[PR #4049] [MERGED] fix(kuna): verify a warm export once for concurrent first requests #6019

New issue

Closed

opened 2026-10-31 01:11:12 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 01:11:12 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/4049
Author: @kaoru0822-kitauji
Created: 10/15/2026
Status: ✅ Merged
Merged: 10/15/2026
Merged by: @kaoru0822-kitauji

Base: main ← Head: fix/kuna-share-export-adoption

📝 Commits (1)

9944762 fix(kuna): verify a warm export once for concurrent first requests

📊 Changes

3 files changed (+46 additions, -4 deletions)

View changed files

📝 docs/kuna-provider.md (+1 -1)
📝 src/kuna/KunaSessionClient.ts (+13 -2)
📝 tests/boundary/providers/kuna/kunaProvider.test.ts (+32 -1)

📄 Description

Summary

A Kuna session decided whether to reuse a cached export by checking #project and then awaiting cache.lookup(). Concurrent first requests all passed that check before any lookup finished, so each one ran its own lookup. On a fresh server a lookup re-hashes every export file and re-parses the .c and inventory. batch_decompile resolves its addresses in parallel, so it verified a warm export once per address. Refs #1957.

Problem and expected behavior

Smallest trigger: warm cache, new MCP server, open_binary with provider_id: "kuna", then one batch_decompile with N addresses. Expected: one verification per session, as for a single first request. Observed: N verifications; the cost grows with N × export size.

Change and scope

#startProject now keeps the in-flight lookup in #starting. Concurrent first requests await that one lookup, which then adopts the export or starts the background build. The promise is cleared when it settles, so a failed lookup is retried by the next request, and the existing eviction recovery (#invalidateProject followed by #startProject) works as before. One line in docs/kuna-provider.md states the behavior.

Not included: a lookup shared across sessions of one process. Verification stays per session; concurrent sessions for one target in one process are uncommon, and sharing it would need waiter-owned cancellation in the cache.

Scope: an ordinary fix within the existing Kuna provider boundary; no maintainer scope decision applies.

Contract and boundary impact

Semantic owner and earliest changed stage: KunaSessionClient.#startProject (session-level adoption of the cache's verified export).
Contribution scope fit and linked maintainer decision: not applicable.
CLI and MCP/tool-catalog contract: unchanged.
Provider, bridge, target-format, or platform compatibility: Kuna only.
Evidence, artifact, provenance, or reconstruction contract: unchanged. Answers still name the same verified export and file spans.
Target selection, authority, upstream behavior, and cleanup details: no new processes or writes. The lookup still uses the session signal, so close_binary cancels it as before.
Generated metadata, package, or installation impact: none.

Expected effects:

Read existing files or captured data.
Launch or attach to a process.
Inject instrumentation or run scripts/code.
Change process memory, registers, files, or configuration.
Contact network services.
Recover protected code or work around protections.
Documentation or usability only; no target interaction.
Not sure yet; investigation or maintainer guidance needed.

Evidence and regression coverage

Tests added or updated: kunaProvider.test.ts gets "verifies a warm export once for concurrent first requests on a fresh server". It builds an export, opens a new KunaProvider (a fresh cache instance, like a new server) over the same cache root, sends four concurrent pseudocode requests (names, an address and a duplicate), and asserts all four come from project-export with exactly one KunaProjectCache.lookup call.
Base reproduction: the new test fails on the base revision (git stash of src/kuna/KunaSessionClient.ts) at toHaveBeenCalledTimes(1).

Real Kuna 1.833, macOS arm64, fresh MCP server, warm 29 MiB export of mDNSResponder thinned to arm64e (1.7 MiB, 3750 inventory entries, REA_KUNA_JOBS=8), batch_decompile of 12 sized entries, 3 runs per build:

Build	`batch_decompile`	Whole process wall	Whole process user CPU
main	668 / 671 / 671 ms	1.20 / 1.20 / 1.21 s	1.43 / 1.44 / 1.44 s
this PR	91 / 94 / 100 ms	0.64 / 0.61 / 0.62 s	0.87 / 0.79 / 0.80 s

For comparison, a single first procedure_pseudo_code on the same warm export takes 88–93 ms on both builds. Response bytes are identical (156318). Process figures come from /usr/bin/time -p around the harness and include the MCP client process. Model tokens were not measured. The time saved grows with export size and the number of addresses.

Remaining proof gaps: no Linux host run of this change; cross-session duplicate lookups in one process remain (see scope).
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

npm run test:focused -- tests/boundary/providers/kuna/*.test.ts: 90 passed (10 files), three consecutive runs.
New test on the base revision: failed as described above.
npm run check:fast: typecheck and lint passed.
npm run docs:check: passed.
REA_KUNA_COMMAND=… npm run verify:kuna (real Kuna 1.833, darwin/arm64): passed, process lineage verified, cleanup: "complete".
Real MCP measurements above.

Compatibility, safety, and release

Breaking changes or migration steps: none.
Real Hopper/Ghidra, browser, or OS coverage: real Kuna on macOS arm64 only.
Package or release metadata impact: none.
Security, privacy, process, or containment review: no new side effects. The same verification runs fewer times.

