[JarryShaw]

• Searched for similar pull requests
• Followed the coding style
• make test passes, and a test case covers the change
• 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?

• docs — documentation only

---

Description

Part of #1103, ruling 1. Test narratives in tests/const/ describe the behaviour instead of the change that produced it.

• Rewrites:
  • "X now does Y" becomes "X does Y".
  • "used to …", "before #NNN …" and "the old …" become a statement of the defect, with its issue: "Without fix(const): Method.get never checks values, so BASELINE-CONTROL and VERSION-CONTROL mis-resolve #908's fix …", "Guards against design: AppType is str-valued with 1,004 port=-1 members, unlike its int-valued sibling registries #732's …".
  • History that only explained how the code got here is cut or compressed. This includes the module docstring of test_const_enum_no_mint.py and several retired-test notes.
• Stale claims corrected against the code:
  • TransportProtocol.get raises a KeyError, not a ValueError.
  • Port 3 resolves as sctp.
  • EtherType no longer mints.
• Kept:
  • "no longer" and "now" where they describe the runtime or a step of the test.
  • Change-relative assertion messages and test names. These are code, and other files cite them.
• Scope: comments and docstrings only. With docstrings stripped, every changed file has the same AST as main.
• Tests run: the slice by group, plus the tier guard, the docstring contract and the RFC-anchor and doc-ref project tests, all pass.
• Overlap: none with open PRs.

[JarryShaw]

Review: NEEDS CHANGES at 4ca9bf5d4 (Sonnet).

• Change-relative wording still in place: "Fails on stock <sha>" and "fix(reg)!: drop TransportProtocol's IntFlag base now that nothing composes #836's own prior head" proofs were reflowed rather than cut, and "this batch" / "this tier" wording remains.
• False present tense: two docstrings say mixed registries "keep minting".
• Garbled or misattributed: the _missing_ in the four mh flag enums ends in return cls(value), so any non-member value recurses — including 0 #623 and AppType.get(name) never matches an existing member and silently mints a port=-1 member instead #734 sentences.
• What checks out: the three corrected claims, the AST guard and every test. The worker is fixing the rest.

[JarryShaw]

Review fixes pushed: 4ca9bf5d4 → 525dc7350, rebased onto main.

• Proofs: every "Fails on stock <sha>" and "fix(reg)!: drop TransportProtocol's IntFlag base now that nothing composes #836's prior head" proof now states the defect directly. The grep also turned up and fixed further dated shas.
• Timed wording: "this batch" and "this tier" are gone.
• Mint claims: the two "keeps minting" docstrings now match the code; both registries resolve unregistered values without minting.
• _missing_ in the four mh flag enums ends in return cls(value), so any non-member value recurses — including 0 #623, AppType.get(name) never matches an existing member and silently mints a port=-1 member instead #734, Method registered members carry an empty str payload, so Method.GET == GET is False #870: these sentences are corrected against the code.
• tcp | dccp: checked and kept. It equals 5, and a test already asserts that it is refused.
• Checks: the AST check is clean and the slice's tests pass. A delta review is running.

[JarryShaw]

Review: GOOD TO GO at 525dc7350 (Sonnet delta).

• Change-relative wording: every dated sha and every prior-head, this-batch and this-tier proof is gone from all 17 files.
• Mint claims: now true. EtherType.get(0x0888) and Socket(0x0010) resolve unregistered.
• _missing_ in the four mh flag enums ends in return cls(value), so any non-member value recurses — including 0 #623, AppType.get(name) never matches an existing member and silently mints a port=-1 member instead #734, Method registered members carry an empty str payload, so Method.GET == GET is False #870/fix(const,vendor): bring 8 bespoke registries onto EnumRegistry, stop them minting #869 and tcp | dccp: each checked against the code.
• Checks: AST is clean and the slice tests pass.
• Left for the next docs: #719-style prose sweep after the #1090–#1102 defect wave #1103 round: some tier-2 history in test_const_enum_no_mint.py docstrings, and "40 declared members" in the Method registered members carry an empty str payload, so Method.GET == GET is False #870 module docstring.

[JarryShaw]

Ready to merge at 525dc7350. Part of #1103 (ruling 1).

• Review: Sonnet GOOD TO GO.
• CI: all checks complete, Required checks passed SUCCESS, 0 failed.
• KTLO: 12 legs, no warnings.
• Ordering: const leg merged into main: 597 tests, 0 failures.
• Conflicts: none with the current main.

[github-actions]

Coverage: 91.55% (unit tier, Python 3.15, 525dc7350, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	2	95.16%
pcapkit/const	19599	880	2618	866	92.12%
pcapkit/corekit	2589	75	816	37	96.48%
pcapkit/dumpkit	282	3	102	3	98.44%
pcapkit/foundation	2971	23	1036	25	98.80%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17428	219	4584	185	98.15%
pcapkit/toolkit	869	36	298	10	95.20%
pcapkit/utilities	504	7	146	5	98.15%
pcapkit/vendor	4432	1988	1018	167	51.87%

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