Fix resource leaks when TeleopSession setup fails

[fallintoplace]

Description

TeleopSession.__enter__ acquired OpenXR, DeviceIO, and plugin contexts onto the session-owned ExitStack. If a later setup step raised, the with body never started and Python did not call TeleopSession.__exit__, leaving the resources acquired so far alive.

This change makes session startup transactional:

• acquire resources on a temporary ExitStack
• on failure, unwind the temporary stack in reverse order and clear the stale OpenXR reference
• keep propagating the setup error even if a child suppresses it or raises during rollback; rollback failures are logged
• transfer stack ownership to the session only after every setup step succeeds
• reject nested entry so an active session stack cannot be overwritten

There is no public API change.

The parameterized regression test fails OpenXR, DeviceIO, the first plugin, and a later plugin in turn, checking exact-once reverse-order cleanup. Additional tests cover rollback cleanup failure and nested entry while preserving normal reuse after exit.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Testing

• ctest --test-dir build -R "^teleop_session_manager_" --output-on-failure --parallel 4 (4/4 passed, macOS arm64, Release, Python 3.11)
• SKIP=check-copyright-year pre-commit run --all-files

Checklist

• I have read and understood the contribution guidelines
• I have run the linter and formatter with SKIP=check-copyright-year pre-commit run --all-files
• Documentation is not required because the public API is unchanged
• I have added tests that prove the fix works
• I have signed off all commits (git commit -s) per the DCO

[github-actions]

📝 Docs preview is not auto-deployed for fork PRs.

A maintainer with write access to NVIDIA/IsaacTeleop can deploy a preview by
commenting /preview-docs on this PR. Once deployed, the preview
will live at:

https://nvidia.github.io/IsaacTeleop/preview/pr-782/

[coderabbitai]

📝 Walkthrough

Walkthrough

TeleopSession.__enter__ now acquires OpenXR, DeviceIO, and plugin resources through a temporary ExitStack. Failed acquisition closes the partially populated stack, clears the OpenXR session, and re-raises the exception. Successful acquisition stores the stack and marks setup complete. A parametrized test verifies reverse-order cleanup and state reset across multiple failure stages.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

• NVIDIA/IsaacTeleop#770: Both changes update TeleopSession and OpenXR session lifecycle cleanup around context-manager teardown.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
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.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly matches the main change: transactional TeleopSession setup to prevent resource leaks on failure.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 2

🤖 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 `@src/core/teleop_session_manager/python/teleop_session.py`:
- Around line 926-932: Update the exception-handling block in the teleop session
setup flow to catch failures from stack.__exit__, log the rollback error, and
explicitly re-raise the original error so cleanup cannot replace it; keep
self._oxr_session reset in the finally path. Add a regression test covering an
already-entered resource whose __exit__ fails during unwind and verify the setup
exception is preserved.
- Around line 922-936: Update the session context-manager entry flow around
_enter_resources to reject a nested __enter__ when the session is already
active, checking _setup_complete and/or the existing exit stack before mutating
state. Raise the appropriate re-entry error while preserving the current
setup-error cleanup behavior for the initial entry.

🪄 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: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: ac60bea7-30af-41eb-a241-d17d51498a62

📥 Commits

Reviewing files that changed from the base of the PR and between 2a67ed4 and 7321e7d.

📒 Files selected for processing (2)

• src/core/teleop_session_manager/python/teleop_session.py
• src/core/teleop_session_manager_tests/python/test_teleop_session.py