fix: shorten the duplicate packet window to 800 ms - #1930

PR Review — fix: shorten the duplicate packet window to 800 ms

Merge-ready: the change is small and does what the description says, with one comment nit.

Strengths:

• Both dedup paths in _process_datagram_at_time (the last_time check and _recent_packets) read the same constant, so a single-value change moves both consistently.
• Keeping _DUPLICATE_QUESTION_INTERVAL at 999 is correct. _history.py and the QM requery spacing in info.py:377-383 use it independently of the packet window, so dropping the old "1ms less" comment loses no enforced constraint.
• The new test would fail with the old 1000 ms value, and it also pins that a 5 ms repeat is still dropped. The new constant comment cites RFC 6762 §8.3.
• Responder rate limiting still holds. Repeated QM queries at 800-999 ms now reach async_response, but _has_mcast_record_in_last_second sends recently multicast records to the delayed queue (_core.py:201).

Needs attention:

• Stale "~1s" wording remains in _listener.py:254 and tests/test_listener.py:392 (non-blocking).

Accepted trade-offs:

• Under a sustained flood, the cap rises from one parse per payload per 1000 ms to one per 800 ms.
• A peer's repeat at 800-999 ms now refreshes the question history for questions we can answer. Our own identical query can therefore stay suppressed slightly longer.

Not verified: the shell for this review blocked Python, so I did not run the tests. The test and bench results in the description are the author's claims.

Checklist

• No hardcoded secrets
• Input validation at boundaries
• Timing change cites the relevant RFC section
• New behaviour covered by a test that fails on the old value
• Duplicate suppression stays bounded under flood
• No Cython .pxd changes required

Automated review by Kōan (Claude) HEAD=047dbc4 2 min 4s

Thanks for contributing! A quick note: this project is in the process of moving from LGPL-2.1 to Apache 2.0, and may keep LGPL-2.1-or-later as an OR option so GPL-2.0-only projects can keep using the library, see #1835.

By continuing with this pull request you agree that this new contribution is offered under both Apache 2.0 and LGPL-2.1-or-later, so it is covered under either outcome.

This notice does not affect your previous contributions. If you would like to consent to relicensing those as well, please post the consent statement on #1835; silence is never treated as consent for past work.

If offering this contribution under these licenses does not work for you, no hard feelings; reply with the exact phrase I do not agree to the Apache 2.0 license or simply close this pull request.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.87%. Comparing base (bd107c6) to head (f267a03).
⚠️ Report is 3 commits behind head on master.

@@           Coverage Diff           @@
##           master    #1930   +/-   ##
=======================================
  Coverage   99.87%   99.87%           
=======================================
  Files          33       33           
  Lines        3913     3913           
  Branches      568      568           
=======================================
  Hits         3908     3908           
  Misses          3        3           
  Partials        2        2           

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

Merging this PR will not alter performance

✅ 24 untouched benchmarks

_{Comparing shorter_duplicate_window (f267a03) with master (bd107c6)}

Fixed the stale "~1s" wording in _listener.py and tests/test_listener.py so the comments refer to the suppression interval rather than a fixed second.