfe19 and the decoder and found none, and the fix is two table lines.

Problem and expected behavior

Smallest trigger, a method body with this IL:

02              ldarg.0
FE 19 04        no. nullcheck
7B 01 00 00 04  ldfld 0x04000001
26              pop
2A              ret

Observed on main: status: "malformed", issue: "Unsupported CIL opcode 0xfe19 at IL offset 1", normalized_il_sha256: null.

Expected: status: "present", five decoded instructions, issue: null, as for the other prefixes.

Change and scope

Add 0xfe19 as no. with a one-byte operand (short-i, the same kind unaligned. uses).
Add no. to PREFIX_OPCODES, so it shares an instruction boundary with the opcode it modifies and an exception region cannot start between them.

Not included: validation of the check-mask bits or of which opcodes no. may precede. The decoder does not validate that for the other prefixes either.

Contract and boundary impact

Semantic owner and earliest changed stage: CIL instruction decoding in src/dotnet/ManagedMemberInstructionDecoder.ts.
CLI and MCP/tool-catalog contract: none. No schemas change; no. can now appear as an opcode name in opcode_counts.
Provider, bridge, target-format, or platform compatibility: method bodies using no. are decoded instead of marked malformed. Digests of bodies that already decoded are unchanged.
Evidence, artifact, provenance, or reconstruction contract: none.
Process execution, authorization, cleanup, or containment impact: none.
Generated metadata, package, or installation impact: none.

Evidence and regression coverage

Tests added or updated: one test in src/dotnet/ManagedCilDecoding.test.ts ("managed exception region prefix boundaries"). It checks that the body above decodes, and that a try range starting between no. and ldfld is refused as misaligned.
Base reproduction or other evidence: executed. With the decoder change reverted the new test fails with Unsupported CIL opcode 0xfe19 at IL offset 1.
User-visible CLI/MCP output: not applicable beyond the fields above.
Remaining proof gaps: the IL is a hand-built fixture. Mainstream compilers rarely emit no., and I did not have an assembler-produced binary containing it.

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

npx vitest run src/dotnet tests/boundary/dotnet — 11 files passed
npm run check — passed
npm run verify:test-discovery — "Discovered all 936 test files exactly once."

Compatibility, safety, and release

Breaking changes or migration steps: none.
Real Hopper/Ghidra, browser, or OS coverage: Linux x64 only. No provider runs; this path is static parsing.
Package or release metadata impact: none.
Security, privacy, process, or containment review: parsing only; the operand read is bound-checked by the existing decoder path.

Review checklist

This PR addresses a concrete problem or an agreed enhancement.
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

https://claude.ai/code/session_012pRGskbPGjT444CgJfuhhj

_{🔄 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/2241 **Author:** [@Angelopgit](https://github.com/Angelopgit) **Created:** 10/12/2026 **Status:** ✅ Merged **Merged:** 10/13/2026 **Merged by:** [@morluto](https://github.com/morluto) **Base:** `main` ← **Head:** `fix/managed-no-prefix` --- ### 📝 Commits (1) - [`48ffdd9`](https://github.com/morluto/rea/commit/48ffdd9a62c8a81ccc61e9b441ea724ddefeeae4) fix(managed): decode the CIL no. prefix ### 📊 Changes **2 files changed** (+27 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `src/dotnet/ManagedCilDecoding.test.ts` (+25 -0) 📝 `src/dotnet/ManagedMemberInstructionDecoder.ts` (+2 -0) </details> ### 📄 Description ## Summary `inspect_managed_members` reports a method as malformed when its IL uses the `no.` prefix. The CIL opcode table has five of the six ECMA-335 prefix opcodes and omits `no.` (`FE 19 <unsigned int8>`, Partition III §2.2). No issue is linked: I searched open and closed issues and PRs for `0xfe19` and the decoder and found none, and the fix is two table lines. ## Problem and expected behavior Smallest trigger, a method body with this IL: ``` 02 ldarg.0 FE 19 04 no. nullcheck 7B 01 00 00 04 ldfld 0x04000001 26 pop 2A ret ``` Observed on `main`: `status: "malformed"`, `issue: "Unsupported CIL opcode 0xfe19 at IL offset 1"`, `normalized_il_sha256: null`. Expected: `status: "present"`, five decoded instructions, `issue: null`, as for the other prefixes. ## Change and scope - Add `0xfe19` as `no.` with a one-byte operand (`short-i`, the same kind `unaligned.` uses). - Add `no.` to `PREFIX_OPCODES`, so it shares an instruction boundary with the opcode it modifies and an exception region cannot start between them. Not included: validation of the check-mask bits or of which opcodes `no.` may precede. The decoder does not validate that for the other prefixes either. ## Contract and boundary impact - Semantic owner and earliest changed stage: CIL instruction decoding in `src/dotnet/ManagedMemberInstructionDecoder.ts`. - CLI and MCP/tool-catalog contract: none. No schemas change; `no.` can now appear as an opcode name in `opcode_counts`. - Provider, bridge, target-format, or platform compatibility: method bodies using `no.` are decoded instead of marked malformed. Digests of bodies that already decoded are unchanged. - Evidence, artifact, provenance, or reconstruction contract: none. - Process execution, authorization, cleanup, or containment impact: none. - Generated metadata, package, or installation impact: none. ## Evidence and regression coverage - Tests added or updated: one test in `src/dotnet/ManagedCilDecoding.test.ts` ("managed exception region prefix boundaries"). It checks that the body above decodes, and that a try range starting between `no.` and `ldfld` is refused as misaligned. - Base reproduction or other evidence: executed. With the decoder change reverted the new test fails with `Unsupported CIL opcode 0xfe19 at IL offset 1`. - User-visible CLI/MCP output: not applicable beyond the fields above. - Remaining proof gaps: the IL is a hand-built fixture. Mainstream compilers rarely emit `no.`, and I did not have an assembler-produced binary containing it. 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 - `npx vitest run src/dotnet tests/boundary/dotnet` — 11 files passed - `npm run check` — passed - `npm run verify:test-discovery` — "Discovered all 936 test files exactly once." ## Compatibility, safety, and release - Breaking changes or migration steps: none. - Real Hopper/Ghidra, browser, or OS coverage: Linux x64 only. No provider runs; this path is static parsing. - Package or release metadata impact: none. - Security, privacy, process, or containment review: parsing only; the operand read is bound-checked by the existing decoder path. ## Review checklist - [x] This PR addresses a concrete problem or an agreed enhancement. - [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) https://claude.ai/code/session_012pRGskbPGjT444CgJfuhhj --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-31 01:04:01 +00:00