fix(ecc2): resolve duplicate kill_process definition that breaks the Windows build

[mohameddsh3ban]

Problem

cargo build fails on Windows with a duplicate-definition error:

error[E0428]: the name `kill_process` is defined multiple times
    --> src\session\manager.rs:3638:1
     |
3609 | fn kill_process(pid: u32) -> Result<()> {
     | --------------------------------------- previous definition of the value `kill_process` here
...
3638 | async fn kill_process(pid: u32) -> Result<()> {
     | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ `kill_process` redefined here
     |
     = note: `kill_process` must be defined only once in the value namespace of this module
For more information about this error, try `rustc --explain E0428`.
error: could not compile `ecc-tui` (bin "ecc-tui") due to 1 previous error

ecc2/src/session/manager.rs defines three cfg-gated kill_process variants:

• #[cfg(unix)] — sync, libc signals
• #[cfg(windows)] — sync, taskkill
• #[cfg(not(unix))] — async, taskkill

On Windows, both cfg(windows) and cfg(not(unix)) evaluate true, so the sync and async versions both compile in and collide. This affects every Windows target — the overlap is independent of toolchain, so x86_64-pc-windows-msvc hits the identical collision as x86_64-pc-windows-gnu. Unix builds are unaffected because only cfg(unix) matches there — which is why Linux/macOS CI never caught it.

Fix

Delete the stray #[cfg(not(unix))] async definition. It was unusable regardless of the collision — both call sites are synchronous and never .await it:

• enforce_session_heartbeats_with<F> is bound where F: Fn(u32) -> Result<()> (a sync fn) and receives kill_process as a plain fn-pointer, invoked as terminate_pid(pid) with no await;
• stop_session_recorded calls kill_process(pid)? synchronously.

The surviving #[cfg(windows)] sync version is also strictly better than the deleted async one: it passes /T to taskkill (terminates the whole child process tree, so spawned agent subprocesses aren't orphaned) and surfaces the actual exit status on failure. The async version had neither.

One file, 18-line deletion. Verified on Windows (x86_64-pc-windows-gnu): builds clean, cargo check passes, and no other cfg(windows) / cfg(not(unix)) overlaps exist in the crate.

---

Summary by cubic

Fix Windows build by removing a duplicate kill_process definition. The async #[cfg(not(unix))] version is deleted so only the Windows sync taskkill implementation remains.

• Bug Fixes
  • Removed the #[cfg(not(unix))] async kill_process that collided with #[cfg(windows)] (E0428).
  • Windows builds now succeed; callers remain sync and use the taskkill /T /F implementation.

^{Written for commit 545fcf4. Summary will update on new commits.}

Summary by CodeRabbit

• Refactor
  • Streamlined internal process management logic by removing deprecated code paths and consolidating platform-specific implementations.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: de44f520-08ee-42de-bf2b-d24f6e2522b7

📥 Commits

Reviewing files that changed from the base of the PR and between 90dfd95 and 545fcf4.

📒 Files selected for processing (1)

• ecc2/src/session/manager.rs

💤 Files with no reviewable changes (1)

• ecc2/src/session/manager.rs

---

📝 Walkthrough

Walkthrough

This PR removes an unused async process-termination function. The #[cfg(not(unix))] variant of kill_process that invoked taskkill is deleted from the session manager. Platform-gated implementations for Unix and Windows process handling remain in place.

Changes

Process Termination Cleanup

Layer / File(s)	Summary
Remove non-Unix kill_process implementation ecc2/src/session/manager.rs	Deleted the async taskkill-based process termination function gated for non-Unix systems. Remaining platform-specific process termination logic remains elsewhere in the file.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Poem

A taskkill that never took flight,
Now gone from the code, out of sight,
Process cleanup done right,
One function removed with delight, 🐰

🚥 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 directly and specifically addresses the main change: removing a duplicate kill_process definition that was breaking Windows builds due to cfg attribute conflicts.
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.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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.}

[cubic-dev-ai]

No issues found across 1 file

_{Re-trigger cubic}

[alexraphael07-prog]

🤖 Bezalel Engine — Automated Code Review

1. Summary

PR #2195 intends to unblock the Windows build by removing the redundant async fn kill_process that shadows the sync fn kill_process shared by all platforms. The linter fails because only one definition may exist across cfg-blocks.

2. Issues

No bugs or security issues are introduced. However, the PR is incomplete:

• Lines 3613–3635: the sync kill_process remains only guarded with #[cfg(not(windows))].
  On Windows this leaves zero implementations →