ADE-60: Local runtime: ensureProject() bypasses isClosed()/retry, so first read after daemon recycle still fails raw

[arul28]

Fixes ADE-60

Summary

• Retry local runtime project registration once when the cached daemon connection is already closed or drops during projects.add.
• Reset the stale connection before retrying so the next connect reaches the live daemon.
• Add regression coverage for both the pre-call isClosed path and the projects.add dropped-connection path before a read action.

Validation

• npx vitest run src/main/services/localRuntime/localRuntimeConnectionPool.test.ts
• npm --prefix apps/desktop run typecheck
• npx eslint src/main/services/localRuntime/localRuntimeConnectionPool.ts src/main/services/localRuntime/localRuntimeConnectionPool.test.ts

ADE PR: https://ade-app.dev/open?type=pr&repo=arul28%2FADE&number=438

Summary by CodeRabbit

Release Notes

• Bug Fixes
  • Improved reliability of local runtime connections with automatic reconnection handling. When a cached runtime connection drops or becomes unusable, the system now automatically retries project registration up to two times, ensuring smoother operation during temporary connection interruptions.

Greptile Summary

This PR adds retry logic to ensureProject() so a stale/closed daemon connection detected either before or during the projects.add call is transparently recovered from, matching the existing pattern already in callActionForRootUncoalesced.

• ensureProject now loops up to two times: on the first attempt it checks isClosed() before calling projects.add, and catches connection-dropped errors thrown during the call itself; on each failure it calls resetActiveConnection (which clears this.connection and projectsByRoot) and retries with a fresh connection.
• Two regression tests are added: one covering the isClosed() pre-call path and one covering the mid-projects.add drop path that then continues all the way through a successful callActionForRoot.

Confidence Score: 5/5

Safe to merge — the retry logic is narrowly scoped to connection-dropped errors, follows the identical pattern already used in callActionForRootUncoalesced, and both new regression tests pass end-to-end through a successful action call.

The change is a focused two-attempt retry loop that only activates on connection-dropped errors in ensureProject. resetActiveConnection correctly clears all shared state (this.connection, activeConnection, projectsByRoot) before each retry, so no stale data can leak across attempts. The existing callActionForRootUncoalesced retry loop interacts cleanly with this: if ensureProject exhausts its retries it throws immediately, and if ensureProject succeeds the action-call retry path continues to work as before.

No files require special attention.

Important Files Changed

Filename	Overview
apps/desktop/src/main/services/localRuntime/localRuntimeConnectionPool.ts	Adds a 2-attempt retry loop to ensureProject() mirroring the existing isClosed/connection-drop handling in callActionForRootUncoalesced; resetActiveConnection correctly clears connection state and projectsByRoot before each retry.
apps/desktop/src/main/services/localRuntime/localRuntimeConnectionPool.test.ts	Two new test cases cover the isClosed() pre-call path and the mid-call connection-drop path; both correctly mock createConnection to return fresh entries on retry and verify call counts, log events, and end-to-end results.

Sequence Diagram

sequenceDiagram
    participant C as callActionForRootUncoalesced
    participant EP as ensureProject
    participant Pool as ConnectionPool
    participant D1 as Daemon (stale)
    participant D2 as Daemon (fresh)

    C->>EP: ensureProject(rootPath)
    EP->>Pool: connect() → entry1 (stale)
    alt "isClosed() == true"
        EP->>Pool: resetActiveConnection(entry1)
        Note over EP: clears connection + projectsByRoot
    else projects.add drops
        EP->>D1: projects.add
        D1-->>EP: connection dropped
        EP->>Pool: resetActiveConnection(entry1)
    end
    EP->>Pool: connect() → entry2 (fresh)
    EP->>D2: projects.add
    D2-->>EP: project record
    EP-->>C: project (cached)
    C->>Pool: connect() → entry2 (same cached)
    C->>D2: ade/actions/call (projectId)
    D2-->>C: action result

Loading

_{Reviews (8): Last reviewed commit: "Refs ADE-60: fix: document unreachable p..." | Re-trigger Greptile}

[linear-code]

ADE-60

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
ade	Ignored	Preview	May 30, 2026 8:34am

[coderabbitai]

📝 Walkthrough

Walkthrough

The PR adds retry logic to the ensureProject method to handle dropped runtime daemon connections, allowing the connection pool to automatically reconnect and retry project registration instead of failing. Two new test cases validate this behavior across different failure scenarios.

Changes

Connection drop retry mechanism and tests

Layer / File(s)	Summary
Retry logic for dropped runtime connections apps/desktop/src/main/services/localRuntime/localRuntimeConnectionPool.ts	ensureProject wraps the cached connection check and projects.add call in a 2-attempt retry loop. It detects closed connections via isClosed(), logs local_runtime.ensure_project_connection_dropped, resets the active connection entry, and retries; other errors are rethrown immediately, and a final error is thrown after exhausting attempts.
Tests for connection drop retry scenarios apps/desktop/src/main/services/localRuntime/localRuntimeConnectionPool.test.ts	Two test cases validate that when a cached runtime client drops before a read action, the pool reconnects and retries projects.add on a fresh client before performing the read action; and when the client is closed before project registration, the pool reconnects first so projects.add runs only on the new client. Both assert the dropped-connection warning log.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Possibly related issues

• Local runtime: ensureProject() bypasses isClosed()/retry, so first read after daemon recycle still fails raw #415: Addresses the core issue by implementing isClosed checks and retry logic inside ensureProject to handle dropped/closed runtime connections that were previously not caught by the existing retry mechanism.

Suggested labels

desktop

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title directly and specifically describes the main change: adding retry logic to ensureProject() to handle closed daemon connections after daemon recycle.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch ade-60-local-runtime-ensureproject-bypasses-isclosed-retry-so-first-read-after-daemon-recycle-still-fails-raw

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[capy-ai]

Capy auto-review is paused for this organization because the monthly auto-review limit has been reached. Increase the limit or turn it off in billing settings to resume automatic reviews.

[arul28]

@copilot review but do not make fixes