PR #16418: Fix Windows terminal execution so foreground commands reliably return…

• Repository: NousResearch/hermes-agent
• Author: xy200303
• State: open | merged: False
• Link: https://github.com/NousResearch/hermes-agent/pull/16418

Description (problem / solution / changelog)

What does this PR do?

This PR fixes Windows local terminal execution so foreground commands reliably return output instead of appearing to hang or complete with an empty reply.

The fix addresses two Windows-specific issues in the local terminal stack:

1. BaseEnvironment._wait_for_process() used Unix-style select.select() to poll subprocess stdout pipes. That works on Unix, but not reliably for ordinary subprocess pipes on Windows, which could prevent command output from being drained and surfaced back to Hermes.

2. LocalEnvironment had path mismatches between native Windows paths and Git Bash/MSYS paths. This affected:

   • bash executable selection
   • cwd
   • sourced shell init files
   • session artifact paths such as snapshot/cwd temp files

This approach fixes the bug at the correct layers:

• base.py now uses a Windows-compatible stdout polling path
• local.py now bridges Windows and Git Bash paths consistently while still using native Windows paths where subprocess.Popen() requires them

Related Issue

Fixes #

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

• Updated tools/environments/base.py
  • added a Windows-specific stdout pipe polling path using PeekNamedPipe
  • preserved the existing Unix select() polling behavior
  • routed _wait_for_process() through the shared pipe polling helper
• Updated tools/environments/local.py
  • improved Windows bash discovery
  • skipped C:\Windows\System32\bash.exe / sysnative\bash.exe WSL stub
  • preferred real Git for Windows bash.exe
  • searched both bin\bash.exe and usr\bin\bash.exe
  • added Windows <-> Git Bash/MSYS path conversion helpers
  • converted shell init file paths into bash-visible form before sourcing
  • used native Windows paths for subprocess.Popen(..., cwd=...)
  • rewrote local snapshot/cwd artifact paths into bash-visible paths in wrapped commands
  • converted bash-reported cwd values back into native Windows paths before storing state
• Added tests
  • tests/tools/test_base_environment_pipe_poll.py
  • tests/tools/test_local_windows_paths.py

How to Test

1. On Windows, run a foreground local terminal command through Hermes that previously returned no visible reply.
2. Confirm the command output is returned normally instead of appearing to hang or complete with an empty response.
3. Run:
   • python -m pytest tests/tools/test_base_environment_pipe_poll.py tests/tools/test_local_windows_paths.py -q and confirm the targeted regression tests pass.

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

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

Screenshots / Logs

Targeted verification run:

python -m pytest tests/tools/test_base_environment_pipe_poll.py tests/tools/test_local_windows_paths.py -q

Result:

8 passed

Note: I was not able to run the full repo-standard scripts/run_tests.sh / pytest tests/ -q flow in this Windows environment because the local bash/script setup has an independent environment issue. The targeted regression tests for this fix are passing.

#16425

Changed files

• scripts/release.py (modified, +1/-0)
• tests/tools/test_base_environment_pipe_poll.py (added, +47/-0)
• tests/tools/test_local_windows_paths.py (added, +84/-0)
• tools/environments/base.py (modified, +66/-11)
• tools/environments/local.py (modified, +125/-7)
• uv.lock (modified, +161/-2)