fix(tools): prevent terminal probe hang and kill orphaned processes on Windows

[ericmaddox]

What does this PR do?

When testing whether candidate Git Bash executables on Windows can successfully launch MSYS child processes, _bash_starts runs an external program probe (/usr/bin/true; /usr/bin/cat --version >/dev/null).

On Windows, Git for Windows' <root>\bin\bash.exe is a wrapper shim that spawns <root>\usr\bin\bash.exe. When the probe child deadlocks (e.g. MSYS2 fork deadlock), subprocess.run(..., timeout=15) times out and calls proc.kill(). However, on Windows proc.kill() (via TerminateProcess) only terminates the direct wrapper process, leaving the re-exec'd usr\bin\bash.exe process alive and holding open the stdout/stderr pipe handles.

Because Python's standard library subprocess.run implementation calls proc.communicate() without a timeout inside its except TimeoutExpired block to drain pipes, it blocks indefinitely waiting for EOF on the still-open pipes. This causes the entire terminal environment creation to hang for minutes past the timeout.

Fix

1. Implement _run_probe_command with explicit process tree termination on Windows (taskkill /F /T /PID <pid>) upon timeout before killing the process, ensuring all child/grandchild processes are stopped and pipes are closed.
2. Tighten probe timeout from 15s to 5s (a healthy probe finishes in <0.2s).
3. Update _windows_bash_candidates to probe <root>\usr\bin\bash.exe directly alongside <root>\bin\bash.exe, avoiding the extra wrapper process layer.
4. Expand _MSYS_SPAWN_FAILURE_MARKERS to recognize timeout/deadlock probe diagnostics as MSYS spawn failures, ensuring actionable remediation guidance is presented if all candidates fail.

Related Issue

Fixes #103398

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• tools/environments/local_gitbash_probe.py: Replace unbounded subprocess.run with _run_probe_command which executes taskkill /F /T /PID <pid> on Windows timeout. Tighten timeout to 5s. Add timeout/deadlock markers to _MSYS_SPAWN_FAILURE_MARKERS.
• tools/environments/local.py: In _windows_bash_candidates, include usr/bin bash candidate paths and resolve direct usr�in�ash.exe for custom paths.
• tests/tools/test_find_shell.py: Update unit tests to mock _run_probe_command and add test_probe_timeout_records_deadlock_diagnostic asserting that probe timeouts fail cleanly and record MSYS failure diagnostics.

How to Test

1. Run unit tests covering shell resolution and probe execution:
   pytest tests/tools/test_find_shell.py -v
2. Verify all tests pass and test_probe_timeout_records_deadlock_diagnostic validates that a hung probe times out and registers as an MSYS failure.

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: Windows 11 / Python 3.11

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

[alt-glitch]

This was generated by AI during triage.

Related: #83413 (earlier open PR, same tree-kill approach for the orphaned usr\bin\bash.exe grandchild, plus stdin handling under ACP embedding) and issues #74982 / #103398. This PR additionally probes usr\bin\bash.exe directly and tightens the probe timeout. Competing fixes for the same hang — flagging so a maintainer can consolidate.