docs(tests): tighten the shared test-support prose from the post-sweep merges (#719)

[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; prose-only, AST-identical

• make test passes, and a test case covers the change

• 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?

• 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
• chore — anything else

---

Description of your pull request and other information

Final #719 pass over the prose ef0964c81..main added to tests/ outside protocols/, foundation/, corekit/, project/. Comments and docstrings only.

• reimport_once_per_class: the unittest "import left in place" claim now excludes restore=True, which restores the region at class end; timed wording cut.
• time_limit / conftest.pytest_timeout_cancel_timer: CI-history sentences cut; the measurements that justify them stay.
• test_compat: "purged and restored around every test" and "registered before anything else" were stale after the per-class switch.
• test_const_enum_lookup: two classes purge in setUpClass, not setUp.
• _DUNDER_SAMPLE_SIZE is a stride divisor (~50 members of TCP/UDP, all of SCTP/DCCP), not a stride.
• EndToEndTestCase.setUpClass: the claim that restore_module_table snapshots right after it was false. Under pytest the fixture restores the warm session pin before every test, so the purge and re-import never reach the tests (measured by id(sys.modules['pcapkit'])); the once-per-class rationale is now scoped to plain unittest. _pin_module_snapshot's opening now states the current contract, and pytest_sessionstart no longer claims the reload saves ~45s or that a reimporting class keeps its import.
• "used to" / "since tests: stop re-importing pcapkit before every test (purge_modules dominates suite time) #1065" / "merged from" history cut in _helpers, builtin_parity, no_mint, registry_protocol.

Numstat: 9 files, +81/-96. Tests: each touched module, every EndToEndTestCase module, in its own pytest process, plus tests/project (375 passed, 1 skipped).

file (AST without docstrings vs origin/main)	equal
tests/_support.py	True
tests/conftest.py	True
tests/const/test_const_apptype_split_unit.py	True
tests/const/test_const_enum_builtin_parity.py	True
tests/const/test_const_enum_lookup.py	True
tests/const/test_const_enum_no_mint.py	True
tests/const/test_const_registry_protocol.py	True
tests/integration/_helpers.py	True
tests/utilities/test_compat.py	True

[github-actions]

Coverage: 88.74% (unit tier, Python 3.14, f9ca0e627, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18757	1035	2342	846	90.23%
pcapkit/corekit	1874	91	578	22	94.33%
pcapkit/dumpkit	136	0	40	0	100.00%
pcapkit/foundation	2422	143	842	34	92.62%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	15653	187	3942	162	98.19%
pcapkit/toolkit	487	71	144	3	84.15%
pcapkit/utilities	429	4	122	4	98.55%
pcapkit/vendor	4409	2359	1006	158	42.84%

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

[JarryShaw]

Cross-review verdict on f9ca0e627: GOOD TO GO (ran on Sonnet; author Opus)

• No code changed: all 9 files are AST-identical to origin/main, and every hunk is a comment or docstring.
• EndToEndTestCase.setUpClass behaves as the new docstring says, re-measured on both runners:
  • under pytest, tests see the pcapkit object from before setUpClass, which is the session pin;
  • under unittest, they see the one setUpClass imported.
• pytest_sessionstart: it imports pcapkit and then pins, so the pin is warm (329 modules). A module held by a direct reference survives a restore.
• reimport_once_per_class with restore=True vs default: confirmed by probe.
• Sampling counts in tests/const:
  • TCP and UDP use a stride divisor of 128 and 127, giving about 50 evenly spaced members plus the aliased ones;
  • SCTP takes all 91 members and DCCP all 10;
  • neither SCTP nor DCCP has aliases.
• The two setUpClass purgers in test_const_enum_lookup: confirmed.
• The dropped 45s and 0.7s/92/28 figures were unfounded: the saving no longer exists, and the tree has 25 classes and 83 tests.
• Touched modules pass, plus two integration modules.
• Nit: tests/conftest.py:75 is 122 characters.
• Not checked by the reviewer: tests/test_support_helpers.py, tests/project and test_const_enum_no_mint.py. The author's runs gave 32 passed and 375 passed, 1 skipped, and CI is green.