fix(corekit): finalise each non-final Info subclass on its own state (#1490)

[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)

• 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?

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

Info.__new__ now reads __finalised__ from the class's own __dict__ (pcapkit/corekit/infoclass.py:424), so a non-final child is finalised on its first construction whatever its parent's state. Before, it inherited the parent's BASE, kept the __excluded__ copied at declaration, and leaked __map__, __map_reverse__, __multi__ from to_dict(), len() and iteration.

• New tests/corekit/test_info_child_finalisation_1490_unit.py: parent-first and child-first, each in a fresh interpreter, over a synthetic pair and six real models (hopopt/ipv6_opts QuickStartOption, SMFDPDOption, ipv4.QSOption, tcp.MPTCPJoin). On main's infoclass.py all six parent-first cases fail; child-first passes on both.
• test_info_roundtrip_unit.py: the fix(corekit): a non-final Info subclass built after its parent leaks internal keys from to_dict() #1490 Gap row is gone and the root-cause pin asserts ['a', 'b'].
• test_final_enforcement.py: two behaviour changes follow. A @final class under a BASE ancestor is now refused, closing the gap that test documented. A child declared before its parent was finalised now gets its own __init__.

Ran: the new test, tests/corekit, tests/protocols/internet, tests/protocols/transport, test_capture_roundtrip_runtime.py, tests/project, test_docstring_contract.py, test_final_enforcement.py, all green. repr(frame.info.to_dict()) hashes over examples/captures (23 files) are identical to main.

Closes #1490

[JarryShaw]

Verdict: GOOD TO GO at 9e86cf01f. Cross-reviewed on Sonnet; the author ran on Opus.

• The fix works. I measured it myself: on origin/main sources the new tests give 7 failed, 1 passed. At head, the new file, test_final_enforcement.py and test_info_roundtrip_unit.py give 40 passed, 1458 subtests.
• It is order-independent. Building all 489 Info subclasses parent-first leaks internal keys from 32 classes on main, and from none at head. Child-first gives 0 on both. The generated __init__ qualnames match between the two orders at head.
• The @final refusal is a fix, not a contract break. The old test was added by design: #514 recorded two rulings it never implemented — the Base/public collapse and the enforcing @final #778 as a documented gap, not a guarantee; I checked its docstring at ccf5f623b. Importing all 485 modules and building every Info subclass raises 0 InfoError on either tree.
• Regressions: the per-capture to_dict() hashes are identical across all 23 captures. tests/corekit plus enforcement: 501 passed. Internet plus transport: 692 passed. util/run_unittest_leg.py corekit: 660 tests, 0 failures.

Non-blocking: a user subclass with no __init__, under a parent with a hand-written one that has already been built, now gets a generated __init__ rather than inheriting the custom one. That matches what main already did in the other build order, and no shipped class is affected. It is worth a line in the body.

Schema has the same inherited-state check; that is filed separately.

[JarryShaw]

Verdict carried to e7d02d195: GOOD TO GO. This is a pure rebase onto main at 00aa1fc90, after #1493 and #1494 merged. range-diff shows = and the stable patch-id is unchanged (f18a09f7cc93).

[github-actions]

Coverage: 89.65% (unit tier, Python 3.14, e7d02d195, 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	2115	66	680	33	96.17%
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.