test: speed up slow loopback tests (closes #1697) by bluetoothbot · Pull Request #1699 · python-zeroconf/python-zeroconf

bluetoothbot · 2026-05-17T12:20:31Z

What

Cut ~28s from the slowest-20 duration sum (70s → 42s) without changing
what any test asserts. All 336 tests still pass.

Why

Issue #1697 — the slow-test report listed twenty tests taking 2–10s
each, dominated by full mDNS probe/announce/goodbye cycles on
loopback. The project already ships a quick_timing fixture in
tests/conftest.py for this exact case (RFC 6762 timings don't need
to apply when talking to 127.0.0.1) — it just wasn't applied
everywhere.

How

Apply quick_timing to register/announce-heavy tests in
tests/test_asyncio.py, tests/test_services.py,
tests/test_handlers.py, and tests/test_init.py. For
unittest.TestCase-based tests, wired in via
@pytest.mark.usefixtures("quick_timing").
Drop ZeroconfServiceTypes.find timeouts from 2s to 0.5s in
tests/services/test_types.py and test_async_zeroconf_service_types
— find() sleeps the full timeout, and on loopback the response
arrives in milliseconds.
Tighten the explicit TOCTOU wait in
test_service_info_async_request from 3000ms to 500ms;
_is_complete=False forces the request to run out the full timeout
regardless of network latency.
Cut the "allow multicast timers to expire" sleep in
test_integration_with_listener_class from 3s to 0.5s now that
quick_timing makes the preceding broadcasts settle in <100ms.
Shorten the duplicate-update negative-assertion wait in
TestServiceBrowser.test_update_record from wait_time=3 to 0.3s
— the listener is asserted NOT to fire, so the wait always times
out.

Tests left untouched include test_async_updates_from_response
(needs the RFC 6762 §10.2 one-second invalidation window),
test_get_info_suppressed_by_question_history,
test_we_try_four_times_with_random_delay, and the
test_response_aggregation_timings* pair — those measure timing
behaviour and the real delays are the assertion.

Testing

poetry run pytest --durations=20 --timeout=60 — 336 passed,
2 skipped, 83.5s (was 87.4s)
Verified each modified test individually still passes and shows a
drop on --durations.
ruff check + ruff format --check clean.

Per-test before/after on the slowest entries from #1697:

Test	Before	After
`test_service_info_async_request`	9.83s	3.98s
`test_integration_with_listener_class`	6.83s	1.84s
`test_async_zeroconf_service_types`	4.49s	1.47s
`test_integration_with_listener`	4.28s	1.28s
`test_integration_with_listener_v6_records`	4.28s	1.28s
`test_integration_with_subtype_and_listener`	4.28s	1.27s
`test_update_record` (browser)	3.22s	0.54s
`test_register_and_lookup_type_by_uppercase_name`	2.88s	1.25s
`test_async_service_registration`	2.87s	0.37s
`test_verify_name_change_with_lots_of_names`	2.71s	0.70s
`test_async_service_browser`	2.40s	0.30s
`test_async_tasks`	2.39s	0.26s
`test_async_service_registration_with_server_missing`	2.37s	0.26s
`test_async_service_registration_name_conflict`	2.34s	0.70s
`test_name_conflicts`	2.20s	0.52s

Closes #1697

Quality Report

Changes: 6 files changed, 27 insertions(+), 19 deletions(-)

Code scan: clean

Tests: passed (4 PASSED)

Branch hygiene: clean

Generated by Kōan post-mission quality pipeline

…er timeouts Cuts ~28s from the top-20 durations sum (was 70s, now 42s) without changing what the tests assert. - Add quick_timing fixture (introduced for the same purpose in conftest) to register/announce-heavy tests in test_asyncio.py, test_services.py, test_handlers.py, and test_init.py. The fixture shortens _CHECK_TIME/_REGISTER_TIME/_UNREGISTER_TIME to 10ms on loopback so the probe/announce/goodbye cycle no longer pays the RFC 6762 production budget. - Drop ZeroconfServiceTypes.find timeouts from 2s to 0.5s in test_types.py and test_async_zeroconf_service_types — find() sleeps the full timeout, and on loopback the response arrives in milliseconds. - Tighten the explicit TOCTOU wait in test_service_info_async_request from 3000ms to 500ms; _is_complete=False forces the request to run out the full timeout regardless. - Cut the 'allow multicast timers to expire' sleep in test_integration_with_listener_class from 3s to 0.5s now that quick_timing makes the preceding broadcasts settle in <100ms. - Shorten the duplicate-update negative-assertion wait in TestServiceBrowser.test_update_record from wait_time=3 to 0.3s — the listener is asserted NOT to fire, so the wait always times out.

