[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. Narrow runs only, listed below.

• 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

Closes #1680.

• Fix: IPv6.read raises ProtocolError("[IPv6] invalid version: N") for any Version but 6, as IPv4.read does for 4 (owner ruling). The default engine only; the adapters are unchanged, per the best-effort ruling on fix(protocols): the default engine accepts an IPv6 Jumbo Payload option of 65535 octets or less, against RFC 2675 #1687.
• Effect: a frame whose IPv6 header has the wrong Version now stops at its link layer, or at a tunnel's inner header, and rebuilds byte for byte. On main, rebuilding such a frame wrote Version 6 (60... for 40...).
• Captures: none of the 23 generated captures change: 1616 frames, 247 IPv6 layers on both trees.
• Make: IPv6.make already hard-codes Version 6, so make->parse is unaffected.
• Tests:
  • New: test_ipv6_version_1680_unit.py, plus an ipv6/version-4 edge round-trip case. With the check removed, the three changed modules show 78 failures.
  • Updated: test_pcap_tcp_segment_ipv6_version_1666_unit.py pinned the old accept-anything reading. It now checks that the default engine reads only the controls, and checks dpkt and scapy against fixed values.
  • Narrow runs on 3.14: tests/protocols and tests/toolkit pass, as do tests/foundation, tests/integration, tests/interface and tests/cli.

[JarryShaw]

Verdict: GOOD TO GO at db9193d25. A Sonnet cross-review checked it; the author was Opus.

• Matches IPv4: the check is the first statement after the schema read, as in IPv4.read. The message is the same apart from the protocol name, and both raise ProtocolError.
• Mutation: turning the check off gives 78 failed, 12 passed. Restoring it gives 12 passed.
• Captures: an independent script compared the default engine before and after the fix over all 23 captures: 0 of 1616 frames differ.
• Hand-built frames: Versions 0, 4, 5, 7 and 15 behind Ethernet behave like the IPv4 sibling. So do linktype 229, 6in4 and 6in6.
• Make path: ipv6.py:319 hard-codes Version 6, so make->parse cannot produce another Version.
• Scope: no adapter or engine file is touched.
• Non-blocking: the rewritten dpkt/scapy test pins TCP inputs, trace inputs and flows, but no longer the IPv4, IPv6 and datagram aspects. That follows refactor(toolkit): revert the cross-engine agreement tests and the adapter code that only mirrors the default engine #1708.

[github-actions]

Coverage: 89.42% (unit tier, Python 3.15, db9193d25, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	3	20	3	95.16%
pcapkit/const	18860	954	2368	868	91.06%
pcapkit/corekit	2372	67	762	35	96.49%
pcapkit/dumpkit	275	3	100	3	98.40%
pcapkit/foundation	2887	144	1012	34	94.20%
pcapkit/interface	138	7	52	5	93.68%
pcapkit/protocols	17183	226	4592	190	98.07%
pcapkit/toolkit	1313	280	532	5	75.88%
pcapkit/utilities	500	4	150	4	98.77%
pcapkit/vendor	4417	2342	1010	157	43.38%

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