fix: reporter context retrieval

[ruben-arts]

Description

Assuming this might fix #5200

How Has This Been Tested?

I've been testing this with the https://github.com/ruben-arts/ros_workspace as that spawns multiple backends, and it was the only project I could get the error with. I've not been able to recreate the race condition, but they are tricky so this is not a promised fix.

I've extended the stress test by making it multi-env, with pypi sdist and solving for multiple platforms:

[workspace]
channels = [
    "https://prefix.dev/pixi-build-backends",
    "https://prefix.dev/conda-forge",
    "https://prefix.dev/robostack-humble",
]
platforms = ["osx-arm64", "linux-64", "win-64", "linux-aarch64"]
preview = ["pixi-build"]


[dependencies]
ros-humble-ros-core = ">=0.10.0,<0.11"
ros-humble-turtlesim = "*"
ros-humble-navigator = { path = "src/navigator/package.xml" }
ros-humble-navigator-py = { path = "src/navigator_py/package.xml" }
ros-humble-talker-py = { path = "src/talker-py/package.xml" }

[dev]
ros-humble-navigator = { path = "src/navigator/package.xml" }
ros-humble-navigator-py = { path = "src/navigator_py/package.xml" }
ros-humble-talker-py = { path = "src/talker-py/package.xml" }

[pypi-dependencies]
sdist = "*"

[feature.test.dependencies]
pytest = "*"

[environments]
test = ["test"]

I can't seem to break it anymore, but it was hard in the first place, so 🤞

AI Disclosure

• This PR contains AI-generated content.
  • I have tested any AI-generated content in my PR.
  • I take responsibility for any AI-generated content in my PR.
    Tools: Claude

Checklist:

• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have added sufficient tests to cover my changes.

This type of issue is almost impossible to recreate in a test as it's flaky behavior.

[baszalmstra]

Yes this doesnt seem right. Instead we should just cancel the child task as well because most likely the parent errored or was itself cancelled.

[nichmor]

Yes this doesnt seem right. Instead we should just cancel the child task as well because most likely the parent errored or was itself cancelled.

I've added a child cancellation token ( from parent token ).

[baszalmstra]

@nichmor When you have time can you point me to how this now fixes the issue. I dont understand why this would fix the issue. From my perspective the tasks were already cancelled anyway? The child cancellation tokens are also never cancelled manually are they?