Author readiness

Human owner:

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/4049 **Author:** [@kaoru0822-kitauji](https://github.com/kaoru0822-kitauji) **Created:** 10/15/2026 **Status:** ✅ Merged **Merged:** 10/15/2026 **Merged by:** [@kaoru0822-kitauji](https://github.com/kaoru0822-kitauji) **Base:** `main` ← **Head:** `fix/kuna-share-export-adoption` --- ### 📝 Commits (1) - [`9944762`](https://github.com/morluto/rea/commit/9944762dc3f62fd95dc6b829b7bbaa8496e804cc) fix(kuna): verify a warm export once for concurrent first requests ### 📊 Changes **3 files changed** (+46 additions, -4 deletions) <details> <summary>View changed files</summary> 📝 `docs/kuna-provider.md` (+1 -1) 📝 `src/kuna/KunaSessionClient.ts` (+13 -2) 📝 `tests/boundary/providers/kuna/kunaProvider.test.ts` (+32 -1) </details> ### 📄 Description ## Summary A Kuna session decided whether to reuse a cached export by checking `#project` and then awaiting `cache.lookup()`. Concurrent first requests all passed that check before any lookup finished, so each one ran its own lookup. On a fresh server a lookup re-hashes every export file and re-parses the `.c` and inventory. `batch_decompile` resolves its addresses in parallel, so it verified a warm export once per address. Refs #1957. ## Problem and expected behavior Smallest trigger: warm cache, new MCP server, `open_binary` with `provider_id: "kuna"`, then one `batch_decompile` with N addresses. Expected: one verification per session, as for a single first request. Observed: N verifications; the cost grows with N × export size. ## Change and scope `#startProject` now keeps the in-flight lookup in `#starting`. Concurrent first requests await that one lookup, which then adopts the export or starts the background build. The promise is cleared when it settles, so a failed lookup is retried by the next request, and the existing eviction recovery (`#invalidateProject` followed by `#startProject`) works as before. One line in `docs/kuna-provider.md` states the behavior. Not included: a lookup shared across sessions of one process. Verification stays per session; concurrent sessions for one target in one process are uncommon, and sharing it would need waiter-owned cancellation in the cache. Scope: an ordinary fix within the existing Kuna provider boundary; no maintainer scope decision applies. ## Contract and boundary impact - Semantic owner and earliest changed stage: `KunaSessionClient.#startProject` (session-level adoption of the cache's verified export). - Contribution scope fit and linked maintainer decision: not applicable. - CLI and MCP/tool-catalog contract: unchanged. - Provider, bridge, target-format, or platform compatibility: Kuna only. - Evidence, artifact, provenance, or reconstruction contract: unchanged. Answers still name the same verified export and file spans. - Target selection, authority, upstream behavior, and cleanup details: no new processes or writes. The lookup still uses the session signal, so `close_binary` cancels it as before. - Generated metadata, package, or installation impact: none. Expected effects: - [x] Read existing files or captured data. - [ ] Launch or attach to a process. - [ ] Inject instrumentation or run scripts/code. - [ ] Change process memory, registers, files, or configuration. - [ ] Contact network services. - [ ] Recover protected code or work around protections. - [ ] Documentation or usability only; no target interaction. - [ ] Not sure yet; investigation or maintainer guidance needed. ## Evidence and regression coverage - Tests added or updated: `kunaProvider.test.ts` gets "verifies a warm export once for concurrent first requests on a fresh server". It builds an export, opens a new `KunaProvider` (a fresh cache instance, like a new server) over the same cache root, sends four concurrent pseudocode requests (names, an address and a duplicate), and asserts all four come from `project-export` with exactly one `KunaProjectCache.lookup` call. - Base reproduction: the new test fails on the base revision (`git stash` of `src/kuna/KunaSessionClient.ts`) at `toHaveBeenCalledTimes(1)`. - Real Kuna 1.833, macOS arm64, fresh MCP server, warm 29 MiB export of `mDNSResponder` thinned to arm64e (1.7 MiB, 3750 inventory entries, `REA_KUNA_JOBS=8`), `batch_decompile` of 12 sized entries, 3 runs per build: | Build | `batch_decompile` | Whole process wall | Whole process user CPU | | --- | --- | --- | --- | | main | 668 / 671 / 671 ms | 1.20 / 1.20 / 1.21 s | 1.43 / 1.44 / 1.44 s | | this PR | 91 / 94 / 100 ms | 0.64 / 0.61 / 0.62 s | 0.87 / 0.79 / 0.80 s | For comparison, a single first `procedure_pseudo_code` on the same warm export takes 88–93 ms on both builds. Response bytes are identical (156318). Process figures come from `/usr/bin/time -p` around the harness and include the MCP client process. Model tokens were not measured. The time saved grows with export size and the number of addresses. - Remaining proof gaps: no Linux host run of this change; cross-session duplicate lookups in one process remain (see scope). - [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 - `npm run test:focused -- tests/boundary/providers/kuna/*.test.ts`: 90 passed (10 files), three consecutive runs. - New test on the base revision: failed as described above. - `npm run check:fast`: typecheck and lint passed. - `npm run docs:check`: passed. - `REA_KUNA_COMMAND=… npm run verify:kuna` (real Kuna 1.833, darwin/arm64): passed, process lineage verified, `cleanup: "complete"`. - Real MCP measurements above. ## Compatibility, safety, and release - Breaking changes or migration steps: none. - Real Hopper/Ghidra, browser, or OS coverage: real Kuna on macOS arm64 only. - Package or release metadata impact: none. - Security, privacy, process, or containment review: no new side effects. The same verification runs fewer times. ## Author readiness Human owner:  - [ ] 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](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

2026-10-31 01:11:12 +00:00