[timurzvural]

Fixes #160986.

test_popping_python_frame_is_not_native created RemoteUnwinder as soon as the child process could be read, then took a fixed 10,000 samples. On a slow startup those samples all landed during interpreter startup, before leaf() existed, so the test observed zero frames and failed. It now waits until leaf is on the stack, then takes the sample window. The check that a popping leaf frame is never reported as <native> is unchanged.

Per review feedback, this PR only changes the test (no NEWS entry).

This pull request was prepared with Cursor Agent.

• Issue: test_popping_python_frame_is_not_native fails randomly on buildbots #160986

[python-cla-bot]

All commit authors signed the Contributor License Agreement.

[bedevere-app]

Thank you for your interest in contributing to Python. This is a community-driven project, and we want it to be a welcoming place to learn and collaborate.

Contributors are responsible for reviewing and understanding what they submit, checking its claims, and being able to explain changes in their own words. This helps reviewers focus on the contribution and leaves more time to support other contributors.

This contribution appears to have been authored with AI assistance. Please explain your contribution in your own words and describe how you checked its correctness.

Please read our guidelines for using AI tools.

[NekoAsakura]

Thanks! Please keep the change minimal and focused, and it doesn't need a NEWS entry.

[timurzvural]

Thanks — updated.

I removed the NEWS entry and left only the test change in Lib/test/test_external_inspection.py.

What was wrong: The test attached RemoteUnwinder as soon as the child PID was readable, then immediately took 10,000 samples. On a slow machine/buildbot, all of those samples could land during interpreter startup, before leaf() was on the stack. The later assertions then saw zero leaf frames and failed intermittently.

What the test change does: After attaching, it polls get_stack_trace() until some frame's funcname is "leaf". Only then does it run the existing 10,000-sample loop and the same check that a popping leaf frame is never labeled <native>.

How I checked it: Compared against the race described in #160986: the failure mode was "samples finish before leaf exists," so delaying the sample window until leaf appears removes that race without changing what the test asserts about native vs Python frames.

[bedevere-app]

Thank you for your interest in contributing to Python. This is a community-driven project, and we want it to be a welcoming place to learn and collaborate.

Contributors are responsible for reviewing and understanding what they submit, checking its claims, and being able to explain changes in their own words. This helps reviewers focus on the contribution and leaves more time to support other contributors.

We believe this contribution does not meet our guidelines for using AI tools, so we're closing it. If you think we've misunderstood your contribution, please let us know. You're welcome to explain your approach or ask questions about the guidelines.

Please read our guidelines for using AI tools.