[qdaxb]

What changed

• Move the executor Cargo target-directory setup out of the job-level expression context.
• Export CARGO_TARGET_DIR through GITHUB_ENV using the runner-provided RUNNER_TEMP environment variable.

Why

The Tests workflow referenced runner.temp from a job-level env block, where the runner context is unavailable. GitHub rejected the entire workflow during parsing, so none of the unit-test jobs, including Test Wework, were created while the separate Wework E2E workflow continued to run.

Impact

Pull requests and protected-branch pushes can run the normal Tests workflow again. Executor build artifacts remain on the runner temporary disk.

Validation

• actionlint -ignore 'shellcheck reported issue' -ignore 'runner of .* action is too old' .github/workflows/test.yml
• git diff --check origin/main...HEAD

Summary by CodeRabbit

• Bug Fixes

  • Improved macOS executor signal-handler compatibility.
  • Increased reliability of runtime, IPC, backend, and streaming behavior tests.
  • Improved handling of interrupted executor tasks and runtime authentication paths.

• Documentation

  • Clarified required module checks for pull requests.
  • Documented the distinction between module tests and separate end-to-end test workflows.

• Chores

  • Updated continuous integration coverage reporting and Rust environment setup.
  • Improved test failure handling and build-cache verification.

[coderabbitai]

📝 Walkthrough

Walkthrough

Updates CI workflows and testing documentation, while aligning executor source and contract tests with current runtime, IPC, backend path, streaming, synchronization, and signal-handling behavior.

Changes

CI testing and documentation

Layer / File(s)	Summary
Workflow execution and coverage updates .github/workflows/test.yml	CI jobs source Cargo environment files, configure CARGO_TARGET_DIR through GITHUB_ENV, handle shared-test exit code 134, and upgrade Codecov actions.
Workflow test documentation docs/en/developer-guide/testing.md, docs/zh/developer-guide/testing.md	Testing guides document ordered module checks, summary-job behavior, and separate E2E workflows.

Executor contract tests

Layer / File(s)	Summary
Executor compatibility contracts executor/src/bin/wegent-executor.rs, executor/src/browser_mcp.rs, executor/tests/build_entrypoint_contract.rs	Signal-handler casting, browser request arguments, and dev-sidecar cache assertions are updated.
Runtime transcript contracts executor/tests/app_runtime_work_contract.rs	Obsolete transcript-cache and running-tool coverage is removed, provider paging coverage is renamed, and transcript-read expectations are updated.
Send flow synchronization and fixtures executor/tests/app_runtime_work_send_contract.rs	Environment locking, transcript assertions, cancellation timing, and fake Codex turn scripts are updated.
IPC environment locking and shutdown executor/tests/local_app_ipc_contract.rs	IPC tests use standard-library environment locks and await aborted server tasks before cleanup.
Backend paths and streaming events executor/tests/local_backend_contract.rs	Backend tests derive runtime auth paths, serialize environment changes, collect large-message deltas, and broaden environment helper compilation.

Estimated code review effort: 4 (Complex) | ~45 minutes

Possibly related PRs

• wecode-ai/Wegent#1843: Related changes to the executor CI job’s Rust target-directory configuration.
• wecode-ai/Wegent#1901: Related dev-sidecar executor cache and layout assertions.
• wecode-ai/Wegent#1968: Related executor runtime contract changes for transcript paging and caching.

Suggested reviewers: ficohu

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: restoring the CI test workflow so it executes again.
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.

✨ 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 fix/ci-test-workflow

---

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.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@executor/tests/app_runtime_work_send_contract.rs`:
- Line 8: Update the test lock initialization and acquisition around StdMutex
and StdMutexGuard so a poisoned mutex is recovered instead of causing a later
.expect(...) failure. Preserve the lock’s serialization behavior by accepting
the guard from the poison error and returning a usable StdMutexGuard to callers.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 96a72981-bda0-43ad-a188-9159a4a419ed

📥 Commits

Reviewing files that changed from the base of the PR and between 2bfd2bd and 8c0fec0.

📒 Files selected for processing (10)

• .github/workflows/test.yml
• docs/en/developer-guide/testing.md
• docs/zh/developer-guide/testing.md
• executor/src/bin/wegent-executor.rs
• executor/src/browser_mcp.rs
• executor/tests/app_runtime_work_contract.rs
• executor/tests/app_runtime_work_send_contract.rs
• executor/tests/build_entrypoint_contract.rs
• executor/tests/local_app_ipc_contract.rs
• executor/tests/local_backend_contract.rs

[coderabbitai]

🩺 Stability & Availability | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Map the test file and inspect the relevant section with line numbers.
ast-grep outline executor/tests/app_runtime_work_send_contract.rs --view expanded || true
echo '--- file excerpt ---'
sed -n '1,120p' executor/tests/app_runtime_work_send_contract.rs | cat -n

echo '--- search for env_lock usage ---'
rg -n "env_lock\(|StdMutex|OnceLock" executor/tests/app_runtime_work_send_contract.rs executor/tests -g '!**/target/**' || true

Repository: wecode-ai/Wegent

Length of output: 32066

---

Recover from a poisoned test lock.

A panic while this guard is held poisons the mutex, so later tests fail at .expect(...) even though the lock is only for serialization.

Proposed fix

         _guard: LOCK
             .get_or_init(|| StdMutex::new(()))
             .lock()
-            .expect("environment lock should be available"),
+            .unwrap_or_else(|poisoned| poisoned.into_inner()),

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@executor/tests/app_runtime_work_send_contract.rs` at line 8, Update the test
lock initialization and acquisition around StdMutex and StdMutexGuard so a
poisoned mutex is recovered instead of causing a later .expect(...) failure.
Preserve the lock’s serialization behavior by accepting the guard from the
poison error and returning a usable StdMutexGuard to callers.