[JarryShaw]

Please follow the guide below

• You will be asked some questions, please read them carefully and answer honestly

• Put an x into all the boxes [ ] relevant to your pull request (like that [x])

• Use Preview tab to see how your pull request will actually look like

• Searched for similar pull requests

• Followed the coding style (make pylint, make mypy; make isort not run)

• make test passes, and a test case covers the change (only tests/foundation/test_extraction_ownership.py was run locally; CI covers the rest)

• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible — N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

What is the purpose of your pull request?

Tick the commit type your subject line carries.

• fix — corrects a defect
• feat — adds a feature
• perf — changes performance, not behaviour
• refactor — changes neither behaviour nor performance
• test — tests only
• docs — documentation only
• ci — workflows or build tooling
• release — bumps the version or rolls up a distribution
• chore — anything else

---

Description of your pull request and other information

Closes #2016.

collect_until itself is sound: it checks the predicate before the deadline. The trouble was the test, which ran the engine's real gc.collect against a 5 s wall-clock grace. It now patches gc.collect, time.monotonic and time.sleep with fakes, each collection costing one fake second. It also covers two edges that were untested: a predicate that holds with no grace (CPython's case), and one that never holds and stops once the grace is spent.

Evidence, with gc.collect patched to sleep 3 s: the old test fails with False is not true, matching the GraalPy leg, and the new one passes in 0.05 s. The new test also fails against a mutant that checks the deadline first and against one that ignores the grace. FINALISER_GRACE is left untouched, so this does not conflict with #2006.

[JarryShaw]

GOOD TO GO at 18d3284b8.

A Sonnet cross-review:

• confirmed the fakes patch gc.collect, time.monotonic and time.sleep at the names collect_until actually calls, and undo them when the with block exits;
• slowed gc.collect to 3 s: the old test failed after 9.27 s, the new one passed in 0.05 s;
• swapped in a collect_until that checks the deadline before the predicate, and the new no-grace case caught it;
• found the FINALISER_GRACE line untouched;
• got 19 passed on the PR head.

GraalPy itself was not run; the graalpy-25.4 leg will be the first real check.

[github-actions]

Coverage: 91.55% (unit tier, Python 3.15, 18d3284b8, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	2	95.16%
pcapkit/const	19597	881	2618	867	92.11%
pcapkit/corekit	2542	71	810	37	96.54%
pcapkit/dumpkit	276	3	100	3	98.40%
pcapkit/foundation	2930	23	1022	25	98.79%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17501	222	4612	187	98.13%
pcapkit/toolkit	872	36	298	10	95.21%
pcapkit/utilities	523	2	158	4	99.12%
pcapkit/vendor	4431	1995	1018	167	51.73%

Per-file detail: the coverage-html artifact of this run.