codspeed-hq · 2026-05-17T12:22:25Z

Merging this PR will not alter performance

✅ 6 untouched benchmarks

_{Comparing bluetoothbot:koan/speed-up-slow-tests (36f20b0) with master (dfa4e00)}

codecov · 2026-05-17T12:22:39Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.76%. Comparing base (dfa4e00) to head (36f20b0).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1699   +/-   ##
=======================================
  Coverage   99.76%   99.76%           
=======================================
  Files          33       33           
  Lines        3410     3410           
  Branches      464      464           
=======================================
  Hits         3402     3402           
  Misses          5        5           
  Partials        3        3

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

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

bdraco · 2026-05-17T13:55:50Z

@bluetoothbot rr

bluetoothbot · 2026-05-17T14:02:51Z

PR Review — test: speed up slow loopback tests (closes #1697)

Clean, well-scoped speed-up that exploits an existing fixture (quick_timing) rather than introducing new test infrastructure. Each change is justified in the PR body with before/after numbers, the untouched timing-sensitive tests (test_async_updates_from_response, test_response_aggregation_timings*, test_we_try_four_times_with_random_delay) are correctly excluded, and the one place where the new constant is non-obvious (negative-assertion wait in test_update_record) has an explanatory comment per project convention. No production code changes, coverage unchanged, CodSpeed unchanged. Only flag is potential CI flakiness on PyPy / 3.14t for the ZeroconfServiceTypes.find() 0.5s timeouts — that's a watch item, not a blocker. Merge-ready.

🟢 Suggestions

1. Watch for flakiness on slow CI matrix entries (`tests/services/test_types.py`, L50-53)

These tests register via zeroconf_registrar.registry.async_add(info) (direct registry add, no broadcast), so the announce path isn't the variable here — the only thing that has to happen inside the 0.5s window is the find() query and its response. On loopback that's milliseconds, so 0.5s is a 100x+ margin and almost certainly safe. The CI matrix runs PyPy 3.10 and the free-threaded 3.14t build though, which can be 5–10x slower under load. If you see flakes on those matrix entries after this lands, the cheap fix is to bump just these four find() calls to 1s rather than reverting — that's still 2x faster than the original and gives PyPy headroom. Not a blocker; flagging because none of the other tightened tests use find().

service_types = ZeroconfServiceTypes.find(interfaces=["127.0.0.1"], timeout=0.5)

2. Consider a one-liner comment for the magic number (`tests/test_services.py`, L117)

The existing comment on line 116 (# short pause to allow multicast timers to expire) is now misleading — 0.5s is no longer a "short pause" in absolute terms relative to what's being waited on, it's deliberately calibrated against the quick_timing budget (preceding broadcasts settle in <100ms, so 5x margin). The PR description captures this rationale, but a future reader looking only at this file won't see it. Per CLAUDE.md ("Don't remove existing comments unless the code they describe is gone") the existing comment should stay, but consider tightening it to something like # allow multicast timers to expire — quick_timing settles broadcasts in <100ms. Same story as your test_browser.py comment, which does the right thing.

            # short pause to allow multicast timers to expire
            time.sleep(0.5)

Checklist

Summary

Clean, well-scoped speed-up that exploits an existing fixture (quick_timing) rather than introducing new test infrastructure. Each change is justified in the PR body with before/after numbers, the untouched timing-sensitive tests (test_async_updates_from_response, test_response_aggregation_timings*, test_we_try_four_times_with_random_delay) are correctly excluded, and the one place where the new constant is non-obvious (negative-assertion wait in test_update_record) has an explanatory comment per project convention. No production code changes, coverage unchanged, CodSpeed unchanged. Only flag is potential CI flakiness on PyPy / 3.14t for the ZeroconfServiceTypes.find() 0.5s timeouts — that's a watch item, not a blocker. Merge-ready.

