test(corekit): field, schema and Info round trips (#1202)

[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort) — not run as make; only a pylint unused/undefined-name check on the new files (10.00/10)
• make test passes, and a test case covers the change — the whole suite was not run; the three new modules, tests/corekit, tests/project, test_tier_guard and test_docstring_contract pass
• Added a changelog entry — 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?

• test — tests only

---

Description of your pull request and other information

The corekit slice of #1202: field pack/unpack in both directions for every concrete corekit field, Schema.unpack(bytes(s)) re-packing and from_dict(to_dict()) for all schemas under pcapkit/protocols/schema/, and Info.from_dict(info.to_dict()) for all 487 Info classes, including repeated keys, __short_read__ and __truncated_raw__.

Module	Subtests	Time	Peak RSS
tests/corekit/test_field_roundtrip_unit.py	735	3.0 s	163 MiB
tests/corekit/test_schema_roundtrip_unit.py	883	4.2 s	144 MiB
tests/corekit/test_info_roundtrip_unit.py	1,436	4.5 s	175 MiB

Known failures, checked in both directions: #1487 (BitField unnamed bits), #1488 (NumberField bit_length narrower than its octets), #1489 (ListField without item_type), #1490 (non-final Info child leaking internal keys). Standalone ListFields are named by hand because of #1491.

Refs #1202

[JarryShaw]

Verdict: GOOD TO GO at c55de8a77. Cross-reviewed on Sonnet; the author ran on Opus. Test-only: four files under tests/corekit/.

• The known-failure table is enforced both ways. I re-checked one direction myself: deleting the fix(corekit): a ListField without item_type cannot pack what it unpacks #1489 row makes 5 subtests fail as unexpected. The reviewer checked the other:
  • Fix direction: a mutation standing in for a fix(corekit): a NumberField whose bit_length is narrower than its octets drops the high bits #1488 fix fails 3 tests, each asking for its row to be deleted.
  • New defect: an asymmetry added to BytesField.pack fails 6 tests as unexpected.
• The tests are not vacuous. They cover 431 of the 447 Schema classes as plain cases; the other 16 are 6 dispatchers (reached through 21 dispatch cases) and 10 fieldless bases. They also cover every concrete Field class and every annotated Info subclass, with UNBUILDABLE empty. Every caught exception becomes a named outcome that fails unless it is listed.
• Order independence. The tests define no Info or Schema subclass at module level. Running the new modules first or last gives the same result (118 passed, 3321 subtests). util/run_unittest_leg.py corekit gives 658 tests, 0 failures.
• Other checks: I re-ran the three modules myself: 20 passed, 3054 subtests. The tier and docstring guards pass, and no capture files are read.

Non-blocking: the #1487 and #1488 rows have an empty failure fragment, so they match on status alone. Their labels are narrow enough for that to be safe.

[github-actions]

Coverage: 89.65% (unit tier, Python 3.14, c55de8a77, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18795	958	2356	871	90.99%
pcapkit/corekit	2107	68	672	29	96.15%
pcapkit/dumpkit	242	2	88	2	98.79%
pcapkit/foundation	2648	98	930	48	94.80%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	17086	238	4628	190	97.98%
pcapkit/toolkit	539	74	168	3	84.58%
pcapkit/utilities	431	4	124	4	98.56%
pcapkit/vendor	4409	2342	1006	157	43.25%

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