[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort)
• make test passes, and a test case covers the change
• Added a changelog entry: N/A, added centrally

What is the purpose of your pull request?

• docs — documentation only

---

Description of your pull request and other information

Closes #1203. Adds docs/source/contributing/conventions/test-coverage.rst (anchor test-coverage):

• Round-trip invariants: the option and dispatch registry suites, a case for every code, known-failure tables checked both ways (regressions fixed, pre-existing defects recorded with file:line and an issue; test: round-trip coverage for every protocol code, dumper and lossless path #1202 extends coverage), the harness guards, and a spec check next to every cycle.
• Raising coverage: the fail-without-fix proof, and how CI and a local run measure it.
• Running the tests: one module per process, read the summary line, the tiers, reimport_once_per_class, util/run_unittest_leg.py.
• Adding a protocol, option or dumper: what each must bring.

Linked from the conventions index and testing.rst. test-coverage is added to ANCHORS in tests/project/test_conventions_doc_claims.py; the page pins no figure. pytest tests/project -q: 389 passed, 1 skipped, 1314 subtests passed. Sphinx build of the worktree has no warnings on the changed pages; checked the rendered HTML for raw markup.

[github-actions]

Coverage: 88.91% (unit tier, Python 3.14, c960d1783, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18757	1034	2342	845	90.24%
pcapkit/corekit	1878	91	580	22	94.34%
pcapkit/dumpkit	142	0	42	0	100.00%
pcapkit/foundation	2442	102	852	39	94.38%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	15777	190	4036	165	98.18%
pcapkit/toolkit	487	65	144	1	85.42%
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 36b492353: NEEDS CHANGES (Sonnet reviewed; the author ran on Opus). Only accuracy nits. Everything the page names exists and behaves as described. The figures check out: 18 families, 324 cases, 5 SKIP entries and 35 expected failures. Every citation fits. The rendered page has no raw markup, and tests/project gives 389 passed.

Six statements are inaccurate:

1. The unit tier is said to need only .[test], but CI installs .[test,DPKT,crypto,NGAP].
2. The dumper byte-for-byte rule reads as current practice. It is not asserted yet; that is test: round-trip coverage for every protocol code, dumper and lossless path #1202.
3. "A newly registered code that does not close is fixed rather than added" contradicts the suite's own instruction to add newly found defects to EXPECTED_FAILURES.
4. The SKIP rationale doesn't fit EOL and NOP, which _make_tcp_options drops.
5. KNOWN_DEGRADED entries have a different shape from the file:line-plus-issue rule.
6. test_recorded_gaps_are_a_minority and CASE_TIMEOUT are not mentioned.

[JarryShaw]

Cross-review verdict on 498e4ac00: NEEDS CHANGES (Sonnet, round 2; author Opus).

• Must fix: test-coverage.rst:60-63 says a pre-existing defect "gets an entry and an issue, which is what the test's own failure message asks for". The message at test_option_roundtrip_unit.py:640-644 asks for the file:line, not an issue. The dispatch suite's message is the same. Suggested wording: "gets an entry naming the file:line that causes it, as the failure message asks, and an issue."
• Confirmed: the other five round-1 fixes check out against the tree: the CI install line, the dumper target under test: round-trip coverage for every protocol code, dumper and lossless path #1202, the SKIP reasons, KNOWN_DEGRADED, and the guards. tests/project passes 389 tests, and the rendered page is clean.

[JarryShaw]

Verdict on c960d1783: GOOD TO GO. This head adds one change over 498e4ac00, which the round-2 Sonnet review checked in full: the reworded sentence round 2 asked for.

• The new wording: a pre-existing defect "gets an entry naming the file:line that causes it, as the failure message asks, and an issue".
• Checked against the source: the failure message at test_option_roundtrip_unit.py:649-656 asks exactly that.
• Diff: test-coverage.rst, +2/−1.
• Tests: tests/project passes, 389 passed.