Automated review by Kōan36f20b0

bluetoothbot · 2026-05-17T14:04:04Z

Rebase with requested adjustments

Branch koan/speed-up-slow-tests was rebased onto master and review feedback was applied.

Stats

6 files changed, 27 insertions(+), 19 deletions(-)

Actions performed

Already-solved check: negative (confidence=high, reasoning=Commit d5e1f01 added quick_timing fixture and applied it to some tests, but this PR extends applicat)
Rebased koan/speed-up-slow-tests onto upstream/master
Pre-push CI check: previous run passed
Force-pushed koan/speed-up-slow-tests to origin
CI check enqueued in ## CI (async)

CI status

CI will be checked asynchronously.

Automated by Kōan

bdraco · 2026-05-17T14:14:17Z

@bluetoothbot rebase

bdraco · 2026-05-17T14:33:40Z

@bluetoothbot rr

bluetoothbot · 2026-05-17T14:56:21Z

Rebase with requested adjustments

Branch koan/speed-up-slow-tests was rebased onto master and review feedback was applied.

Stats

6 files changed, 27 insertions(+), 19 deletions(-)

Actions performed

Already-solved check: negative (confidence=high, reasoning=Master commit d5e1f01 added quick_timing fixture and applied it to some tests, but this PR extends a)
Rebased koan/speed-up-slow-tests onto upstream/master
Pre-push CI check: previous run passed
Force-pushed koan/speed-up-slow-tests to origin
CI check enqueued in ## CI (async)

CI status

CI will be checked asynchronously.

Automated by Kōan

bluetoothbot temporarily deployed to release May 17, 2026 12:25 — with GitHub Actions Inactive

bdraco marked this pull request as ready for review May 17, 2026 15:02

bdraco approved these changes May 17, 2026

View reviewed changes

bdraco merged commit dd341a3 into python-zeroconf:master May 17, 2026
37 checks passed

bluetoothbot deleted the koan/speed-up-slow-tests branch May 17, 2026 15:40

bdraco mentioned this pull request May 17, 2026

test: drop pending multicast responses before TOCTOU assertion #1701

Merged

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test: speed up slow loopback tests (closes #1697)#1699

test: speed up slow loopback tests (closes #1697)#1699
bdraco merged 1 commit into
python-zeroconf:masterfrom
bluetoothbot:koan/speed-up-slow-tests

bluetoothbot commented May 17, 2026 •

edited

Loading

Uh oh!

codspeed-hq Bot commented May 17, 2026

Uh oh!

codecov Bot commented May 17, 2026 •

edited

Loading

Uh oh!

bdraco commented May 17, 2026

Uh oh!

bluetoothbot commented May 17, 2026

Uh oh!

bluetoothbot commented May 17, 2026

Uh oh!

bdraco commented May 17, 2026

Uh oh!

bdraco commented May 17, 2026

Uh oh!

bluetoothbot commented May 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

bluetoothbot commented May 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Why

How

Testing

Quality Report

Uh oh!

codspeed-hq Bot commented May 17, 2026

Merging this PR will not alter performance

Uh oh!

codecov Bot commented May 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

bdraco commented May 17, 2026

Uh oh!

bluetoothbot commented May 17, 2026

PR Review — test: speed up slow loopback tests (closes #1697)

🟢 Suggestions

Checklist

Summary

Uh oh!

bluetoothbot commented May 17, 2026

Rebase with requested adjustments

Stats

CI status

Uh oh!

bdraco commented May 17, 2026

Uh oh!

bdraco commented May 17, 2026

Uh oh!

bluetoothbot commented May 17, 2026

Rebase with requested adjustments

Stats

CI status

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

bluetoothbot commented May 17, 2026 •

edited

Loading

codecov Bot commented May 17, 2026 •

edited

Loading