Break cyclic references at connection close when there was a traceback

[bdraco]

What do these changes do?

Clears the exception on the DataQueue and WebSocketDataQueue when the connection is closed to break cyclic references.

Are there changes in behavior for the user?

fixes a potential memory leak

Is it a substantial burden for the maintainers to support this?

no

Related issue number

fixes #10535

Checklist

• I think the code is well written
• Unit tests for the changes exist
• Documentation reflects the changes
• If you provide code modification, please add yourself to CONTRIBUTORS.txt
  • The format is <Name> <Surname>.
  • Please keep alphabetical order, the file is sorted by names.
• Add a new news fragment into the CHANGES/ folder

  • name it <issue_or_pr_num>.<type>.rst (e.g. 588.bugfix.rst)

  • if you don't have an issue number, change it to the pull request
    number after creating the PR

    • .bugfix: A bug fix for something the maintainers deemed an
      improper undesired behavior that got corrected to match
      pre-agreed expectations.
    • .feature: A new behavior, public APIs. That sort of stuff.
    • .deprecation: A declaration of future API removals and breaking
      changes in behavior.
    • .breaking: When something public is removed in a breaking way.
      Could be deprecated in an earlier release.
    • .doc: Notable updates to the documentation structure or build
      process.
    • .packaging: Notes for downstreams about unobvious side effects
      and tooling. Changes in the test invocation considerations and
      runtime assumptions.
    • .contrib: Stuff that affects the contributor experience. e.g.
      Running tests, building the docs, setting up the development
      environment.
    • .misc: Changes that are hard to assign to any of the above
      categories.

  • Make sure to use full sentences with correct case and punctuation,
    for example:

    Fixed issue with non-ascii contents in doctest text files
    -- by :user:`contributor-gh-handle`.

    Use the past tense or the present tense a non-imperative mood,
    referring to what's changed compared to the last released version
    of this project.

[bdraco]

I'm not sure this is a true memory leak as I expect them to eventually get cleaned up.. but maybe not since the reproducer does gc.collect() and they are still there

edit: #10535 (comment) seems to indicate its a real leak since after del , objgraph can still find the object by gc

[codspeed-hq]

CodSpeed Performance Report

Merging #10556 will not alter performance

_{Comparing break_cyclic_refs (2107c72) with master (d067260)}

Summary

✅ 47 untouched benchmarks

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.70%. Comparing base (d067260) to head (2107c72).
Report is 1 commits behind head on master.

✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #10556      +/-   ##
==========================================
- Coverage   98.70%   98.70%   -0.01%     
==========================================
  Files         122      124       +2     
  Lines       37290    37335      +45     
  Branches     2063     2063              
==========================================
+ Hits        36808    36850      +42     
- Misses        335      338       +3     
  Partials      147      147              

Flag	Coverage Δ
CI-GHA	98.58% <100.00%> (-0.01%)	⬇️
OS-Linux	98.25% <100.00%> (-0.01%)	⬇️
OS-Windows	96.18% <97.77%> (+<0.01%)	⬆️
OS-macOS	97.36% <97.77%> (-0.02%)	⬇️
Py-3.10.11	97.27% <97.77%> (+<0.01%)	⬆️
Py-3.10.16	97.81% <100.00%> (+<0.01%)	⬆️
Py-3.11.11	97.90% <100.00%> (-0.01%)	⬇️
Py-3.11.9	97.35% <97.77%> (-0.01%)	⬇️
Py-3.12.9	98.35% <100.00%> (+<0.01%)	⬆️
Py-3.13.2	98.34% <100.00%> (+<0.01%)	⬆️
Py-3.9.13	97.15% <97.77%> (+<0.01%)	⬆️
Py-3.9.21	97.69% <100.00%> (+<0.01%)	⬆️
Py-pypy7.3.16	87.01% <76.92%> (-4.21%)	⬇️
VM-macos	97.36% <97.77%> (-0.02%)	⬇️
VM-ubuntu	98.25% <100.00%> (-0.01%)	⬇️
VM-windows	96.18% <97.77%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bdraco]

Writing a test for this is proving difficult, because we need complete isolation to do the GC test

[Dreamsorcerer]

I haven't really looked at the implementation, but does it make sense to put some weakrefs in, rather than trying to break the cyclic references?

[bdraco]

I haven't really looked at the implementation, but does it make sense to put some weakrefs in, rather than trying to break the cyclic references?

I think setting to None probably makes the most sense here.

It seems to be the prescribed solution for asyncio as well

Lib/asyncio/base_events.py:        # Needed to break cycles when an exception occurs.
Lib/asyncio/base_events.py:        handle = None  # Needed to break cycles when an exception occurs.
Lib/asyncio/events.py:        self = None  # Needed to break cycles when an exception occurs.
Lib/asyncio/selector_events.py:            # Needed to break cycles when an exception occurs.
Lib/asyncio/tasks.py:                    self = None  # Needed to break cycles when an exception occurs.
Lib/asyncio/tasks.py:            self = None  # Needed to break cycles when an exception occurs.
Lib/asyncio/tasks.py:            self = None  # Needed to break cycles when an exception occurs.
Lib/asyncio/tasks.py:        self = None  # Needed to break cycles when an exception occurs.
Lib/asyncio/windows_events.py:            # Needed to break cycles when an exception occurs

[patchback]

Backport to 3.11: 💚 backport PR created

✅ Backport PR branch: patchback/backports/3.11/9d4e11617a1e596926d11e483280183224fc7ee6/pr-10556

Backported as #10566

🤖 @patchback
I'm built with octomachinery and
my source is open — https://github.com/sanitizers/patchback-github-app.

[patchback]

Backport to 3.12: 💚 backport PR created

✅ Backport PR branch: patchback/backports/3.12/9d4e11617a1e596926d11e483280183224fc7ee6/pr-10556

Backported as #10567

🤖 @patchback
I'm built with octomachinery and
my source is open — https://github.com/sanitizers/patchback-github-app.

[bdraco]

On a side note, I still haven't had luck fixing #10548.

I'm going to build an xfail test for it using some of the tooling I just built here. Maybe a weakref will work there. Hopefully I'll figure it out soon