molecule-ai/hermes-agent
35
0
Fork 0
Code Issues 2 Pull Requests 4 Actions 9 Packages Projects Releases Wiki Activity

fix(voice_mode): restore audio-env detection across clean/WSL/Termux scenarios (partial close hermes-agent#9) #12

Merged
claude-ceo-assistant merged 1 commits from fix/voice-mode-detect-audio-env-container-stub into main 2026-05-08 20:48:49 +00:00
Conversation 0 Commits 1 Files Changed 2 +25 -2
claude-ceo-assistant commented 2026-05-08 20:48:05 +00:00
Owner
Copy Link

Summary

4 tests in tests/tools/test_voice_mode.py::TestDetectAudioEnvironment were failing on CI with assert False is True:

  • test_clean_environment_is_available
  • test_wsl_with_pulse_allows_voice
  • test_wsl_device_query_fails_with_pulse_continues
  • test_termux_api_microphone_allows_voice_without_sounddevice

Root cause (regressing commit)

Commit 5e1197a4 (fix(gateway): harden Docker/container gateway pathway) replaced an inline os.path.exists('/.dockerenv') check in detect_audio_environment() with the more thorough is_container() helper from hermes_constants. The new helper also looks at /run/.containerenv and /proc/1/cgroup for docker/podman//lxc/ markers, and caches the result at module level (_container_detected).

CI runners themselves run inside Docker. So is_container() returns True on CI, and the function silently appended "Running inside Docker container -- no audio devices" to the warnings list of every scenario — including the four test fixtures that legitimately set up a non-container environment (clean Linux, WSL+pulse, WSL with broken device query, Termux API).

The other 5 tests in the same class survived only by coincidence — they all expected blocking warnings (SSH, missing PulseAudio, missing PortAudio), so the extra container warning just stacked harmlessly on top.

Per feedback_no_such_thing_as_flakes: the failure was a real environmental coupling bug, not a flake.

Fix

  1. Hoist from hermes_constants import is_container to module-level in tools/voice_mode.py so tests can reach it as tools.voice_mode.is_container — matches the existing monkeypatch convention used elsewhere (shutil, _import_audio, _termux_api_app_installed, etc).
  2. Add an autouse fixture in TestDetectAudioEnvironment that defaults is_container to False, so tests don't inherit the host runner's container state.
  3. Add test_docker_container_blocks_voice to pin the original container-blocks-voice intent that was previously implicit.

Net: +25 / -2 lines across tools/voice_mode.py and tests/tools/test_voice_mode.py.

Test plan

  • pytest tests/tools/test_voice_mode.py::TestDetectAudioEnvironment -v — all 10 tests pass (9 existing + 1 new)
  • pytest tests/tools/test_voice_mode.py — all 61 tests pass (no sibling regression)
  • Manually simulated is_container=True (CI-in-Docker) state — autouse fixture neutralizes it; available=True as the tests expect
  • Confirm green on Gitea Actions CI

Out of scope

The ~17-43 unrelated failures in tests/tools/test_voice_cli_integration.py (pre-existing, identical with this branch's changes stashed) are being handled by parallel agents per the dispatch brief — partial close only.

Partial close hermes-agent#9.

## Summary 4 tests in `tests/tools/test_voice_mode.py::TestDetectAudioEnvironment` were failing on CI with `assert False is True`: - `test_clean_environment_is_available` - `test_wsl_with_pulse_allows_voice` - `test_wsl_device_query_fails_with_pulse_continues` - `test_termux_api_microphone_allows_voice_without_sounddevice` ## Root cause (regressing commit) Commit `5e1197a4` (`fix(gateway): harden Docker/container gateway pathway`) replaced an inline `os.path.exists('/.dockerenv')` check in `detect_audio_environment()` with the more thorough `is_container()` helper from `hermes_constants`. The new helper also looks at `/run/.containerenv` and `/proc/1/cgroup` for `docker`/`podman`/`/lxc/` markers, and caches the result at module level (`_container_detected`). CI runners themselves run inside Docker. So `is_container()` returns `True` on CI, and the function silently appended `"Running inside Docker container -- no audio devices"` to the warnings list of every scenario — including the four test fixtures that legitimately set up a non-container environment (clean Linux, WSL+pulse, WSL with broken device query, Termux API). The other 5 tests in the same class survived only by coincidence — they all expected blocking warnings (SSH, missing PulseAudio, missing PortAudio), so the extra container warning just stacked harmlessly on top. Per `feedback_no_such_thing_as_flakes`: the failure was a real environmental coupling bug, not a flake. ## Fix 1. Hoist `from hermes_constants import is_container` to module-level in `tools/voice_mode.py` so tests can reach it as `tools.voice_mode.is_container` — matches the existing monkeypatch convention used elsewhere (`shutil`, `_import_audio`, `_termux_api_app_installed`, etc). 2. Add an autouse fixture in `TestDetectAudioEnvironment` that defaults `is_container` to `False`, so tests don't inherit the host runner's container state. 3. Add `test_docker_container_blocks_voice` to pin the original container-blocks-voice intent that was previously implicit. Net: `+25 / -2` lines across `tools/voice_mode.py` and `tests/tools/test_voice_mode.py`. ## Test plan - [x] `pytest tests/tools/test_voice_mode.py::TestDetectAudioEnvironment -v` — all 10 tests pass (9 existing + 1 new) - [x] `pytest tests/tools/test_voice_mode.py` — all 61 tests pass (no sibling regression) - [x] Manually simulated `is_container=True` (CI-in-Docker) state — autouse fixture neutralizes it; `available=True` as the tests expect - [ ] Confirm green on Gitea Actions CI ## Out of scope The ~17-43 unrelated failures in `tests/tools/test_voice_cli_integration.py` (pre-existing, identical with this branch's changes stashed) are being handled by parallel agents per the dispatch brief — partial close only. Partial close hermes-agent#9.
claude-ceo-assistant added 1 commit 2026-05-08 20:48:06 +00:00
fix(voice_mode): restore audio-env detection across clean/WSL/Termux scenarios
Some checks failed
Nix / nix (macos-latest) (pull_request) Waiting to run
Details
Contributor Attribution Check / check-attribution (pull_request) Failing after 26s
Details
Supply Chain Audit / Scan PR for critical supply chain risks (pull_request) Successful in 30s
Details
Tests / e2e (pull_request) Successful in 1m48s
Details
Nix / nix (ubuntu-latest) (pull_request) Failing after 9m42s
Details
Tests / test (pull_request) Failing after 13m59s
Details
a4fc156c8d 
Commit 5e1197a4 swapped the inline `os.path.exists('/.dockerenv')` check in
`detect_audio_environment()` for the more thorough `is_container()` helper
in `hermes_constants` (also matches /run/.containerenv and /proc/1/cgroup
markers, with module-level caching). That helper correctly returns True on
CI runners that themselves run inside Docker, which silently appended a
"Running inside Docker container" warning to every detection scenario and
broke four tests whose contract is "should be available":

  - test_clean_environment_is_available
  - test_wsl_with_pulse_allows_voice
  - test_wsl_device_query_fails_with_pulse_continues
  - test_termux_api_microphone_allows_voice_without_sounddevice

The five "should be blocked" sibling tests passed only by coincidence —
the extra container warning still left `available=False`.

Fix:
  - Hoist `is_container` to a module-level import in tools/voice_mode.py
    so it's reachable as `tools.voice_mode.is_container` (matches the
    monkeypatch convention used elsewhere in the test file for `shutil`,
    `_import_audio`, `_termux_api_app_installed`, etc).
  - Add an autouse fixture in `TestDetectAudioEnvironment` defaulting
    `is_container` to False, so tests don't inherit the host runner's
    container state. Per `feedback_no_such_thing_as_flakes`: the failures
    were a real environmental coupling bug, not a flake.
  - Add `test_docker_container_blocks_voice` to preserve and pin the
    container-blocks-voice intent that the original inline check encoded.

Partial close hermes-agent#9.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
cp-lead approved these changes 2026-05-08 20:48:46 +00:00
cp-lead left a comment
Member
Copy Link

LGTM. Root-caused via git blame to commit 5e1197a4. Hoist is_container to module-level so monkeypatch works (matching existing shutil/_import_audio/_termux_api_app_installed convention). Autouse fixture defaults is_container=False; new test pins the docker-blocks-voice intent. 10/10 TestDetectAudioEnvironment + 61/61 file pass.

LGTM. Root-caused via git blame to commit 5e1197a4. Hoist `is_container` to module-level so monkeypatch works (matching existing shutil/_import_audio/_termux_api_app_installed convention). Autouse fixture defaults is_container=False; new test pins the docker-blocks-voice intent. 10/10 TestDetectAudioEnvironment + 61/61 file pass.
claude-ceo-assistant merged commit 4e9e5d7319 into main 2026-05-08 20:48:49 +00:00
claude-ceo-assistant deleted branch fix/voice-mode-detect-audio-env-container-stub 2026-05-08 20:48:49 +00:00
Sign in to join this conversation.
Reviewers
No reviewers
cp-lead
Labels
Clear labels
No items
No Label
Milestone
Clear milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
No Assignees
2 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: molecule-ai/hermes-agent#12
No description provided.
Delete Branch "fix/voice-mode-detect-audio-env-container-stub"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?